<![CDATA[ Latest from TechRadar in Opinion ]]>

<![CDATA[ Latest from TechRadar in Opinion ]]> https://www.techradar.com Tue, 27 Jan 2026 00:00:00 +0000 en <![CDATA[ RGB TVs are a huge danger to OLED TVs — and that should be the best news OLED fans have heard in years ]]> 2026 is going to be the year of RGB TVs. The biggest TV makers have sets coming out using next-gen RGB backlight technology, which basically replaces the blue or white backlight of a traditional mini-LED LCD TV with one that has full RGB color support, meaning that these TVs are more efficient, offer richer and more accurate colors, and can suffer less light leakage into dark areas.

Obviously, we need to wait and see just how well these TVs perform in practice – but TechRadar's TV team has seen several of them in early demos, and they're incredibly impressive. Wide viewing angles (often an issue with LCD TVs), vibrant yet realistic colors, inky deep black tones, powerful HDR highlights… they look like a huge danger to the best OLED TVs.

That's in no small part because RGB TVs are set to arrive at only slightly higher prices than standard mini-LED technology, and that's in their first year – think about how quickly standard mini-LED prices have dropped in the few years since their inception, meaning you can get a large-screen, great-quality mini-LED set for under $500 / £500 these days.

My colleague Al Griffin, a TV industry veteran of nearly 30 years, even went so far as to say the budget OLED TV is doomed. He thinks that RGB TVs' contrast performance is so close to non-flagship OLED TVs, with richer colors and far superior brightness, that cheaper OLED TVs will become "irrelevant".

I agree with him in principle, although the news that LG Display's new, cheaper, and simultaneously brighter new cheap OLED panel is coming this year probably means that reports of budget OLED's demise are slightly exaggerated.

But I can see a different potential path forward, I hope: while I agree with Al's sentiment about the current state of the industry, I think the aggressive expansion of RGB TVs is the best thing that could happen for OLED TV fans.

LG C5 (left) and LG B5 (right) showing orange butterfly on screen — The LG C5 and LG B5: have they had it too good for too long? (Image credit: Future)

At the high end of OLED TVs (like the LG G5 and the Samsung S95F), there's been significant development over the last few years – micro lens array, QD-OLED, Primary RGB Tandem, next-gen quantum dots.

The progress in mini-LED TVs, plus the arrival of competition for the creation of OLED panels between LG and Samsung, lit a fire under the high-end OLED world. It's improving faster now than at any time in the 13 years OLED TVs have been mainstream.

But at the more affordable end? Not so much.

The LG C5 OLED uses essentially the same panel as the LG C4, and as the LG C3 – and the LG C6 also has it (although there will be a new LG C6H at larger sizes with a better panel).

The LG B5 affordable OLED uses basically the same panel as the LG B4, the B3, and… you get it.

And that's been okay, because the TVs have been very high quality, and mini-LED TVs haven't really been able to match OLED in the key area that people buy them for: the perfect black tones maintained down to the individual pixel.

That truly cinematic look just hasn't come across the same in mid-range mini-LEDs… but RGB TV could change that. Sony told me that one of the advantages of using colorful backlights is that some light wavelengths are more easily absorbed in black areas than others, meaning that there will naturally be less 'blooming' from light areas to dark, as you get in current mini-LED TVs.

If mid-range RGB TVs get close to the contrast of cheaper OLEDs while outperforming them in color and brightness, even purists may turn to the RGB side.

Samsung's Micro RGB TV at CES 2026, with a sign saying Micro RGB 130 — The power of the colors in RGB TVs is something to behold… (Image credit: Future)

Why is this still a problem for OLED?

The problem for OLED TVs has always been in the manufacturing. The materials needed for the organic pixels haven't changed enough in price over time, and the complicated nature of depositing the material hasn't shifted enough either. Yield rate is another major issue that still holds it back – the manufacturing process just isn't reliable enough to keep prices from dropping.

There have been breakthroughs in areas like the holy grail new blue phosphor material, or the developments in inkjet-printed OLED materials – but I've been told by insiders that the latter is probably three years away from even starting to be used on TV-sized panels, and blue phosphor just won't be enough on its own.

The costs of making OLED panels just haven't changed enough over time to have ever made them truly cheap TVs, and while there's no magic wand for making them less expensive suddenly now, there's certainly more incentive.

Nothing motivates like an existential threat – and RGB TVs certainly have the potential to be just that for today's affordable OLEDs.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/televisions/rgb-tvs-are-a-huge-danger-to-oled-tvs-and-that-should-be-the-best-news-oled-fans-have-heard-in-years Fxwx9TBajmGuDHKjfJMFhk Tue, 27 Jan 2026 00:00:00 +0000 Mon, 26 Jan 2026 20:14:37 +0000 <![CDATA[ 3 risks hindering enterprise-ready AI — and how low-code workflows help ]]> You’ve likely heard of agentic AI by now: systems that can autonomously plan, execute, and adapt across tasks with “minimal” human oversight (quotations intentional here). In this, we’ve seen a shift from AI as a mere tool to now AI as a collaborator.

However, there is also a growing tension: Agentic AI needs deep access to your data to act autonomously, but that level of access makes it hard to deploy safely and responsibly at scale.

Data access without proper guardrails makes agentic AI harder to audit and integrate into environments where trust, data governance, and maybe even operational complexity, are non-negotiable.

Most enterprises don’t struggle to build agentic systems — they struggle to trust them.

In this article, we’ll explore three risks holding agentic AI back from enterprise-readiness, especially when it comes to agents working with data, and then three ways intuitive, low-code workflows can turn these systems into reliable colleagues.

Workflows don’t limit the “intelligence” of agentic AI, but rather, act as a “safe layer” between agentic AI and your data, making it possible to operationalize agentic AI in the enterprise.

Risk #1: No transparency in decision making

Most AI agents today rely on LLMs as the planners or “brains” behind the scenes. This means most agents don’t follow predefined blueprints or logic. The way they work is dynamic and always changing. Their actions are based on likelihood derived from vast datasets — not knowledge.

We’ve all seen it: an AI telling us what we want to hear instead of what’s true, simply because we (subconsciously or not) led it in the wrong direction with our responses. Think of that famous example where someone convinced an AI that 2+2=5.

The result? These actions are difficult to inspect, explain, or trace. Without a clear, visible audit trail, enterprises cannot confidently answer a critical question: “Why did the agent do that?” Especially when the agent takes an unexpected action.

Ultimately, this makes debugging challenging, and likely miserable. Instead of systemic debugging, enterprise teams have the time-consuming task of second-guessing the agent’s behavior. This slow, manual, error-prone, unscalable process of “prompt forensics” is ineffective for enterprises.

If you can’t trace it, you can’t trust it. And on the topic of trust…

Risk #2: Indeterminism means no operational trust

Agentic AI is not deterministic, which means it doesn’t produce consistent, repeatable outputs. Identical tasks could yield different actions. In addition, agents could hallucinate actions based on what could be plausible, but end up being just plain wrong.

And there is often no built-in layer to enforce or constrain what an agent can or can’t do.

This is particularly high-risk for enterprises, like financial systems or anything touching personal data, where data leakage is unacceptable. Especially in those cases, lack of consistency, transparency, explainability, and control ultimately leads to lack of trust.

Risk #3: No clear boundary between data and AI

In traditional enterprise systems, data and logic are clearly separated. IT teams know where the data is stored, how it’s accessed (whether through permissions or trust), and there’s an explicit code or set of rules that govern how this data is used.

Agentic systems break these rules. They blend reasoning, knowledge, and actions into an opaque process. Drawing a clear boundary line between what information the agent has access to versus what the agent does can be challenging, and in some instances, impossible.

The lack of separation is not only high-risk — it’s a dealbreaker. Enterprises are legally required to meet compliance and governance standards. This lack of a clear boundary discourages enterprises from AI adoption.

So, what can we do to mitigate these risks and (safely) benefit from agentic AI — and encourage adoption — in business? Or better yet, how can agents reliably work with data?

Workflows as a unifying language and bridge for enterprises and agentic AI

The answer lies in transparency. Intuitive, low-code workflows bring in that transparency, acting as a clear separation between agents and your data. Workflows force agents to interact with tools, not directly with the data.

While agentic systems are powerful because they can reason with minimal human input, workflows rein in that power and build trust by setting a defined, structured path for how these agentic systems can operate. Workflows bring control, clarity, and repeatability to dynamic and uncertain systems.

1. Workflows allow for auditability

Workflows being visual in nature, each step — and each potential failure point — is more visible. The decision-making process is more clearly documented. The outputs are controllable and explainable.

Also, the visual nature of workflows makes it an intuitive format. It allows teams with varying levels of technicality to have the ability “to speak the same language,” in contrast to a mess of SQL, Python, or other code that other solutions may come with.

This makes debugging and monitoring much more straight-forward for enterprise teams.

2. Workflows allow for trustworthy guardrails and reusability

Workflows reduce risk because workflows set what data and tools agentic systems can access, and in what level of detail. Decision-makers can define this explicitly company-wide.

Additionally, once these approvals and logic have been set, workflows allow for reusability and scalability. Enterprises can reuse these validated blueprints and implement workflows into other parts of the business without reinventing the wheel — or at the very least, be a reliable starting point for other projects.

3. Workflows allow for governance and accountability

Workflows enforce guardrails, observability, and accountability. By being the clear separation between data and AI (Remember, what the agent knows versus what the agent does), enterprises have complete governance. Organizations can protect data, monitor data access, and audit data lineage.

Put simply: Workflows make sure agentic AI uses your data properly… and doesn’t abuse it!

Agentic AI is undeniably valuable in the enterprise context. Even with these risks, there doesn’t have to be a tradeoff between transparency and complexity. By enforcing workflows as the safety layer to your agentic work, you allow for visual, modular, and governable ways to build intelligent agents that enterprises can trust and scale.

Again: You don’t give agents access to your data. You give agents access to your tools, which keeps your data protected from attacks or misuse. Agentic AI is not limited by workflows.

Rather, these systems have more “freedom” to do cool things when operating within the data-safe boundaries of well-defined workflows as a language. And as a sidenote: This is nicely in line with newer trends to provide agents with sets of skills instead of detailed instructions on how to use hundreds of tools.

Just like Pat Boone once said, “Freedom isn’t the absence of boundaries. It’s the ability to operate successfully and happily within boundaries.”

We've featured the best business intelligence platform.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

]]> https://www.techradar.com/pro/3-risks-hindering-enterprise-ready-ai-and-how-low-code-workflows-help QwwCNNSCZrpJtYhLfMQADh Mon, 26 Jan 2026 15:27:47 +0000 Mon, 26 Jan 2026 15:27:47 +0000 <![CDATA[ AI projects worldwide are failing in businesses because of this simple reason ]]> Artificial intelligence is dominating conversations in boardrooms across the country. CEOs often see it as a transformative tool capable of reshaping businesses, creating efficiencies, and opening new opportunities.

Meanwhile, IT leaders - those responsible for bringing such visions to life - frequently approach AI with caution, aware of the complexities involved in integration, security and scaling.

In mid-market companies, where leadership teams are smaller and more tightly connected, one might assume alignment would come naturally. Yet differences in perspective can still stall progress before it begins. CEOs often focus on ambition and opportunity, while IT leaders focus on practicality and feasibility.

Without a shared approach, AI initiatives risk remaining aspirational rather than impactful.

Understanding the evolving IT landscape

The role of IT has fundamentally changed. Once, CIOs held specialized expertise that set them apart. Today, technology is central to almost every aspect of business and leaders at all levels recognize its importance.

At the same time, technology is increasingly familiar to everyone, from smartphones to AI-powered tools, which raises expectations and lowers tolerance for slow or inflexible systems.

In mid-market teams, where skills often overlap and people wear multiple hats, this creates both opportunity and tension. CEOs may have experience with business systems or even basic coding, while IT leaders may have operational experience beyond technology.

But their roles will often mean they get their information and advice from very different places. Without clear alignment, these overlapping perspectives can blur responsibilities and create misaligned expectations.

Building a shared language

Bridging the gap begins with creating a shared language for discussing AI. Leaders need to move beyond abstract notions of “transformation” and focus on specific outcomes that matter to the business.

Framing AI in terms of tangible goals - efficiency gains, improved customer experiences, or enhanced decision-making - helps everyone understand both the ambition and the constraints.

Equally important is transparency around risk and readiness. By openly discussing infrastructure limitations, integration challenges, and workforce capabilities, boards can set realistic expectations without dampening enthusiasm.

IT leaders can play a critical role in translating technical constraints into business terms, while executives can clarify strategic priorities, ensuring everyone is focused on the same objectives.

Creating a practical AI roadmap

A successful AI plan is both aspirational and actionable. Mid-market boards can adopt a phased approach: identifying high-impact opportunities, piloting projects in manageable scope and scaling gradually based on measurable results.

Clear milestones, KPIs and accountability ensure initiatives do not stall, and regular review cycles allow strategies to evolve alongside business needs.

To support adoption, boards can encourage cross-functional collaboration. Bringing finance, operations, marketing, and IT together ensures that AI initiatives are practical, achievable, and aligned with overall business strategy.

Small, cross-disciplinary teams can surface issues early, avoid duplication, and ensure that pilots are designed for success before broader rollouts.

Encouraging cultural alignment

Technology adoption is as much about people as it is about systems. Boards need to foster a culture where experimentation is safe, learning is continuous, and failure is seen as an opportunity to iterate. Leaders can encourage this by championing small wins, celebrating early successes and sharing learnings openly across teams.

When the executive team models curiosity and openness to AI, it sets the tone for the wider organization.

Mid-market companies can also invest in upskilling staff, helping both technical and non-technical team members understand the possibilities and limitations of AI. This shared understanding reduces friction, empowers decision-making and ensures that AI is used effectively across departments rather than siloed within IT.

Turning alignment into impact

The gap between CEOs and IT leaders is real, but it is not insurmountable. Boards that create a shared understanding, define tangible goals, and build phased, measurable plans will see AI move from a buzzword to a practical business lever.

Mid-market organizations, with their lean teams and agility, are uniquely positioned to adopt AI rapidly and scale it efficiently, delivering outcomes that truly matter.

Bridging the divide requires commitment, communication and clarity. It requires executives to appreciate technical realities and IT leaders to think strategically about business impact.

The reward is clear: a board that speaks the same language can turn AI from ambition into action, potential into performance, and ideas into measurable results.

We've featured the best AI website builder.

]]> https://www.techradar.com/pro/ai-projects-worldwide-are-failing-in-businesses-because-of-this-simple-reason vb3UdYwLc8G3JjgHY8YMNS Mon, 26 Jan 2026 15:07:33 +0000 Mon, 26 Jan 2026 15:07:34 +0000 <![CDATA[ I tested the three most popular Nintendo Switch 2 cameras, and this is the only one I’d recommend ]]> When the Nintendo Switch 2 was first shown off last year, one of the most intriguing additions to the system was its GameChat functionality. Just press the C button, and you can chat with family or friends, and soak in every reaction – whether you’re overtaking on the last lap in Mario Kart or winning a minigame in Mario Party.

In order to enjoy GameChat at its best, I’d suggest picking up a camera for your Nintendo Switch 2. This will open up video capabilities, which aren’t only used for GameChat, but also for certain content in titles such as Super Mario Party Jamboree + Jamboree TV and the upcoming Switch 2 version of Animal Crossing: New Horizons.

But which Nintendo Switch 2 camera is the best? After testing three of the most popular models, I have one clear answer. That’s right, I’ve tried out the official Nintendo Switch 2 Camera, the Hori Piranha Plant Camera, and the Hori USB Camera to help you select the best model for your shiny new console.

I’ll compare all of these across three key metrics: design, performance, and value for money. Then, I’ll give my verdict on which camera is the best Nintendo Switch 2 accessory, and what truly sets it apart. With all of that said, let’s get right into our comparison.

Design: zooming in on the details

Nintendo Switch 2 camera, Hori Piranha Plant Camera, and Hori USB Camera all together — (Image credit: Future)

Now, in terms of design, there are a lot of similarities between Hori’s USB and Piranha Plant cameras. One key benefit these share is that they’re optimized for both TV and handheld play. When playing docked, you can place the cameras on their respective base and either mount them or place them on a flat surface. When you’re on the go, you can instead plug the webcam directly into the system’s top USB-C slot, which is incredibly practical.

Meanwhile, the official Switch 2 camera isn’t quite as versatile. It comes with a secure stand, which is optimal for docked gameplay, but isn’t detachable – yep, no specialized handheld mode here, folks.

Even if the official cam is a little more rigid than its rivals, one area it propels itself above them is in build quality. The stem and base of this camera are metallic and very well-made, giving the product a premium look and feel. It isn’t a very vibrant or exciting-looking piece of kit, but again, it’s got a sleek, elegant build, which feels on-brand with the Switch 2 console.

On the other hand, I’m not a massive lover of the Hori USB camera’s aesthetic. Sure, it’s ultra-compact, making it easy to throw in a case like the official Nintendo Switch 2 Carrying Case, but it does look a little bland.

The camera with the most pizzazz is clearly the Hori Piranha Plant Camera for Nintendo Switch 2. This model, which beautifully replicates the classic Super Mario baddie, has a closable mouth, which acts as a privacy shutter, an attachable leaf part, and even a base that looks like its iconic plant pot. It's colorful and eye-catching, which combined with its flexibility, makes it the best-designed model of the three, in my view.

Performance: lights, camera, action

Upon its release, the Hori Piranha Plant Camera was met with a fair share of criticism. Not for its design or for its price… but for its performance. See, this thing’s resolution is just 480p, which will seem very archaic to a lot of users.

What does this mean in practice? Well, you can expect unrefined results, with blurry faces, obscured details, and washed-out colors. What’s more, this remains the case for the Hori USB camera as well, meaning I can’t recommend either for those seeking out good picture quality.

These cameras also have a fairly limited field of view – 85 degrees to be exact. So, in the situation that you want to capture a bunch of players in your living room, the scope of these models may be too limiting. The only aspect that these cameras excel in, performance-wise, is tracking. They’ll frame your face pretty effectively in games like Mario Kart World – just don’t expect a sharp picture, especially from a distance.

It’s a very different story with the official Nintendo Switch 2 camera. Its 1080p resolution is a gigantic upgrade on the Hori models, producing a more detailed, color-accurate, and dynamic picture time after time. Still, 1080p is by no means revolutionary, and I picked up on some surroundings appearing a tad fuzzy, but this camera is certainly sharp enough for most users.

The official Switch 2 camera also has a better field of view – about 110 degrees, which makes it ideal for capturing larger groups. It runs at 30fps – the same as both Hori models, and all three cameras offered consistent motion in my testing. All things considered, then, the official Switch 2 camera is the clear winner when it comes to performance.

Value: which camera is really worth your money?

Finally, let’s talk about value for money, which is arguably the most important category of all. These cameras vary a little in price, with the Hori Piranha Plant typically coming in at $39.99 / £24.99 / AU$64.95 (after a widespread price-cut). Hori’s USB Camera usually sits at around the same mark, but the official Switch 2 camera has a list price of $54.99 / £49.99 / AU$69.95.

In this case, though, I think the additional spend is well worth it. The official camera may not be as versatile and compact as the Hori USB Camera or as groovy as the Piranha Plant, but its picture quality is so much better than theirs. It also has a robust, sleek design, and a neat privacy shutter of its own. As a result, it’s the only model I’d recommend outright for anyone looking to take their GameChat experiences up a notch.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/gaming/nintendo-switch/i-tested-the-three-most-popular-nintendo-switch-2-cameras-and-this-is-the-only-one-id-recommend vsRVMoFkdvrtQuDC7gWrK5 Sun, 25 Jan 2026 19:00:00 +0000 Fri, 09 Jan 2026 15:24:14 +0000 <![CDATA[ What ‘vibe coding’ means for API platforms and the future of DevRel ]]> AI-assisted code is becoming a standard part of many developers’ daily workflows, and AI-driven tools are now directly targeting the broader software development lifecycle.

For instance, at Amazon Web Services’s 2025 re:Invent in December, AWS launched a new class of autonomous, long-running ‘frontier agents’, including a coding agent, a security agent, and a DevOps agent, each designed to work for hours or days on behalf of development teams.

These developments reflect a broader shift: organizations are increasingly seeing AI not just as an assistive typing tool they can use to develop proof-of-concepts and prototypes, but as a partner they can use throughout the development life cycle, capable of generating integration code, handling security reviews, or even triaging operations issues automatically.

As a result, what started as ‘vibe coding’, the informal, exploratory use of AI code generation, is rapidly becoming intrinsic to many teams’ development practices.

A new dual audience for API platforms: humans and AI agents

With AI agents actively participating in creating code, testing, and operations, the ‘consumer’ of your API platform is now extended beyond just human developers. Platforms now need to be built not only for humans, with rich narrative documentation, guides, and tutorials - but also for machines.

AI agents benefit from structured, predictable APIs: clear endpoint definitions, consistent naming, unambiguous parameter types, and machine-readable metadata.

If an API is easy for a human to read but ambiguous for a tool (e.g., inconsistent naming, missing schema, edge-case behaviors omitted), the first integration attempt from an AI-driven tool may fail or misbehave.

This means API providers should treat machine-readability as a first-class design goal, as part of the ‘definition of done’ - not optional. In effect, documentation, SDKs, discovery models and metadata outputs should be optimized for both human and agent ingestion.

Industry research supports that this shift is already underway: while 89% of developers now use generative AI in their work, only 24% of organizations currently design APIs with AI agents in mind.

This gap suggests many platforms remain optimized solely for human users - a misalignment that may cost them relevance as agentic development becomes more common.

What this means for API-first platforms and DevRel

Platform teams should now view AI readiness as a core element of API design. This means greater discipline around endpoint consistency, schema stability and naming conventions, supported by documentation and metadata that can be consumed programmatically.

When these foundations are in place, machine agents are far more likely to produce correct integration code on the first attempt, which reduces friction for both humans and their AI counterparts.

The discovery surfaces that platforms expose also matter more than before.

Auto-generated OpenAPI or Swagger schemas, structured metadata endpoints and machine-friendly SDKs give agents the clarity they need to understand available functionality and select the right paths through an API. In practice, this means treating metadata as a strategic asset rather than a by-product of engineering.

Teams should also anticipate that first impressions will increasingly be shaped by automated agents rather than human developers.

The moment an AI agent successfully returns a 200 OK is becoming as important as a developer reading a polished README, because it determines whether the agent continues to attempt deeper integration or quickly turns elsewhere.

For DevRel and developer experience teams

Developer Relations and DevX teams will need to reassess how they measure impact in a world where agents initiate a growing share of platform usage.

Metrics like forum activity, tutorial completions or SDK downloads may no longer offer a full picture of adoption. Instead, teams should track how often AI systems attempt integrations, how frequently those integrations succeed and where agent-driven errors occur.

This shift opens up a new responsibility to provide AI-friendly tooling that guides both developers and their copilots. Machine-readable reference documentation, prompt templates, example snippets designed for code generation and environments that help teams audit or refine AI-generated code will all become increasingly useful.

Above all, DevRel teams should begin to think of agents as a first-class audience. That means investing in predictable schema design, clear behavioral models and error handling that is explicit enough for an agent to learn from.

Supporting developers now means supporting both the humans doing the building and the AI systems helping them do it.

First-mover advantage for ‘AI-Ready’ APIs

As agentic AI tools continue to grow in popularity, platforms that adapt early to machine-readability will gain a competitive edge. Their APIs will be easier for AI agents to integrate, more predictable, and more likely to be the first successful target the agent tries - giving them early adoption advantage.

Teams that wait risk being bypassed, ignored, or causing friction that pushes developers (or their agentic copilots) elsewhere.

Over time, ‘vibe coding’ will simply become ‘coding’. The software development lifecycle (SDLC) will increasingly include AI agents as first-class participants - and platform readiness for those agents will be a key differentiator.

We've featured the best text editor for coding.

]]> https://www.techradar.com/pro/what-vibe-coding-means-for-api-platforms-and-the-future-of-devrel sCA6mPy9bKtv5pPuh8FExV Sun, 25 Jan 2026 15:00:00 +0000 Fri, 23 Jan 2026 15:02:06 +0000 <![CDATA[ We now know the magic behind Samsung new brighter OLED TV tech, but it still hasn’t solved the biggest problem ]]> OLED TVs are among the most popular on the market. Thanks to OLED’s self-emissive pixels, it can produce perfect blacks while also delivering crisp textures, bold colors and strong contrast. Traditional TV OLED tech, known as Tandem WOLED now, has always struggled with its brightness, measuring significantly dimmer than mini-LED TVs.

Enter QD-OLED. Introduced by Samsung in 2022, QD-OLED adds quantum dots into the OLED equation, enabling very high brightness. While WOLED itself is getting brighter, QD-OLED continues to promise higher brightness increases than WOLED can generally manage each year – and we seem to know now the secret sauce that's allowed a huge 35% claimed jump in 2026, which I'll come back to shortly.

But QD-OLEDs advances aren't helping it take over the industry. Both OLED and QD-OLED are used in some of the best TVs, but WOLED continues to be the most widely used, even though QD-OLED is the brighter of the two panel technologies. In 2023, I said that WOLED was dying thanks to the arrival of QD-OLED and since then, a QD-OLED TV has been named TechRadar’s TV of the Year each year since, last year won by the Samsung S95F.

So, why hasn’t QD-OLED fully taken over? Let’s explore.

QD-OLED vs OLED brightness

Samsung S95F vs LG G5 with peacock feather on screen — The Samsung S95F (left) exhibits higher perceived brightness, despite measuring lower than the LG G5 (right) (Image credit: Future)

When it comes to brightness, QD-OLED usually comes out on top. Last year however, LG introduced a new OLED panel: the Primary RGB Tandem OLED, also known as ‘four-stack’ OLED. It delivered higher brightness and richer colors, and it even beat QD-OLED in one area when we tested LG and Samsung’s flagship models last year.

We measured the LG G5’s peak HDR brightness at 2,268 nits, which was higher than the Samsung S95F’s peak HDR brightness of 2,132 nits. However, where QD-OLED had the advantage was fullscreen HDR brightness, with the S95F hitting 390 nits and the G5 clocking in at 331 nits, and this is more impactful measure for real-life viewing.

It’s worth noting that the RGB Tandem OLED was reserved for the flagship models. The mid-range LG C5, and its older WOLED OLED EX panel, hit 1,180 nits peak brightness – whereas the 65-inch Samsung S90F, which used a QD-OLED panel, hit 1,400 nits. The mid-range is where the QD-OLED made the most difference.

QD-OLED vs OLED’s future

Image 1 of 2

The Samsung S95H at CES 2026 — The Samsung S95H could be up to 35% brighter than last year… (Image credit: Future)

Image 2 of 2

LG G6 TV at CES 2026 — …whereas the LG G6 (2) is set to be 20% brighter. (Image credit: Future)

At CES 2026, LG revealed its 2026 OLED TV lineup, announcing that both the LG G6 and LG C6H, the latter of which is basically the LG C6 at 77-inch and 83-inch sizes, will use a next-gen Primary RGB Tandem 2.0 panel. LG says that the G6 will be 20% brighter than its predecessor the G5.

While this is impressive, Samsung revealed its own 2026 OLED TV lineup and we learned that the Samsung S95H, its 2026 flagship, should deliver 35% higher brightness than its predecessor the Samsung S95F. That’s a big boost and dwarfs LG’s 20% higher brightness claim. How is such a boost possible?

Well, we actually learned from Nanosys, the manufacturer behind quantum dots, that QD-OLED will get brighter thanks (at least in part) to a new version of the Quantum Dot Color Conversion (QDCC). QDCC is a type of ink which is used in QD-OLED and this new version is more efficient, enabling higher brightness levels. Other factors probably play in too, but this seems like the key technical development.

So, with all these exciting updates, why is QD-OLED not the most widely used OLED tech?

QD-OLED’s downfall: price

Samsung S95F with canyon on screen — For all their positives, QD-OLED TVs (such as the Samsung S95F pictured) are still pricier and available in less sizes than WOLED TVs. (Image credit: Future)

It’s no secret that OLED TVs are some of the most premium priced on the market. While there are ‘budget’ options available such as the LG B5 and Samsung S85F, they’re still much pricier than mini-LED TVs. But, while WOLED TVs are getting cheaper and are available in more budget models, QD-OLED isn’t.

Price is QD-OLED’s biggest downfall. As QD-OLED appears to be a more costly panel to make, it hasn’t made its way to cheaper models. In fact, it’s also not available (at the time of writing) in anything smaller than 55-inch, leaving WOLED as the sole OLED panel tech in smaller 42 and 48-inch models. QD-OLED also isn't available at 83-inch sizes.

At launch, a 65-inch Samsung S95F cost $3,399 / £3,399 / AU$5,295, very similarly priced to the 65-inch LG G5 cost $3,399 / £3,299 / AU$5,299. And while QD-OLED’s prices are similar in its available sizes, 55-77-inch, the lack of a smaller, more budget friendly model is noticeable. Samsung actually uses traditional WOLED in its entry-level S85F model, and in some sizes of the Samsung S90F, with QD-OLED seemingly not able to keep up with Samsung's own demands.

What's also very notable is how other manufacturers don't seem desperate to use QD-OLED. Panasonic and Philips use LG's Tandem WOLED displays. Sony has used QD-OLED for its flagship OLED TVs, but Sony has always sold TVs at a price premium compared to other manufacturers – so it being the only user feels telling.

With LG Display recently announcing a lower-cost OLED panel with impressive brightness, and no announcements from Samsung Display that it'll be able to offer cheaper QD-OLED panels or new sizes (so far), it looks like LG still has the lead with other manufacturers where it matters: bang for buck.

I put the LG G5 and Samsung S95F side-by-side and it was an extremely close fight, with both producing incredible pictures. Putting these two models in a four-way OLED showdown, they both came out on top, with the S95F scoring higher in most categories. But, as long as QD-OLED is pricier, I can’t help feeling it’s still one step behind WOLED. If it can get its price down, all bets are off.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/televisions/samsung-qd-oled-nanosys-price-problem rcWczCizHpNrtSGgu3EKJR Sun, 25 Jan 2026 13:00:00 +0000 Fri, 23 Jan 2026 17:20:52 +0000 <![CDATA[ Are robot window cleaners any good? I tested one – here's what you need to know before you buy ]]> Most people have heard of robot vacuums. Perhaps you're even aware of the existence of robot lawn mowers. But did you know that now you can even get your hands on a robot for cleaning your windows?

The first time I saw one of these window bots in action at a tech show, I was mesmerized. I watched, transfixed, as the bot made its way methodically across a pane of glass, gently buffing as it went. I knew I wanted to try one out in my own home, to see if it could move outside its comfort zone of an already-clean bit of demo glass and tackle my grimy windows instead.

For those whose homes have numerous glass windows and doors, a robot window cleaner is undoubtedly an appealing prospect. Manual window cleaning is messy, and generally requires you to hire a professional – in which case you're at the mercy of their schedule. But will a window bot do as good a job? And are there any downsides to outsourcing this task to a robot? Here's my take...

How do robot window cleaners work?

A robot window cleaner sticks to your window, and makes its way methodically across its surface, cleaning dirt and grime as it goes. Typically, such models work by lightly spraying a mix of water and cleaning solution ahead of them, after which their mop pads vibrate to gently buff off any dirt, leaving the surface clean and shiny. Don't expect lots of water or soap suds – this is a relatively dry setup. The machine will adhere to the surface using suction, with the mop pads helping to form a seal around the suction area, which is usually in the middle-centre of the underside of the bot (there will be a safety cable supplied as a backup). To keep the weight down, most window bots are mains-powered.

Hobot S7 Pro window cleaning robot on outside of reviewer's window, photographed from inside — (Image credit: Future)

#1. Cleaning quality can vary

How well does a window bot actually clean, then? Unfortunately I can't give a definitive answer on this one. The window bot I tested did a pretty good job of clearing away surface dirt, but it left visible track marks wherever it pivoted on the glass (read my Hobot S7 Pro windowbot review for more). However, my colleague tested a different model and had nothing but good things to say about its cleaning prowess (more details in her Ecovacs Winbot W2 Pro Omni review).

What I can say for sure is that these cleaning bots aren't really for tackling the likes of bird poop or mud splashes – the cleaning is quite gentle, with not much water applied, so they're really only for surface dirt rather than anything more challenging.

WinBot W2 Omni cleaning window — (Image credit: Future)

#2. You will still need to supervise

While other robot helpers – robot vacuums, for instance – can be left alone to do their thing, window bots require human intervention. For starters, you'll need to be present to place the bot on each pane of glass and switch it on, and then remove it or shift it to the next pane once it's done. While using a window not will certainly save you some effort compared to manual cleaning, you'll need to be around while it's doing so.

There's also the question of how secure it is on the surface you're cleaning. One of the things that impressed me most when I first saw my window bot in action is how it adhered to the surface – it looked almost like magic. I was so confident in its stickiness that I stopped bothering to attach the safety cable when I left it to clean. So it came as quite surprise when, one morning, it simply dropped off the window with no warning.

Hobot S7 Pro window cleaning robot on outside of reviewer's patio door window — (Image credit: Future)

I caught the rope and no damage was done, but there could have been – it was on the outside of a window on the third floor of a town house at the time. It only happened the once, and it seems like a freak occurrence, but I'd say it's still worth being diligent about attaching your safety rope to something, and keeping an eye on the window bot if it's in an area where a fall could potentially do some harm. Note that window bots will generally have a backup battery to keep them stuck for a short period, in the case of a power outage.

#3. They can get places a human can't

Aside from saved effort, the other big benefit of window bots is that they can reach places that human cleaners can't. For instance – the home in which I tested this bot has full-length windows that sit behind protective railings on one floor. For a person trying to clean behind those railings would prove a challenge (to put it mildly), but the windowbot snuck behind them with ease.

Oh, and they're not just for windows, either. Robot window cleaners can theoretically be used on any hard, flat surface. That means you can deploy them on your bathroom mirror or shower screen, saving you effort indoors as well as outside.

]]> https://www.techradar.com/home/small-appliances/are-robot-window-cleaners-any-good-i-tested-one-heres-what-you-need-to-know-before-you-buy 5n6GyEmb9ag9hc5eFrnPxZ Sun, 25 Jan 2026 11:00:00 +0000 Fri, 23 Jan 2026 12:39:36 +0000 <![CDATA[ The NexPhone could already be the most important hardware launch of 2026 — here's why ]]> This week, the NexPhone has been making serious waves. In case you missed the revelation of this piece of hardware, it's a three-operating-systems-in-one smartphone, a seriously clever idea that's set to ship in Q3 2026, and it caught me completely by surprise.

The 6.6-inch NexPhone runs Android, as well as desktop Linux and Windows 11. As was made clear when we interviewed the CEO of the maker Nex Computer, it's essentially trying to bridge the gap between the phone and the PC by turning that phone into a computer.

So, it's a phone that can be a phone (obviously), but it can also boot a desktop OS and effectively be a computer, either running that environment on the small phone display – with tweaks so the OS works better here – or more ideally, connected to an external monitor of some kind, with Windows or Linux depicted in its full glory.

It's not a new idea, but the way it's done here is very remarkable. You can boot to Android or to Linux (Debian) and still access all your files on the Android system. Or you can go in another direction entirely, and boot Windows 11 (without Android file access, mind) – but this isn't some cut-down flavor of Microsoft's desktop OS. This is full-blown Windows 11.

Why am I so stoked about this prospect? Well, firstly, it's just really cool to think you can have your Windows 11 setup accessible on your phone, to the point that it's just a matter of plugging in the handset to a monitor, and away you go (with some additional input peripherals, naturally).

You'll be able to run all your Office apps on the NexPhone when it's acting as your Windows 11 PC and carry out other basic computing tasks. I've seen people worried that the admittedly somewhat old mid-range ARM mobile CPU (it goes back half a decade or so) might struggle running Windows 11, but based on YouTube footage of the device in action, which has been aired this week, it seems like the OS should work fine with the basics.

Of course, you won't be able to do anything remotely demanding – but for emails, browsing, dipping into Office, and so on, it all looks relatively okay in performance terms. There's 12GB of RAM backing up the Qualcomm CPU, incidentally, which should be enough. It's also worth noting that the CPU has been picked for its long-term support, with 10 years of that in the pipeline, a comforting prospect.

Money talks

NexPhone being held in a hand, rear camera showing, outside and it's raining — (Image credit: Nex Computer)

The second thing that excites me about the NexPhone is its potential to save people money. While it's more targeted as a secondary phone, or work smartphone (it's a rugged model, incidentally), I can see a market for this later in 2026 for people who are priced out of buying a PC, perhaps.

We've got the RAM crisis, of course, and SSD prices are also heading skywards, so it's no secret that PCs are going to become a good deal more expensive this year. Indeed, these price hikes are widely predicted to get even worse before the situation improves (or even stabilizes), and the reality is that some folks may need to buy a new computer in 2026, but may not be able to afford one. Even though their current PC may be in dire need of replacement and on its very last legs.

If they're also mulling a new (affordable) phone in the near future, the NexPhone could be a tempting middle ground that gives them entry to both worlds. The $549 price will effectively secure two devices: a new mid-range smartphone and a Windows 11 (or Linux) PC. That could seem quite the bargain in a time of spiralling prices (RAM has got truly ridiculous now).

For this niche set of buyers, the NexPhone could be a way to sidestep this whole PC pricing nightmare. It's your phone, it's your PC, and if you get (or already have) a portable display (and Bluetooth keyboard), it could be your laptop, too.

Just scratching the Surface

The final reason the NexPhone is a truly intriguing development is not the device itself, but what it points to in the future. Now this hardware exists – and if the reality of how useful it seems actually pans out, it'll surely pave the way for others to make similar devices.

Even if performance does feel a bit wonky with the Windows 11 side of the equation on the NexPhone, which is one of the main points of skepticism – that and 'who wants Microslop Copilot on a phone' (just use the Linux desktop, then) – a beefier version of such a phone can remedy that.

Further to that, here's another thought: maybe this could prompt Microsoft to consider creating its own higher-end version of this concept? I may be wrong, but I've got a feeling the NexPhone could be a very important step forward in terms of bringing together the worlds of mobile and desktop computing, which, one way or another, is going to be the future.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/computing/windows/the-nexphone-could-already-be-the-most-important-hardware-launch-of-2026-heres-why zwAtXARybKEq4iiMzdwCEQ Sat, 24 Jan 2026 17:00:00 +0000 Fri, 23 Jan 2026 18:52:35 +0000 <![CDATA[ I tested dozens of home gadgets last year – these are my 3 absolute favorites ]]> As one of TechRadar's Homes Editors, I test dozens of gadgets and appliances each year. And 2025 was an especially strong year. So many products really impressed me that I wanted to showcase them in their own article.

What's that you say? We're way too far into January to still be looking back on last year? You might have a point, but I'm doing it anyway. And the good news is that because these picks aren't brand-spanking new any more, I'm starting to see price drops on many of them.

There were too many great products to list here. Particular shout-outs should go to the Dyson Airwrap i.d. (which is clever enough to turn even me into a capable hair stylist) and the Hoover HW3 wet floor cleaner (which proves products don't need to be ultra-flashy or expensive to make your life tons easier). But I've brutally narrowed down all the products I tried out last year to just my absolute favorites. Read on for the top 3 gadgets that truly blew me away...

#1. Dyson V8 Cyclone

Dyson V8 Cyclone cordless stick vacuum — (Image credit: Future)

What is it?

An updated version of Dyson's iconic V8 cordless vacuum, which came out almost a decade ago now but remains a bestseller. Dyson has kept the essence of that design, but added some key upgrades (longer runtimes, a removable battery, stronger suction) to turn it into a cordless vac that's fit for modern times.

Why do I love it?

It's not the fanciest vacuum around, but the V8 Cyclone is one mean cleaner. In my group test to find the best cordless vacuums around, this model sailed through my cleaning trials, performing way better than I'd expect for the price. Speaking of which – I love that Dyson has kept the cost relatively affordable here. It's only slightly pricier than the original V8, even though it's a way better vacuum all round.

Read my full Dyson V8 Cyclone review

#2. Roborock Saros 10

Roborock Saros 10 robot vacuum in its dock, in reviewer's home — (Image credit: Future)

What is it?

One of Roborock's 2025 flagship robot vacuums. The headline feature is a streamlined design that can get beneath low furniture, partly thanks to a LiDAR puck that can descend when the robot senses it's entering a low-height space.

Why do I love it?

Firstly, it can clean the space under my sofa, which otherwise would be getting a twice-yearly vacuum at most. Otherwise, it's just a super-capable robot vacuum. Even though dozens of robovacs make their way through my doors each year, this is the one that's earned a permanent spot in my home – it vacuums and mops extremely well, it's reliable and very user-friendly. Oh, and I love the swish, glossy dock, despite it being a fingerprint magnet. Because there's a newer version out now (the Roborock Saros 20 Sonic), I've spotted some big discounts on this older model, too.

Read my full Roborock Saros 10 review

#3. Segway Navimow X3 Series

Segway Navimow X350 on reviewer's lawn — (Image credit: Future)

What is it?

A robot lawn mower from a sister brand of the guys that make the self-balancing, two-wheeled transporters so beloved by tourists and hen dos. This model is designed for massive lawns. It's built to mow quickly, efficiently and precisely, leaving large yards looking smart with minimal effort.

Why do I love it?

After testing a few robot mowers that caused me big headaches, this Segway bot was a breath of fresh air. It was easy to set up, the app was nicely designed and actually made sense, and it did the job it was designed to do. It mowed my sister's massive garden quickly and with no fuss, taking tree coverage and awkward perimeters in its stride. The brand has announced its 2026 lineup now and it packs some cool-looking upgrades, but the X3 is still an outstanding product.

Read my full Segway Navimow X3 Series review

Find out where to buy a Segway Navimow X3 Series

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/home/i-tested-dozens-of-home-gadget-last-year-these-are-my-3-absolute-favorites LFezMXWghktz5RFj3yGCgm Sat, 24 Jan 2026 15:00:00 +0000 Fri, 23 Jan 2026 16:07:35 +0000 <![CDATA[ The Telco AI factory: co-shaping the future with intelligent communications platforms ]]> The telecommunications industry is at the forefront of a transformative era, moving beyond its traditional role as a mere connectivity provider. This profound evolution is driven by the Telco AI Factory, a concept being actively pioneered by hardware platform providers.

This operational model fundamentally reshapes a telco's infrastructure into an intelligent, hyper-automated, and data-driven system. It leverages AI across every layer of the network: operations, product development, and customer experience, while also being a key revenue engine.

This visionary shift is a collaborative endeavor currently involving two core pillars: Hardware Platform Providers which supply the foundational computing and Telcos, which contribute their vast infrastructure and trusted customer relationships.

The third emerging pillar which is crucial for co-shaping this model are Cloud Communications Providers. They provide the contextual real-time, voice data enabling telcos to combine with network insights and develop AI-centric business cloud solutions.

Voice AI, which encompasses and extends well beyond Generative AI, will fuel a massive new market for telcos. According to Goldman Sachs, generative AI alone has the potential to add a staggering $7 trillion to the global GDP over the next 10 years.

Cloud communication innovators are key to unlocking this immense potential in Voice AI-driven services, transforming the way enterprises communicate.

Why Telcos are uniquely positioned to build AI Telco factories

Historically, Central Offices (COs) were the switching hubs forming the bedrock of local connectivity. Today, COs are rapidly evolving into AI hubs that will transform telcos into providers of real-time Agentic services.

While hyperscalers have traditionally dominated by hosting many telco workloads, telecom providers possess inherent, undeniable advantages that uniquely qualify them to lead this transformation:

Sovereignty: Telcos are trusted by nations, organizations, and government entities. This enables them to offer localized AI solutions tailored to meet national requirements for sovereign AI.

Extensive, local Infrastructure: They own vast, interconnected networks, including regional and central data centers. The COs' close proximity to users provides the low-latency edge crucial for AI workloads.

Access to Vast, Contextual Data: Telcos have access to immense volumes of network KPIs and conversational metadata allowing for deep, trusted context and cross-optimization across both private and public 5G/6G networks.

This control over infrastructure and data is critical, as AI inference often cannot be centralized.

The role played by cloud communications providers

While hardware vendors provide the essential compute infrastructure and telcos offer the physical network and localised trust, it is these specialised Cloud Communications Providers that inject the "intelligence" into the Telco AI Factory's offerings. They are critical partners in enabling telcos to move up the value chain via several key dimensions:

Conversational intelligence: AI engines analyze all communication data (voice, video, text) to perceive, reason, and act, transforming raw data into actionable insights. Sophisticated, domain-specific AI Agents provide human-like, empathetic service, adapting to nuances and user profiles across all interactions.

Converged network & application APIs: Network exposure functions with real-time network context, such as device location, SIM swap status, and QoS APIs can be integrated with AI agents to support advanced capabilities like secure authentication, personalized services, fraud detection, and QoS, adhering to global standard APIs like CAMARA and robust governance frameworks.

These integrated capabilities allow telcos to quickly deploy new AI-centric services, creating a seamless, personalised, and proactive end-to-end business engine across the entire communications estate (internal collaboration to external customer service).

Ultimately, telcos can significantly uplevel their business offerings, gain incremental revenue, and increase customer satisfaction and retention.

Enabling industry-specific AI applications: The healthcare example

The Telco AI Factory offers a compelling opportunity for telcos to expand into AI-driven vertical markets, supporting local economies through sovereign AI solutions. Healthcare, a sector plagued by escalating costs and administrative burdens, is a prime example.

The U.S. healthcare system alone experiences an estimated $760 billion to $935 billion in wasteful spending annually, with a significant portion (~30%) attributed to administrative complexity.

Telcos can deploy specialized, empathetic AI agents built on local, trusted infrastructure with the Telco AI Factory.

These agents offer 24x7, omni-channel access in any language. Critically, they connect directly to existing healthcare data foundations (Cerner, Epic, FHIR APIs) and use trusted, specialized LLMs at the edge, enhancing communications with network intelligence, ensuring patient data adheres to local regulations and data sovereignty mandates.

This trusted, localized approach positions telcos as indispensable partners in transforming patient care and reducing operational waste.

The strategic imperative for telecom providers

It’s clear that the transformation into the Telco AI Factory is not merely an option but a strategic imperative for telcos in a new era where pervasive AI integration will define enterprise competitiveness.

By embracing this model and innovating with Cloud Communication Providers, telcos can fundamentally enhance their operational capabilities, secure their position as key players in the AI economy, and drive measurable results for their business customers.

It's an opportunity to leverage their unique assets – trust, infrastructure, and data – to become the essential architects of intelligent, context-aware communication that powers the next wave of global economic growth.

We've featured the best small business server.

]]> https://www.techradar.com/pro/the-telco-ai-factory-co-shaping-the-future-with-intelligent-communications-platforms iG8Xz3LBP3mxerwtScEBrD Sat, 24 Jan 2026 15:00:00 +0000 Tue, 20 Jan 2026 15:10:07 +0000 <![CDATA[ The new Fable trailer almost brought a tear to my eye — could it finally be the game that was promised two decades ago? ]]> Nostalgia can be a potent feeling, but by Albion did Playground Games’ Fable, aka Fable 4, trigger a dose of it during the first Xbox Developer Direct showcase of 2026. As someone who can’t remember the last time he cried, the sheer potency of nostalgia for a game I adored in my younger years almost brought a tear to my eye.

One of my first and most enduring Xbox memories was waiting for the original Fable; with the promise of a game with a world you could shape through your actions and decisions.

While the options to influence story outcomes via decisions and morality systems was nothing new, as there were plenty of CRPGs with similar systems – Baldur’s Gate 2 stands out to me as one notable example – the idea of a somewhat open world with such decision making in, for the time, attractive 3D graphics, was tantalising. I remember Lionhead boss Peter Molyneux touting how you could plant a tree in the game and, as the years passed, watch it grow.

However, when it was released in 2004, Fable lacked a lot of the dynamic features promised. Still it was far from a bad game: it had a whiff of 3D Zelda with open areas to explore and secrets to uncover, yet also mixed in some systemic elements – for example, people would react to what you did, and if your character ate too much he’d get fat.

I distinctly remember exploiting it all in fun and slightly dark ways. For example, playing as a renowned hero you could flirt with all manner of people in various villages and towns, to the extent you could ask them to follow you. I used this to lure people outside of their respective homes and villages and then murder them without alerting the guards. This would then mean their property would be put up for sale, which I’d then buy and rent out; from this I built up a real estate empire across a slice of Albion.

Other times, simply for fun and chaos, I’d buy a load of beers from a local tavern and use an emote system to give them to other punters, often ending up turning most of the village of Oakwood into staggering drunks.

All this was wrapped up in a fairy tale setting and story, rather than more traditional fantasy fare. With a big helping of very British humor on top, it made for a fresh and wonderful game back two decades ago, and I think it’s still worth playing today.

In the decades since Fable, we’ve seen some incredible games with fantastic systems and far more expansive worlds. With Breath of the Wild, Skyrim, and Red Dead Redemption 2 being just a few I could name, it perhaps leaves Fable and its sequels feeling a more basic and constrained when viewed through a modern lens.

Grounds for a reboot

a screenshot from Fable — (Image credit: Microsoft / PlayGround Games)

But this is where Playground Games comes in. Rather than go for a sequel, this Fable is effectively a reboot, using different characters and a new story centred around being a hero.

Seeing a Fable rendered in what looks to be cutting-edge graphics and set in a proper open world – something Playground Games is adept at giving us in its Forza Horizon games – is amazing. Spotting hobbes and balverines, and heroes kicking chickens in the deep-dive video, takes me back to a time when life was a lot simpler.

However, the really exciting thing is it looks like Playground Games has the technology to actually bring to fruition the dynamism Molyneux touted for the original game.

Not only does it look like the ability to mess with villagers and townspeople is back, including a romance system and deeper insight into how they perceive you, there’s what Playground Games is calling the “Living Population” where there’s more than 1,000 unique characters with their own routines and personality; in the original Fable, such people had day-night routines, but it was somewhat rudimentary, as was the romance system.

It’s also looking like the morality system will be a lot less binary and more subjective, which should interact interestingly with the living population.

All this is very intriguing and exciting to me; while we’ve seen plenty of gorgeous open-world games – recently the likes of Starfield and Avowed – they arguably lack the depth needed to make interacting with the world engaging.

I also like the look of the combat system. It seems like it will have the straightforward approachability of the other Fable games but likely more depth for advanced players; I’m not expecting a Dark Souls or Elden Rin level of skill required, but more nuance and tools at a hero’s disposal is certainly attractive.

It’s probably safe to say the past 12 months hasn’t been a slam dunk for Xbox, with layoffs and game cancellations. Keeper gave a solid dose of creativity out of the Microsoft-owned Double Fine, but beyond that the brand hasn’t exactly galvanised the industry. But if Playground Games can nail this Fable reboot, I think it could be a fresh trailblazer for Xbox Game Studios. And the fact it's coming out on PS5 too, could bring Fable to a whole new audience.

So I’m crossing my fingers and toes that Fable finally lives up to its promise come this Fall, as it’ll be nice to have something else other than Grand Theft Auto 6 to get me through the darker, colder months.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/gaming/the-new-fable-trailer-almost-brought-a-tear-to-my-eye-could-it-finally-be-the-game-that-was-promised-two-decades-ago SdUheCrfUC4mEJZxfDJ8Jf Sat, 24 Jan 2026 10:30:00 +0000 Fri, 23 Jan 2026 18:21:44 +0000 <![CDATA[ A practical blueprint for scaling AI in financial services ]]> Generative AI, and increasingly AI agents, are quickly taking center stage in financial services.

What was formerly limited to experimentation has since evolved into systems capable of data analysis, real life action and large-scale decision making.

Many businesses are already feeling the effects of this transformation; according to KPMG research, more than half (51%) of the financial sector say AI is reshaping their business. On the other hand, almost three-quarters (72%) are concerned about data quality.

This is when strategic risk creeps in, stemming from fragmented or poorly governed data, that ultimately delays the transition from pilot to production.

Financial institutions must shift their focus if they want to see true value from AI. With solid data foundations that are backed by robust infrastructure and unified governance they will then be in a better position to implement AI safely and successfully.

The true challenge now is not what AI can accomplish, but rather how businesses can take the right steps to enable it to operate at enterprise level.

Building the foundations for enterprise-scale AI

Most AI pilots fail both because the data beneath them are fragmented, poor quality or locked away in silos and because their AI agents do not have a focus on measuring and improving quality and accuracy. In order to successfully deploy AI, the infrastructure must be set up correctly to harness results.

For leaders in the financial sector to close the AI adoption gap, a structured roadmap should be in place to enable their business to move from experimentation to scaled impact.

The first step is to unify data silos under a single platform to eliminate duplications, reduce inefficiencies and build reliable, trusted models from a single source of truth.

From there, governance must be embedded to manage lineage, access and audit trails. For AI agents, governance is far more than a mere compliance exercise. A unified governance model treats agents with the same rigor as human staff, applying robust access controls and security measures.

Prioritizing explainability is equally crucial. In a highly regulated market, businesses need accessible, transparent models that demonstrate how results are produced.

Additionally, adopting a "start small, scale fast" strategy demonstrates impact early, fosters internal trust and establishes a replicable model for safely and responsibly expanding AI across the company.

Closing the gap between AI vision and execution

Leaders in the financial industry are no longer asking where AI works, but instead, where it can deliver the most impact. The potential is enormous, but the gap between ambition and execution is slow to close.

At present, hype is outpacing reality. A recent Gartner survey shows that finance AI adoption jumped from 37% in 2023 to 58% last year, however momentum is now slowing, showcasing the gap between experimentation and enterprise scale.

Despite varying regulatory environments, firms across banking, payments, capital markets and asset management align on the same strategic objectives driving AI adoption.

Businesses must acknowledge that to deliver these ambitions consistently at scale, the challenge is not in the vision, but in bringing together fragmented data assets and legacy infrastructure.

How AI delivers sustainable growth

The financial industry recognizes the value that AI technology can offer by boosting efficiency and driving growth, we can see that in the uptick of adoption of the technology.

Smarter customer segmentation and hyper-personalization allow enterprises to differentiate their brand and elevate customer experience, creating a significant advantage over their competitors.

In payments and mortgages for example, AI-powered product innovations such as real-time fraud prevention and property valuation models are transforming journeys and reshaping how institutions deliver their services.

However, implementing individual use cases is not enough to translate these skills into long-term revenue development; a clear business strategy is also necessary.

Financial organizations must prioritize use cases with quantifiable ROI, align AI operations to particular business goals, and make sure that data foundations enable models to be constantly refined.

How AI agents are redefining risk management

In financial services, risks can appear in minutes, from cyber threats to fraud disruption. The speed, complexity and sheer volume of these problems are too much for traditional manual methods to handle.

AI agents are quickly becoming the new competitive frontier to improve quality and accuracy. Unlike static models, these systems can act almost like virtual employees that take actions autonomously.

In mission-critical areas such as fraud detection, anti-money laundering (AML) and cybersecurity, agents monitor, orchestrate and conduct checks with far greater speed and reliability than manual teams.

Operating in one of the strictest regulatory industries, AI agents provide a means for organizations to keep ahead of the risks while preserving the integrity of key operations. Rather than replacing human judgement, AI agents enhance it; enabling teams to react with greater assurance.

Reimagining operations with AI

Advanced AI tools are changing the game for financial services, driving innovation and agility. AI agents can automate repetitive business processes allowing institutions to “do more with less”, reducing workloads which allow teams to focus on higher-value, customer orientated work.

AI-driven customer service assistants are also already delivering measurable impact. Trained on enterprises’ own data, they can answer questions accurately and automate much of the triage process. The results are fewer manual bottlenecks, elevated customer experiences and a more resilient operational model.

Building the future of financial services

The majority of financial organizations understand how and why AI will affect them in the future. However, trying to operationalize it in a way that is dependable, compliant, and long-lasting is less evident. Data architecture and governance must be seen as fundamental in a company’s data and AI strategy to achieve a competitive edge.

For AI agents to operate safely within rigorous boundaries, systems must be built with established controls and continuous monitoring. AI success will favor the institutions that adopt a disciplined approach and scale with confidence.

We've featured the best small business software.

]]> https://www.techradar.com/pro/a-practical-blueprint-for-scaling-ai-in-financial-services kzJM5ks28HeEpvaWp54vab Sat, 24 Jan 2026 09:00:00 +0000 Mon, 19 Jan 2026 15:08:42 +0000 <![CDATA[ It might sound strange, but The Matrix is actually a happy movie — rewatch it to shake off the January blues ]]> One of the things I enjoy about being a journalist is that stories, ideas, and thoughts can come from all sorts of places: case in point, when on a date and discussing the topic of movies, it was presented to me that The Matrix is a ‘happy’ movie.

I cracked a smile at that; for those who don’t know thrust of The Matrix is that robots rose up against humans and enslaved our consciousness in a virtual world while they use us as biological batteries, with the few people who escaped from the system now entering the titular Matrix to free chosen humans from this virtual prison. From that premise, and the dark, cyberpunk tone of the movie, I’d never have thought to describe it as happy.

But then, as the discussion went on, I was brought round to my date’s thinking.

Ultimately, The Matrix is a happy movie as it’s got a lot of the archetypes of a viewing experience that one leaves happy and uplifted; the good guys win over the bad, Neo is ‘The One’ and a promised saviour, no one major dies, and the scenery-chewing Mr Smith gets ‘killed’ in a most satisfying way.

Of course, there's more to this. Unlike the successive Matrix movies that tried to hammer home the kind of philosophical chatter one might expect from mildly inebriated university students, the first movie was taught, superbly shot, and stylish, albeit in a very late-1990s way.

The pacing was on point, starting with a young Keanu Reeves as Mr. Anderson and his journey to become Neo. At the time it was dark, gritty, and had a few challenges for those of a squeamish nature – the mechanical bug that enters Neo’s body through his belly button, and the waking of real-world Neo in a gross, almost Giger-esque bio-battery pod – yet this was then built up into a feel good moment, where Neo “knows Kung-Fu” and does a good job at fending off a tougher, more experienced Morpheus.

Throughout The Matrix, there are numerous ‘feel good’ and happy moments all set at a satisfying pace. Who can forget the classic “guns, lots of guns” moment, the utterly brilliant and innovative bullet-time slow-mo corridor shootout, Neo’s ability to dodge bullets, and the rescue of Morpheus? And of course, the ending, which is basically a cluster of heroic, air-punching scenes.

I could argue that, in the face of more recent sci-fi/cyberpunk storytelling, The Matrix is a bit old-hat and almost one-note. But then that’s what makes it a happy movie: you don’t have to overly work your grey matter to have an enjoyable time. And its action, camera work, and direction all stand the test of time; I reckon the color palette helps nullify any CGI that could look a little ropey in 2026.

So with all that in mind, and given January is considered to be a rather depressing month in the Northern Hemisphere, I’d suggest giving The Matrix a watch this weekend if you want a bit of a pick-me-up.

It should be available to stream in the US on HBO Max and Apple TV+, as well as on other on-demand services, and on Now TV in the UK if you opt for the Entertainment tier.

Equally, for what it’s worth, may I suggest looking at grabbing The Matrix on 4K Blu-ray. I think it's one of those movies that will always be evergreen and worth an addition to one’s physical media collection; in 4K, the action should look rather crisp, too.

Check out the widget below for any 4K Blu-ray deals on The Matrix in your area.

And then head to the comments to let me know if you agree that The Matrix is a happy movie, and tell me what movies you'd recommend as left-field happy, feel-good flicks.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/streaming/it-might-sound-strange-but-the-matrix-is-actually-a-happy-movie-rewatch-it-to-shake-off-the-january-blues 4cWTCDRJmCdFjPLy5mkSZe Sat, 24 Jan 2026 02:00:00 +0000 Fri, 23 Jan 2026 20:13:54 +0000 <![CDATA[ Cracking the AI code: realizing AI's true value in finance ]]> Finance teams have largely moved beyond the experimentation phase of AI adoption and are now focused on delivering tangible results.

In fact, our recent research shows 99% of UK finance leaders view AI as essential, with 85% of UK finance teams already integrating it. Investment in the technology is also set to continue rising through 2026 as organizations double down on AI capabilities.

While AI is delivering automation, efficiency, mitigating risk, and providing deeper insights from vast datasets, the value realized across the finance function can vary depending on how initiatives are deployed.

Recent research suggests that many AI projects fall short, not because of the technology itself, but due to challenges aligning implementation with regulatory complexity and fast-changing economic conditions.

The adoption AI tools accelerates, expectations are rising in parallel. This has prompted finance leaders to take a more considered approach to AI risk, with 78% acknowledging they have some concern about the risks associated with AI.

Rather than slowing progress, this growing focus on risk and trust reflects a maturing market - one that recognizes sustainable value comes from deploying AI with the right controls, governance and oversight in place.

As AI reshapes customer expectations, finding the right balance is even more crucial.

Finance teams are increasingly under pressure to deliver faster, more personalized and higher-quality service. AI systems only built on trust, transparency, and human oversight can scale these improvements without increasing risk – whatever challenge lays ahead.

The role of AI is evolving in finance

AI’s role in finance is evolving from completing simple, repetitive tasks into a strategic partner which guides decision-making using data-driven insights and analytics. Only when embedded into core processes, AI enables finance leaders to respond to real-time changes and market shifts with greater agility.

It can also reassure finance leaders that long-term planning and resilience can be built through scenario modelling, enhanced forecasting, supply chain analytics, and optimized investments.

Trust and visibility unlock AI’s value in finance

Gaining value from AI is a balancing act. Finance teams that onboard AI too quickly risk introducing new vulnerabilities, while those that delay adoption altogether may find hesitation becomes a liability.

According to OnPhase, manual processes, such as lagging approvals, unpredictable payments, and time-consuming repetitive tasks continue to consume multiple hours each day. However, AI is perfectly positioned to eliminate such inefficient ways of working.

At the same time, inadequately regulated AI introduces risks. A lack of oversight can leave systems exposed to malware and data integrity issues, undermining trust and creating challenges for auditors.

This concern is highlighted in recent research, finding that over three quarters (78%) of UK finance leaders express some level of concern about AI risk, including 11% who are extremely concerned, 30% who are somewhat concerned, and 37% who are a little concerned.

The key to unlocking AI’s true value lies in visibility and control in finding the right balance between autonomy and oversight. Teams can thrive by building trust into every AI-related decision without compromising on the essential human intervention.

While AI is widely used across UK finance teams for forecasting (60%), financial analysis (58%), reporting and insights (59%), and fraud detection (57%), over half of respondents say they need to review AI actions, and 48% say retaining decision-making control is essential.

Real value requires resilience through global politics

Once trust and visibility are established, AI must also be resilient to global political and regulatory change.

Organizations need confidence that AI systems will adapt to regulatory shifts and geopolitical tensions, without compromising speed, accuracy, and control. For finance teams to rely on AI long- term and for companies to integrate it more widely, it will need to guarantee tangible, measurable value.

This is particularly important as AI maturity varies widely across organizations. When it comes to AI agents, many businesses are taking a deliberate phased approach - initially deploying them within one or two functions to build confidence.

At present, with 10% report scaling AI agents organization wide, highlighting that most organizations are still in an evaluation and optimization phase.

Distinguishing between different AI use cases is therefore critical. Each application brings its own considerations around value, risk and control, and organizations that take time to define these parameters early are better positioned to scale responsibly and unlock measurable, long-term impact.

Global policy shifts, such as tariffs and other policy changes, directly affect profit margins, cost structures, operational adjustments, and liquidity pressures. AI systems training on static assumptions may require retraining when conditions change, especially with many models having a knowledge cutoff point.

Policy changes, like Tariffs, can also introduce new financial reporting risks, increasing scrutiny around revenue recognition.

In high-pressure environments molded by political and economic uncertainty, AI’s effectiveness hinges on how well it has been implemented and how efficiently it has been onboarded.

When finance teams are prepared to adapt processes in real time, AI can support more informed decision making; without this readiness, it risks becoming a constraint rather than an advantage.

What’s expected for AI to deliver on results

Both generative AI and agentic AI have the potential to boost efficiency and improve insights for finance functions, but as AI is more widely integrated, humans need to trust, oversee, and guide its outputs to ensure it delivers real value.

Teams must balance AI autonomy with human intervention, and ensure systems remain resilient amid regulatory shifts and geopolitical tensions.

Ultimately, AI will only display its full potential when finance teams feel they can confidently rely on it to support faster decisions, drive insights, and build long-term resilience, while retaining the ability to intervene when necessary.

We've featured the best business plan software.

]]> https://www.techradar.com/pro/cracking-the-ai-code-realizing-ais-true-value-in-finance 9LZpk6rNfmrwNVHkRaxARP Fri, 23 Jan 2026 15:17:48 +0000 Fri, 23 Jan 2026 15:17:49 +0000 <![CDATA[ The human paradox at the center of modern cyber resilience ]]> Cybersecurity headlines continue to be punctuated by high-profile, highly disruptive breaches. In fact, 2025 has already seen some of the most damaging incidents on record, with organizations such as M&S and Co-Op suffering severe and costly disruption.

If the estimated £100 million-plus cost to M&S wasn’t bad enough, it was eclipsed by the incident at JLR. widely reported as the most expensive cyber incident in UK history, its overall economic estimated at £1.9 billion.

These and other incidents expose cybersecurity’s most enduring paradoxes, humans are both the strongest and weakest links in the chain. M&S has publicly stated that its breach was a result of “human error”. While the jury is still out at JLR, speculation points towards a credentials compromise, a classic social engineering tactic.

At the same time, employees are also the bedrock of effective cyber resilience, with an engaged, well-informed workforce ideally placed to recognize and stop suspicious activity long before it develops into a full-scale incident.

A bad problem is becoming worse

The problem for security leaders is that social engineering is still the most effective way to bypass otherwise robust technical controls. The problem is becoming more acute as threat actors increasingly use AI to deliver compelling, personalized, and scalable phishing attacks.

While many such incidents never reach public attention, an attempt last year to defraud WPP used AI-generated video and voice cloning to impersonate senior executives in a highly convincing deepfake meeting.

Unfortunately, the risks don’t end there. Even with strong technical controls and a workforce alert to social engineering tactics, risk also comes from employees who introduce tools, devices or processes that fall outside formal IT governance.

Widely known as ‘Shadow IT’, this behavior often stems from good intentions, with people trying to work faster or collaborate more effectively when the official tools don’t meet their needs as well as alternatives they have discovered. The result is security blind spots and data governance risks that continue to cause serious difficulties.

The list of challenges goes on, with everyday habits such as credential reuse, storing files in unapproved locations, sharing data via consumer apps or using unsecured public Wi-Fi all having the potential to undermine organizational security.

The need for a cultural shift

So, where does that leave us? Reducing human risk is not just about eliminating mistakes, it’s about creating an environment where secure behavior becomes the default. It’s clear, for instance, that tick box training where people passively click through generic modules just isn’t fit for purpose.

What’s needed instead is a shift in both mindset and culture, where employees understand not just what not to do, but why their day-to-day decisions, which tools they trust, how they handle unexpected requests, and when they choose to slow down and double check something rather than act on instinct genuinely matter.

From a leadership perspective, it’s much better to foster a culture which people feel comfortable reporting suspicious activity without fear of blame, rather than an environment where taking the risk feels like the easier option.

Driven by the frustration that awareness training just doesn’t cut it, many organizations have implemented layers of security tooling to close human centric security gaps. While no one would dispute that tools are essential, they can only go so far, and without a strong culture, their impact is limited.

Strong culture

But what does this ‘strong culture’ look like? Consider this scenario: An employee receives an unexpected email request that looks routine but feels slightly “off”. Perhaps, its a supplier asking for account details, or a colleague requesting access to documents they don’t usually need.

Instead of acting quickly to avoid delaying work, the employee pauses because the culture has normalized slowing down when something seems unusual.

They also know exactly how to report or verify because the processes are familiar and straightforward, with no confusion about who to contact or whether they’ll be blamed for raising a false alarm.

A quick check with the security team confirms the request isn’t legitimate. The team treats the report as valuable intelligence rather than an inconvenience.

Leadership acknowledges the employee’s actions, reinforcing that thoughtful decisions and early reporting are recognized and valued. The incident becomes a learning example for the wider organization, shared as best practice to help others recognize similar patterns in future.

The underlying point is that this needs to happen much more often. Get the approach right, and organizations of every size stand a far better chance of staying on course when faced with cyberattacks with the potential to escalate into existential threats.

We've featured the best business VPN.

]]> https://www.techradar.com/pro/the-human-paradox-at-the-center-of-modern-cyber-resilience KGCviFNB34Dps3hhKfYw7m Fri, 23 Jan 2026 14:56:52 +0000 Fri, 23 Jan 2026 14:56:52 +0000 <![CDATA[ Dyson's biggest product innovations, ranked from game-changer to gimmick ]]> Dyson is one of the most respected appliance brands around, and has a reputation for invention. In my role as a TechRadar Homes Editor, I've had a chance to try out almost all of this brand's gadgets and appliances (as well as plenty of competitor products, for added context). I'm consistently impressed with Dyson's creativity and innovation... but not all of its inventions are quite as good as others.

Below, I've listed Dyson's big innovations, ranked from best to worst. To narrow things down a bit, I've focused on the brand's vacuum cleaners and haircare gadgets, though, of course, it also has significant stakes in the air purifier and fan markets. Read on for the Dyson features that truly blew me away, and those that turned out to be downgrades in disguise.

#1. A teeny tiny motor

Any handheld gadget typically benefits from being as compact and light as possible. For its Supersonic r hair dryer, Dyson shrunk the motor right down and put it in an incredibly compact shell. It looks weird, but it's a dream to use, with the pipe-like shape directing air with precision – plus, it blasts out air with serious force for such a small gadget.

Dyson PencilVac in tester's home — (Image credit: Future)

Perhaps even more impressive is the new Dyson PencilVac. This is a cordless stick vacuum, except all the internal mechanics have been shrunk down and tucked into the handle, which is 1.5 inches / under 4cm in diameter.

It looks like a broom handle, except inside there's a tiny little motor, alongside filters, a battery, and even a dust cup. It almost defies belief. I've been busy testing this gadget at home, and it's ultra-nimble and light.

#2. Lasers on vacuums

Want to be disgusted by exactly how dusty your floors are? Now you can! Dyson was the brand that first put dirt-illuminating laser headlamps on its vacuum floorheads. They're ultra-bright and positioned at precisely the right angle to cast big shadows on the tiniest particles and finest hairs on your hard flooring.

They're genuinely useful in making sure you get every last speck of dust cleared away. They're also just generally really satisfying to use and helpful in highlighting (literally) the dirt-hotspots that are getting neglected.

Laser illumination on Dyson Gen5detect vacuum cleaner — (Image credit: Future)

These were first introduced on the Dyson V15 Detect, but appear on most of Dyson's newest vacuums, including the Dyson Gen5detect and V16 Piston Animal. Lots of brands have copycat versions these days. I've tested plenty of headlamp-equipped vacuums from different brands, but very few come close to Dyson's version in terms of effectiveness.

#3. Curl automation

Dyson deserves credit for pioneering a new way of curling hair with its OG Airwrap. Now my bathroom is full of Airwrap dupes, and the auto-wrapping, air-powered curling feels like old news, but Dyson has managed to capture my attention once again with a major upgrade to its iconic multi-styler.

The newest models – the Airwrap i.d. and Airwrap Co-anda 2x – have a little spring-loaded switch on the front that automates the curl sequence for you. Flick it upwards, and it kicks off a wrap – curl – cool shot process, so all you need to do is hold the styler in place while it works its magic.

The timings of each part are specific to your hair and determined by a quiz you complete in the companion app. Normally, I'm pretty opposed to apps for things that shouldn't really need one, but here, Dyson gets a pass.

Airwrap i.d. with auto-wrap barrel attached — (Image credit: Future)

Now, this upgrade isn't essential. I could achieve the same results on my own with some focus and clock-watching. It's just that it makes styling so effortless. And because I invariably get bored or distracted when trying to curl without the automated sequence, the results on me always look more polished and last longer when the machine times it all out for me.

#4. Ultra-precise dust detection

Dyson's newer flagship vacuums (from the V15 onwards) have the ability to detect exactly the size and number of particles they're sucking up at any time, and report back on a screen mounted above the handle. They do this in real-time, via a colorful, moving bar graph, and a summary is provided at the end of the cleaning session. The vacuums also use the information to adjust suction to clean away the particles efficiently.

Let's be clear: I do really like this feature. The reason it's further down my list is that I'm not convinced all of it is strictly necessary. The particle detection and automatic suction adjustment is super-useful. I also enjoy watching the bars shift as I clean, and to some extent, it helps me do a more thorough job, as I can see where the floor is especially dirty.

Close up of info screen on Gen5detect vacuum showing summary of particles sucked up — (Image credit: Future)

However, the real-time reports are limited in their practical usefulness, and there's certainly no need for them to be broken up by particle size. I have a sneaking suspicion the bar graph is mainly there for the gratification of the engineers who really, really wanted us to know what was going on under the hood and how much work it took to make it.

#5. Submarine floorheads

Another innovation I'm on the fence about is Submarine floorheads – mopping heads that can be swapped onto certain vacuums to enable them to wet-clean as well as vacuum. In fact, it might be more accurate to say I'm not sure Dyson has really found its way when it comes to wet cleaners in general.

Dyson WashG1 wet floor cleaner — (Image credit: Future)

I haven't been blown away by any of the mopping gadgets I've tested from Dyson so far. The V15s Detect Submarine suffered from a poorly designed Submarine head that had an unfortunate tendency to leak dirty water when disconnected from the machine for emptying. And while the Dyson WashG1 is a very good cleaner, it doesn't do enough to justify its sky-high price tag. There are similarly priced (and cheaper) models in my best wet-and-dry vacuum roundup that do more, and better.

I have a new Clean+Wash Hygiene on its way for testing, so Dyson may still redeem itself on the mopping front.

#6. Conical floorheads

My final entry in this list is one that I genuinely think is a downgrade. On its latest flagship – the V16 Piston Animal – Dyson decided to shake things up on the floorhead front. Instead of having one parallel brushroll, the new-look head has two cone-shaped rollers connected in the center. The idea is that any long strands of hair will make their way to the narrow end of the cone and away into the bin, rather than staying tangled around the roller.

The knock-on effect is that the front of the floorhead becomes slightly pointed. This is a major pain when it comes to cleaning the edges of rooms, because the head won't sit flush to the wall. Getting the perimeter of a room clean now requires a load of awkward maneuvering, whereas before it was straightforward as anything.

What's more, there's a barrier in the center of the floorhead where the two rollers connect, and below which nothing gets sucked up.

Dyson All Floors Cones Sense floorhead on V16 vacuum — (Image credit: Future)

While the new design does work on an anti-hair tangle front, here's the thing: I never had any real issues with hair wrap on Dyson's previous (parallel) floorhead, either. It might make a noticeable difference if you have Rapunzel-like locks, but for most households, the issues it creates will far outweigh any benefits.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/home/vacuums/dysons-biggest-product-innovations-ranked-from-game-changer-to-gimmick Qp33RRc9HN5gQkfqvJdGUG Fri, 23 Jan 2026 02:00:00 +0000 Thu, 22 Jan 2026 20:29:10 +0000 <![CDATA[ The ROI blueprint: turning AI and automation into business value ]]> AI tools have become essential for business, with most companies moving from testing to action. Yet, despite massive investment, only one in five organizations is seeing the results they hoped for, and analysts predict up to 30% of generative AI projects could be scrapped this year.

This churn isn't necessarily a failure of planning, but a natural symptom of the early AI revolution; as business leaders are still discovering which workflows will stick, making it difficult to measure success by efficiency outputs alone.

As budgets tighten and leadership demands proof of value, the definition of ROI must evolve. However, realizing the full potential of these initiatives requires recognizing that in this experimental phase, the most valuable "return" is the creation of an AI-ready culture.

The key is simple: connect what you measure to the reality of the adoption curve, prioritizing organizational agility over immediate speed and results.

Assessing readiness: Culture over infrastructure

Many companies have learned that jumping into AI without proper prep can be costly. However, "readiness" today requires more than just infrastructure updates; it demands a focus on adoption, knowledge and people.

Research confirms that the "human element" is the biggest differentiator. McKinsey found that organizations who emphasize communication and make it a central part of a rollout are seven times more likely to succeed with automation.

It is not enough for employees to know that AI exists; they must understand how to wield it effectively within their specific roles.

To truly build this readiness, you must move beyond passive surveys and prioritize hands-on workshops. These practical sessions allow teams to engage with the tools directly in a controlled environment.

By replacing theoretical training with active experimentation, you bridge the gap between abstract potential and daily utility. This approach not only reveals genuine skill gaps but also empowers your teams to become the driving force of this cultural shift from day one, creating momentum early.

Selecting the right battles

To navigate this revolution, choose initial projects that target high-friction points like repetitive manual tasks or data silos. Start by reviewing ticketing systems and labor costs to pinpoint where AI offers immediate value.

Crucially, prioritize tools that fit natively into your team's existing workflows. If users must constantly switch contexts or toggle screens, adoption will drop. The most successful implementations meet users where they are, enhancing current processes rather than disrupting them.

By targeting these "native" bottlenecks, you eliminate drudgery and make the cultural leap seamless for your staff, while demonstrating clear early wins.

The pivot: From performance KPIs to adoption KPIs

Once tools are live, resist the urge to immediately track efficiency metrics like speed or error rates. In this experimental phase, you cannot optimize a tool that no one uses. Instead, prioritize adoption KPIs to measure the cultural shift.

Focus on usage rates, engagement, and whether teams are actually changing their daily behaviors. Tracking these metrics allows you to catch resistance early and validate that the culture is adapting.

This flexibility is critical; an MIT study found that companies that regularly update their KPIs to match their stage of maturity are three times more likely to see major financial gains over time.

The data difference: Why context drives success

Even with the right culture, the success rate of any AI initiative ultimately relies on the tool's ability to connect to organizational data. Generic AI tools that operate in a vacuum often suffer from low success rates because they lack context. They can generate text, but they cannot solve specific business problems.

In contrast, tools that "know" the organization—by securely integrating with internal data, ticketing history, and documentation—have a significantly higher success rate. This is where data quality becomes a decisive factor. You must look at your data: how clean, complete, and accessible is it?

Organizations that take a thoughtful approach to this integration ensure that their AI isn't just a novelty, but a context-aware engine that drives real business value across multiple departments.

IT as the architect of culture and integration

This integration process places IT at the very center of the transformation. IT teams are no longer just service providers; they are the architects of the organization's future. They are the only ones positioned to bridge the gap between technical data integration and human cultural adoption.

By taking ownership of these initiatives, IT ensures that AI tools are not just "deployed," but are deeply integrated into the company’s unique data landscape. Simultaneously, by using adoption metrics to guide the workforce, IT shapes how the organization works and learns.

The "win" today is twofold: building the technical pipelines that make AI context-aware, and fostering the human habits that make AI usable. By leading with purpose and focusing on these dual outcomes, IT teams turn strategy into lasting business impact and make AI a true business driver in every operational area.

We've featured the best AI website builder.

]]> https://www.techradar.com/pro/the-roi-blueprint-turning-ai-and-automation-into-business-value JDShHZFHWhtqXf6q5UNLmk Thu, 22 Jan 2026 15:22:43 +0000 Thu, 22 Jan 2026 15:22:42 +0000 <![CDATA[ How CIOs can shift from patch and pray to risk-based software change ]]> Let me be clear from the outset: this is not an argument against upgrading. Progress matters. New capabilities matter. Security improvements matter. What’s dangerous is not upgrading, but upgrading by reflex, on someone else’s timeline, in the hope that motion equals safety.

That distinction matters, because recent events have shown what happens when choice, visibility and judgment quietly disappear.

When Airbus grounded A320-family jets over a flight-control fault linked to solar radiation, it looked like a niche aviation issue. It wasn’t. It was a reminder that even industries built on obsessive safety engineering and change control can be exposed by software risk buried inside critical dependencies.

And if that can happen in aviation, with its redundancy, certification and discipline, it should give every CIO cause for pause.

We’ve seen the same pattern play out elsewhere. The CrowdStrike update that rendered millions of Windows machines unusable. The Cloudflare outage that rippled across large parts of the internet.

In each case, a change or failure introduced far upstream triggered immediate consequences downstream. But the deeper issue wasn’t simply the update. It was the dependency, and the assumptions that came with it.

Hidden dependencies and inherited risk

In the Airbus case, modern fly-by-wire aircraft are inherently dependent on tightly integrated software systems. When those systems encountered a failure mode they were not sufficiently hardened against, there was no simple operational escape hatch.

Aircraft were grounded not because engineers were careless, but because safety demanded certainty, and certainty required change before flight could resume. Cloudflare tells a similar story at internet scale.

Cloudflare runs a strong operation and provides critical services. They are not the villain here. But a vast portion of the internet now depends on a small number of shared infrastructure providers for DNS, routing, security and availability.

Organizations that design their services to rely entirely on those platforms inherit their failure modes. When a core service falters, there is often no graceful degradation, just a very hard stop.

This is the uncomfortable reality of modern IT estates. Dependency is not free. When capability is outsourced, control is diluted. Unless systems are intentionally designed with resilience, diversification and fallback in mind, organizations quietly absorb risks they may not fully understand.

Technical leaders must be explicit about what they depend on, how those dependencies fail, and what happens when they do.

Disruption that crosses the line into safety

This is where disruption stops being an inconvenience and becomes a safety issue. Hospitals, emergency services, utilities and financial markets increasingly operate on layers of software stitched together over decades.

Integrations pile up, accountability fragments, and no one has full visibility of end-to-end impact. In that environment, even “routine” change can carry real risk.

A vendor update or external outage can delay access to patient records, disrupt dispatch systems or interrupt services people rely on in moments that matter. Aviation assumes change is dangerous until proven otherwise. Many digital estates still assume change is safe until something breaks.

Yet the dominant narrative remains unchanged: stay current, move fast, trust the vendor.

That narrative is reinforced by incentives. Vendors are rewarded for upgrade momentum. Support models, revenue and product strategies depend on it. The downstream operational impact, retesting safety cases and retraining staff, sits entirely with the customer.

That’s where greed shows up. Not as malice, but as indifference baked into the business model. The people pushing change do not feel the consequences when it goes wrong.

Choice, not obedience, should drive upgrades

Upgrades are optional. They should happen only when an organization wants the new capability they deliver, not because a vendor declares a version “unsupported.” Software does not carry a best-before date.

There is a dangerous belief embedded in modern security practice that once the latest patch is applied, the risk has been addressed and the job is done. In reality, patching is just one possible response to risk, and sometimes the wrong one.

A patch may reduce exposure, but it can also introduce instability, new vulnerabilities or entirely new failure modes.

The strongest form of defense is not speed, but understanding. Knowing what you are exposed to, how that exposure manifests in your environment, and which controls or mitigations actually reduce risk in practice. That requires evidence, not assumptions. Too often, patching becomes a ritual rather than a decision.

Regulatory pressure can even unintentionally reinforce this behavior. Frameworks designed to improve resilience are sometimes reduced to “apply the latest fix and move on.” Compliance becomes a checkbox.

Assurance becomes performative. A freshly patched system can still be poorly understood, poorly monitored and poorly defended.

Risk should be managed deliberately, using the controls that make sense for your architecture and operating model. Sometimes that includes upgrading. Often it means isolating, compensating, hardening or monitoring instead. Evidence, not fear or vendor pressure, should drive the decision.

CIOs may not be able to change vendor incentives, but they can change their own posture. The real upgrade the industry needs is not a new software version. It is a mindset shift, from compliance to comprehension, from speed to substance, and from “patched” to “resilient.” Because snake oil and critical systems still do not mix.

We've featured the best online cybersecurity course.

]]> https://www.techradar.com/pro/how-cios-can-shift-from-patch-and-pray-to-risk-based-software-change tEkPs6gs5mBGozoXHfHKTA Thu, 22 Jan 2026 15:15:16 +0000 Thu, 22 Jan 2026 15:15:16 +0000 <![CDATA[ It’s not a bubble, we’re surfing the AI wave ]]> Every technological revolution brings both opportunity and uncertainty. The rapid rise of AI is no exception. It’s no wonder, then, that alongside the excitement and experimentation comes a familiar question: are we heading towards an AI bubble - and what would happen if it burst?

Investment in AI has accelerated across predictive, generative, and agentic technologies as organizations race to capture value. Slalom’s research shows that 62% of UK&I executives expect a return on AI investment within just two years, a level of impatience that echoes the early days of the internet boom.

Media narratives mirror this tension, fluctuating between optimism and warnings of a dot-com style correction.

The real test now is whether AI can consistently deliver measurable impact and move from promise to performance.

AI: Is it a bubble?

A “tech bubble” typically refers to over-inflated valuations caused by speculation rather than fundamentals. In this case, among companies developing and delivering AI technology.

Naturally, share prices fluctuate in every sector, but tech stocks often remain attractive thanks to ongoing innovation. When interest spikes, investment follows and anything tied to the latest buzzword risks becoming overvalued.

The smaller players especially are most exposed whether through investor overreach or simply the speed of change of AI evolution outdating products. At the same time, new, large players are emerging that could challenge the future AI landscape.

The largest technology firms continue to lead the charge. Alphabet, Microsoft, Amazon, and Meta are expected to invest nearly $370 billion in AI-related initiatives this year, a level underpinned by stable growth, strong financials, and operational efficiencies.

Others are seeing similar momentum; NVIDIA’s record revenues reflect real demand for compute infrastructure, and Sam Altman recently claimed that OpenAI will reach an annualized run rate revenue of $20 billion in 2025 and has plans to reach hundreds of billions in sales by 2030.

This lays the groundwork for an IPO with the potential value of up to $1 trillion. If Altman’s claim is accurate, this potential price to earnings ratio of 50x, while very high, is not unheard of for high growth technology companies with a strong competitive advantage.

Even policymakers see this as a structural shift rather than speculation. Federal Reserve chair, Jerome Powell, recently described the current surge in AI investment as “the beginning of real, profitable businesses” capable of driving sustained economic growth.

Encouraging as this may be, it only makes the competition for success fiercer. The race to turn AI capability into tangible business impact is accelerating and only those who can demonstrate value quickly will maintain momentum.

AI: Is it a wave?

There's no denying the speed and scale of change. The surge of innovation and investment in AI feels overwhelming and that intensity naturally suggests comparisons to a bubble. But if this momentum isn’t building toward a burst, is it instead a building into a wave?

With this much hype, it’s essential to separate substance from noise. Some AI development is being driven by fear of missing out or board-level mandates rather than a clear understanding of value.

Building for the sake of jumping on a trend risks wasted investments and reinforces the perception of a bubble, even when underlying progress remains solid.

Across the industry, however, we’re seeing AI fundamentally reshape how software is delivered, modernized, and scaled. The tools and practices emerging today are shortening development cycles, accelerating legacy modernization, and enabling new forms of automation that simply weren’t possible two years ago. In our work with customers, these capabilities are already unlocking productivity gains and helping organizations deliver value faster and more reliably.

Experimentation remains essential but it must be grounded in a realistic view of business outcomes. Some initiatives will fail, as they should in a healthy innovation cycle. Others will redefine how companies operate and compete.

Slalom data shows 64% of UK&I organizations are creating or planning new AI-related roles, outpacing expectations of workforce reduction. This reinforces what economists call Luddite Fallacy – that technological change tends not to eliminate work long term, but to reshape it and create new forms of opportunity.

The pace of change shows no sign of slowing. Most individuals and organizations have only begun to explore AI’s full potential. Its ubiquity, spanning every role, industry and daily workflow, ensures adoption will deepen rather than decline.

As valuations stabilize and early-stage hype settles, today’s “bubble” will look more like the progress of a sustained wave of transformation.

AI: In summary

AI investment trends show characteristics of a speculative bubble, rapid funding, media hype, and inevitable comparisons to the dot-com era. However, AI is already delivering measurable value.

Organizations are moving beyond pilot projects into production-scale deployments, embedding AI in operations, customer engagement, and decision-making. The coming phase will test which innovations endure, but the broader trajectory points to sustained transformation, not collapse.

We’re not watching a bubble inflate; we’re surfing a powerful, sustained wave of technological change.

We've featured the best AI chatbot for business.

]]> https://www.techradar.com/pro/its-not-a-bubble-were-surfing-the-ai-wave rY94vD5vLcXWMo5ybFXJ8A Thu, 22 Jan 2026 15:07:36 +0000 Thu, 22 Jan 2026 15:07:36 +0000 <![CDATA[ Miss the iPhone mini? Now's the perfect time for small phones to make a comeback ]]> Towards the end of last year, I wrote that 6.3 inches is the best size for a smartphone, with the iPhone 17 and iPhone 17 Pro nailing the phone dimensions I want.

This isn't exactly a revelation: over the past decade, the screen sizes of all smartphones have ballooned. I remember when the Samsung Galaxy S6 Edge+ came out with a 5.7-inch screen, seen as a large screen at the time, yet now they have the Samsung Galaxy S25 Ultra with a 6.9-inch display and a design that pushes the definition of pocketability.

Apple did try to cater to people with smaller hands with the iPhone 12 mini and iPhone 13 mini, but these got discontinued pretty quickly – seemingly down to poor sales – and now the smallest phone Apple has is the iPhone 16e with a 6.1-inch display.

Off the top of my head, I can't think of a single big phone maker that offers a smartphone with a display coming in at under six inches, even on the budget side. The Samsung Galaxy Z Flip and Motorola Razr phones do go some way to making foldable phones with a small cover display and footprint when one needs to quickly scan a notification or two, but they aren’t exactly positioned as small phones.

While I doubt people’s hands have gotten bigger on average over the past decade, it seems like a huge number of us simply went with the flow when it came to accepting bigger phones.

However, jumping to the comments of my recent iPhone article, it appears that there is still a cadre of people who bemoan the rise of large phones.

Readers react

“Phones these days are practically phablets, and the iPhone 17 Pro Max at 6.9 inches is really a mini tablet,” said a commenter with the handle ‘zeino’, echoing a sentiment I somewhat agree with.

“I long for the days of a proper phone, with a 4-inch display (iPhone 5s) that you can pocket without it bulging or making it uncomfortable to sit, and still use one-handed in bed for long reading sessions, thanks to its relatively light weight of 112 grams (under 4oz).”

The commenter went on to theorize how even the iPhone mini with its 5.4-inch display could be considered too large and hence why it didn’t surge in popularity, noting they are now stuck with a first-generation iPhone SE with its 4-inch display in order to have a smaller phone.

Responding to zeino, commenter ‘MrBond007’ noted, “The iPhone 13 mini has a very loyal fanbase. There are still people who look for brand new iPhone 13 minis in 2025. It’s such a shame people overlooked the Mini because it really is an excellent phone, and with devices like MagSafe and portable battery banks, battery life isn’t really that big of an issue like people make it out to be.”

I remember someone on Instagram asking about phones and noting how small their hands are. Yet with no desire to move to Android, they ultimately decided they’d go for a refurbished iPhone 13 mini over newer and more capable Apple phones.

MrBond007 also responded directly to my article, stating: “I disagree completely and think anything above ~5.4 inches is too large and hard to use one-handed. iPhones have become way too large over the past 8-10 years (starting with the iPhone 6/6S Plus or whatever it was).

“It’s not just the size that’s the problem, it’s how heavy phones have become as well. A lot of people use their pinky to support their phone, and there are loads of comments online from people saying they get pain in their pinky finger because phones are so heavy now. I mean, the “pop socket” has been invented so people can instead rest their phone on their middle finger when using it, instead, a solution created for a man-made problem.”

Continuing, MrBond007 noted the iPhone 13 mini is the perfect size and how larger phones are “objectively worse” and linked a consumer obsession with size to a similar reason why people chose SUVs over smaller cars, despite being heavier, harder to maneuver, and offering less driving dynamics.

The end of plus size?

While I don’t think larger phones are objectively worse than their smaller predecessors – I’d say that in terms of performance, battery life, and camera arrays, they are definitely better– I do agree in part that there seems to be an accepted sentiment in the phones world that bigger is better without any solid logic.

On my commute, I often see people with smaller hands than mine tap away on plus-sized phones. It’s as if being forced to use a phone two-handed has become the accepted norm, while a few holdouts like myself try to use the 6-inch plus phones one-handed and then see them slip from my grip and dive bomb into the ground.

Despite there seeming to be an appreciation and desire for small phones, those that were on the market didn't seem to sell in big numbers, and mainstream phone makers don’t seem to be considering them; slimmer phones, yes, smaller, no.

Part of me thinks the likes of the iPhone 12 mini came at the wrong time. It arguably arrived when big phones were getting a lot of attention, with Samsung effectively bringing back the Galaxy Note series with Galaxy S22 Ultra and Apple expanding the screen size of its Pro phones. This surely made bigger phones the zeitgeist, with smaller ones and their smaller batteries left as a bit of an afterthought.

But now there’s a feeling we're pushing towards a time where people don’t necessarily want the biggest and best phones to dominate their life, with the idea of limiting screen time and going on digital detoxes being more prevalent; this is something I’m slowly becoming aware of given how easily I find it is to doom scroll on increasingly toxic social media and short attention-grabbing videos.

This is perhaps why we’ve seen people look back at using older phones, and the likes of Phone Writer Jamie Richard extol the virtues of ‘retro’ phones like the iPhone 5s. So maybe now’s the time for phone makers to explore going back to smaller and simpler phones that feel more like tools than time-sucking pocket computers. An early hint at this feeling I have can be seen with the Ikko MindOne Pro, a compact phone revealed at CES 2026 that looks like a slick antidote to ubiquitous big phones.

Of course, I’m lucky as what I’d call a medium phone (6.1 to 6.3 inches) fits my hands. But I’d not be against seeing a fresh slew of small phones that help cut out the ease to dive into social media and streaming, and instead work well enough to send some messages, follow directions on Google Maps, and snap the odd photo.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/phones/miss-the-iphone-mini-nows-the-perfect-time-for-small-phones-to-make-a-comeback BU6CbMwCven8ZtVrtRrvBa Thu, 22 Jan 2026 02:00:00 +0000 Fri, 23 Jan 2026 08:59:23 +0000 <![CDATA[ Report: Apple does about-face on Siri chatbot — and it might compete directly with ChatGPT and Google ]]> Apple promised us this was never the plan...until we guess, it became the plan: A new report from Apple soothsayer Mark Gurman says Apple will build an all-new Siri chatbot for iOS 27, one that looks and works differently but should remind most of us of an AI chatbot like ChatGPT.

To say this is a possible about-face is an understnatment and it's just one piece of the whirlwind that is apparently now Siri development.

Remember that pledge to deliver in early 2026 the Apple Intelligence Siri we were promised almost two years ago? That's apparently still happening, Gurman claims, via an iOS 26.4 update. If so, Siri will, with your permission, be able to see your screen, understand your actions in apps and other systems, becoming a more proactive and smarter assistant.

It's unclear if that ability is still what Apple meant when it announced earlier this month that it would use Google Gemini models to build its Foundation Models and enable the promised Siri. The only reason to doubt that iOS 26.4 will contain that is the timeline.

For Apple to deliver that update by late May (before it unveils iOS 27 in June), it probably must have been working with Google since late last year. Again, Apple is not commenting, so we have no insight into the development process.

Siri, we have questions

What will power this potential Siri AI chatbot is also a question mark. Will it still be tapping into Gemini foundation models and therefore be more or less a skin for Gemini? That seems unlikely, or at least too much of an admission by Apple that it was never up to the task.

But Gurman claims that not only will the iOS 26.4 update use those promised Gemini models (albeit built into Apple's Foundation Models), but the chatbot update (currently called Project Campos, per Gurman) will also rely in part on fresher, more powerful Gemini models and even use Google cloud servers.

That last bit sounds bonkers because it erases the security promise of Apple's Private Compute Cloud. Sure, since iOS 18, it's transparently sent some queries to either ChatGPT or Google, but to send potentially all Siri Chatbot prompts there is tantamount to abandoning a core privacy promise.

The problem with this Gurman theory is that it contradicts what Federighi told me almost a year ago at WWDC 2025 when we asked about what happened to the development of that long-promised Siri update.

"This wasn't about just building a chatbot," said Federighi last year, adding. "...That was never the goal, and it remains not our primary goal."

A hint of truth

So who are we to believe? Gurman or Apple executives. To be fair to Gurman, the only recent comment we have from Apple is that they are partnering with Google (Gurman claims it's for a $1 billion-a-year price tag, paid to Google). So it stands to reason that there is some truth here.

I agree with Apple that Google's Gemini provides some of the best models, and it is clear that Apple can't get there on its own, but the handing of privacy keys to a chief computer, one who already rules so much of our data, looks pretty anti-Apple. That's why I'll take all these rumors with a massive helping of salt and await Apple's next big AI reveal.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/phones/iphone/report-apple-does-about-face-on-siri-chatbot-and-it-might-compete-directly-with-chatgpt-and-google vdo6maNvMamVuvdBgfL35J Wed, 21 Jan 2026 20:44:05 +0000 Wed, 21 Jan 2026 20:44:05 +0000 <![CDATA[ Hardware phone keyboards are a fad built on iPhone keyboard frustration — and it might not last ]]> Blackberry, Palm, Samsung, Nokia: these were the four horsemen of physical keyboards – but far from instilling dread, back in the day we all enjoyed a love affair with their varied yet wonderful feature-phone keyboards.

Arguably, no one did it better than Blackberry, and I was on my final Blackberry Storm when I made the switch to the iPhone 4. What followed were weeks, if not months, of me retraining my digits to type on imaginary keys, and not send friends and coworkers absolute gibberish.

I wasn't alone in this, and for a time, many steadfastly clung to their Blackberries, Samsung Galaxy with slide-out keyboard phones running Windows Mobile, and Palm Pre handsets (and to this day, I miss Nokia's wild feature phones and their old T9 keyboards). The writing was on the wall for physical keyboards, however, and ultimately, all flagship smartphones turned into flat panes of glass.

Those born after 2005 and coming up in the age of AI likely have no memory of those old phones, and their almost clicky keyboards.

Somehow, though, the hardware-based smartphone keyboard is making a comeback – or at least a whole lot of buzz. At CES, I saw Clicks Communicator and Clicks Power Keyboard. The latter is a Magsafe-ready powerbrick and physical keyboard add-on for iPhones, and it reminds me a bit of Ryan Seacrest's ill-fated Typo keyboard case, which was just a tad too similar to Blackberry's keyboard for, well, Blackberry's taste.

The Clicks Power Keyboard also reminds me of a Blackberry keyboard; but with its ability to work with your iPhone, whether it's in portrait or landscape mode, it's clearly more versatile.

Even more interesting is the Clicks Communicator, a self-contained Android 16 phone with a smallish full-color screen and an ample keyboard – that's the one that's really turned people on. Now, in short order, they have another keyboard phone to consider, the Titan 2 Elite from Unihertz. Where the Clicks Communicator at least tries a more distinctive pill-shaped key design, Unihertz looks a lot like the iconic Blackberry Keyboard; somewhere, a patent lawyer is smiling.

Unihetz Titan 2 Elite — The Blackberry-like Unihertz Titan 2 Elite (Image credit: Unihetz)

I'm wondering: why this sudden fascination with hardware smartphone keyboards.

After all, Gen Z and Gen Alpha – the key market demographic at the dawn of the smartphone age – didn't grow up didn't grow up using physical keys. Even their laptop keyboard keys barely move. That said, those generations are obsessed with nostalgia and lo-fi tech. They like vinyl albums, tiny keychain cameras, film, retro game consoles, and even meet-cutes instead of Tinder.

It's not just them, though. I've spoken to colleagues and contemporaries who are excited to own a Clicks Communicator. Even the guy who helped birth the product, MrMobile (A.K.A. Michael Fischer) has some flecks of gray in his hair.

Looking at you, Apple

If I had to choose a culprit for this wave of excitement over pressing physical keys to control a modern mobile OS, it would have to be Apple. Why is it, I wonder, that after nearly 20 generations of iPhones and iOS, I am worse at typing on the iPhone's keyboard than ever?

Its predictive capabilities are often awful, and it doesn't seem to be able to use AI to understand the intention of my fingers. I spend more time trying to fix my texting and email typos than I do composing messages.

I feel I'm not alone in this. I hear complaints about iOS 26 being buggy, and some of those center around the keyboard. Apple has not made the keyboard better – and I'm sorry, but QuickPath does not cut it for me.

I've done a fair amount of typing on Android, too, and it's better, but not by much.

@techradar
♬ original sound - TechRadar

If Apple, Google, Samsung, and others can't fix typing on in-screen keyboards, they may face a wave of defections to Clicks Communicator and other hardware keyboard alternatives.

There is another scenario, though. This could be a blip; momentary and short-lived nostalgia for typing days gone by. Most of the nostalgia tech waves supported by Gen Z and others do not result in the creation of massive new markets. They make some money, and then people go back to using what everyone else is using.

That may be the case here; but for now, let's bask in the clicky fun of hardware keyboards, and enjoy the moment for as long as it lasts.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/phones/hardware-phone-keyboards-are-a-fad-built-on-iphone-keyboard-frustration-and-it-might-not-last nG4XM4Hs3tC2MwoZtNsH3V Wed, 21 Jan 2026 19:00:00 +0000 Wed, 21 Jan 2026 17:59:08 +0000 <![CDATA[ How to build AI agents that don’t break at scale ]]> The early success of AI tools is creating an illusion of readiness and scale that many organizations are not yet equipped to roll out or sustain.

What’s possible in a couple of carefully selected pilots is rarely applicable to large-scale deployments.

As you scale, workflows become less predictable, attention is spread thin and issues are not caught as quickly as before, which makes things more fragile.

Let’s look at where these gaps appear most often, and what needs to change for AI agents to hold up as expected once they are part of everyday work.

Unclear goals make agents less effective

While AI agent pilots are forgiving and people are more hands-on, one of the biggest issues with any scaling a project is starting without a clear goal. When teams don’t define exactly what they want an AI agent to do, the result is often something that feels unfocused or doesn’t solve a real business problem.

In fact, Gartner claims that over 40% of agentic AI projects will be cancelled by the end of 2027, due to escalating costs, unclear business value or inadequate risk controls.

The teams that see the best results start small and specific. They choose one clear task to automate and set simple expectations, which makes an AI agent easier to train and improve over time. This approach accelerates early wins and provides a clear blueprint for scaling an AI agent into other areas of the business.

Weak data foundations hurt performance

AI agents depend on accurate and up-to-date data. If the data feeding the system is messy or inconsistent, even a great model will struggle. In fact, Gartner predicts that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data.

First, leaders must define what constitutes AI-ready data. Then they must ensure that the data is representative of the AI use case, know whether it is interoperable across the business, how the data should be protected when being fed into AI models, and have a system for automatically detecting sensitive data.

Then data teams must prepare data pipelines to build an AI model dataset for training and for the live data feed to AI production systems based on the requirements gathered and once ready test and monitor it to optimize the models.

Thereafter, data observability processes are then needed to track data patterns and changes, adjusting data requirements as needed.

Lack of transparency erodes trust

Organizations must choose tools that provide visibility into an AI agent’s reasoning and behavior. As soon as AI agent projects are out of the pilot phase, it's not possible for humans to oversee everything. Transparency has to be embedded as an operating feature so that things can be debugged, updated, relied upon and trusted.

Executives are increasingly recognizing the value of AI observability. Platforms and frameworks that surface an AI agent’s reasoning, highlight anomalies, avoid context decay, and give business leaders confidence that the system is behaving as intended.

Stress-testing transparency as you would performance is a must. Instead of asking, “Does this make sense to the team that built it?”, the question should be, “Would this make sense to someone encountering it for the first time six months from now?”

Poor integration slows everything down

AI agents don’t work well in isolation. Even the most capable AI agent cannot deliver value if it cannot interact and orchestrate with the systems that drive the business. They need to talk and take action among systems a company already relies on — CRMs, ERPs, workflow tools, data platforms, and even older on-premise software.

Leaders should view integration as a strategic, composable design structure rather than a post-deployment task.

They must prioritize platforms that can connect seamlessly across modern cloud systems, traditional enterprise applications and legacy infrastructure. The result is not just an AI agent that works — but one that feels native to the organization's existing workflow ecosystem.

Security and governance come too late

As AI agents take on more important tasks, they often handle sensitive business or customer data. Still, many teams only start thinking about security after the AI agent is built.

The strongest approach is to embed security and governance early, such as access controls, audit trails, data protections, and live monitoring. This keeps AI agents safe and predictable as they grow so that what they reason, plan, and act upon is known.

Be explicit about what the agent is allowed to do on its own and where it must always pause and bring a person in. And don’t lock those choices in on day one and forget them. Watch where teams naturally step in or override an AI agent, because that’s usually telling you something important.

It is a company's responsibility to know how its agents are behaving just like with its employees. This proactive stance not only mitigates risk but also accelerates adoption by giving stakeholders confidence that the system is secure and governed to use at scale.

AI agents can’t adapt when business needs change

Business priorities, mandates, rules and policies shift all the time, and AI agents need to keep up. If they can’t evolve, they quickly become outdated. Without intentional mechanisms for retraining, evaluation and feedback, an AI agent that was once well-aligned can quickly become outdated.

AI agents must be treated as living systems that are continuously reviewed. Teams should gather feedback, update models and regularly review performance so that an AI agent keeps improving and is aligned with the current business so they remain strategic assets.

Building AI agents that last

As AI moves deeper into core operations, the organizations that succeed won’t simply deploy AI agents — they’ll cultivate them. Success depends on how honest you are with every AI Agent initiative from the very beginning.

Always ask whether you have done enough in the set up stages, if there are any gaps and if you are truly ready to scale. Accept any pitfalls upfront and act on them, and bring in third party partners as necessary to help at every stage.

We've featured the best AI website builder.

]]> https://www.techradar.com/pro/how-to-build-ai-agents-that-dont-break-at-scale Zefiq3MK8GnQ9ST64Vx45L Wed, 21 Jan 2026 15:28:30 +0000 Wed, 21 Jan 2026 15:28:30 +0000 <![CDATA[ Attackers are getting stealthier – how can defenders stay ahead? ]]> Attackers are increasingly abandoning noisy, direct attacks in favor of more subtle, stealthy tactics. They are flying under the radar and achieving long dwell times with the aid of more modular strategies and complex, multi-stage malware.

As a result, defenders must rethink their approach to expose these hidden threats before they strike.

Is evasion the new normal?

Cyber attackers will typically take the path of least resistance – if a tactic works, they’ll keep using it. When enough companies are armed with effective defenses, the offensive tools and tactics will adapt to get around them.

We’re seeing such a tipping point now. Traditional detection systems can pick up standard attacks reliably enough so criminal groups are exploring more subtle and complex options. Instead of launching large-scale, brute-force campaigns, threat actors are optimizing for persistence.

They want to stay embedded for weeks or months, quietly collecting data or preparing for more impactful attacks.

This stealth-first mindset is also driven by economics. Cybercrime-as-a-Service has made sophisticated evasion techniques widely available, meaning even low-level operators can buy stealthy, modular tools off the shelf.

OPSWAT’s own telemetry showed a 127% increase in malware complexity in just six months, typically in the form of multi-stage, obfuscated payloads designed to bypass static detection.

What are the most striking trends?

The skyrocketing complexity of malware is definitely one of the most alarming developments.

While there is no shortage of basic malware out there, more groups are deploying multi-stage, obfuscated script chains. They often combine PowerShell, JavaScript, and Batch scripts that disguise each stage of execution. These lightweight, modular loaders make detection more difficult.

The average multi-stage sample we analyzed this year contained 18 behavioral nodes, up from eight six months earlier. That leap demonstrates how attackers are layering more obfuscation and decision logic into each chain.

We’re also seeing attackers weaponize platforms that are in widespread use by employees to help their evasive tactics with command-and-control traffic increasingly hiding inside tools such as Google Sheets or Calendar.

Campaigns will also aim to exploit human behavior, such as the ‘ClickFix’ attack, which tricks victims into running malicious Windows Run prompts. There’s no malware payload or malicious URLs involved, so nothing for standard tools to detect.

Why are so many defenders still struggling to detect these threats?

Many organizations are still operating with a “known threats” mindset, relying on static, signature-based tools to flag anything familiar. This has been fairly successful for many years, to the point where attackers have had to switch it up.

But today’s more advanced malware rarely looks the same twice. Obfuscation, encryption, and fileless execution mean there’s often nothing for these systems to match against.

We found that one in 14 files dismissed as benign by public feeds turned out to be malicious when behaviorally analyzed. That’s a huge blind spot for any defensive strategy relying on known threat signatures.

How can organizations cope with these stealth attacks?

Standard detection tools will still be necessary for the myriad lower-level threats making the rounds, but security architects must also adopt a behavior-first strategy.

This focuses on determining what a file does, not just what it looks like. Instead of trusting static indicators, defenders analyze execution in real time to see how a file behaves. This includes assessing what processes it spawns, which registries it touches, and how it interacts with the network or system memory.

Adaptive sandboxing and emulation can safely expose hidden behaviors such as memory-only malware or scripts that decrypt and execute only at runtime.

At OPSWAT, we’ve seen that combining behavioral analysis with machine learning-powered similarity search achieves 99.97% detection accuracy. This approach accelerates detection as well as precision, revealing new threats up to 24 hours before they appear in public OSINT feeds.

What else security leaders can do to build resilience

One of the most important factors in building true resilience is to take a multi-layered approach.

So, while a behavior-first approach is crucial, there are other elements to consider. One important capability is the implementation of data diodes. These are hardware units that enforce unidirectional data flow, providing an effective barrier to the more subtle C2 and exfiltration tactics.

Alongside this, deploying a Managed File Transfer (MFT) solution will provide another layer of defense by automatically blocking and sandboxing potential risks.

Finally, another valuable approach is to implement Content Disarm and Reconstruction (CDR). This treats all incoming files as malicious by default and will rebuild and sanitize them to fully remove any hidden threats before they are allowed into the system.

The most important step is to accept that stealth is becoming the default attack mode. Defenders need to evolve accordingly by prioritizing adaptability across people, processes, and technology. Continuous detection and response should replace point-in-time scanning, while threat intelligence must flow freely between tools and teams.

Security leaders should focus on knowing how a threat behaves. Building behavioral context into every stage of the security pipeline helps teams make faster, smarter decisions.

Above all, successful defense demands a continuous approach to security. The organizations that learn from each incident will always stay ahead.

We've featured the best online cybersecurity course.

]]> https://www.techradar.com/pro/attackers-are-getting-stealthier-how-can-defenders-stay-ahead rj7s2R5zcJmTiFm8skMdiU Wed, 21 Jan 2026 15:06:54 +0000 Wed, 21 Jan 2026 15:06:53 +0000 <![CDATA[ ‘A new era of e-bike use’ or ‘the dumbest ebike law ever’? E-bike restrictions are hitting New Jersey, and I’m worried it’s just the beginning ]]>

New Jersey just introduced tighter e-bike rules

Other states are considering similar restrictions, too

Many e-bike advocates are frustrated by the new law

New Jersey is imposing new e-bike legislation that will seriously impact riders in the state – and could be a sign of tighter restrictions rolling out more widely in the US and beyond, which could have serious ramifications for the adoption of e-bikes and similar technologies like e-scooters.

Under the new law, all classifications of e-bike ridden in New Jersey will require the rider to get a license, get insurance, and register their vehicle – a lot like you would for a car or motorbike. You must also be aged at least 15 and have a motorized bicycle license or be at least 17 years old with a valid driver’s license to own and operate an e-bike.

The state has also banned conversion kits that would allow riders to convert their lower-powered e-bikes to higher-powered ones, and for the next year, it has banned the sale of electric motorized bikes online (the fastest tier of bikes, which can go at speeds over 28mph).

E-bike jump — (Image credit: Canyon Bicycles)

Lawmakers who proposed the bill, and supporters, have heralded this as a major win for e-bike safety, with bill sponsor Senate President Nick Scutari declaring, “We are in a new era of e-bike use” – showcasing the local government’s action to help prevent e-bike-related injury and deaths following some high-profile incidents late last year.

E-bike enthusiasts are a lot less positive about New Jersey’s actions, however. You’ll find various articles and social media posts like this one from Chris Crossed, calling this “the dumbest ebike law ever,” as they point out issues like the lack of insurance options in the current market, as well as the conflict the law has with efforts to cut carbon emissions and reduce traffic congestion.

Though these voices are not entirely opposed to all of the actions being taken. Instead, they’re concerned that the broad, rather than targeted, measures could harm interest in e-bikes, which offer a solid alternative travel method for many. Especially as they fear New Jersey won’t be the last to impose restrictions – case in point, New York State has a similar senate bill in the works.

As e-bike fans would themselves be quick to point out, e-bikes are not a well-defined category. The term can be used to refer to pedal bikes that offer minimal assistance (topping out at speeds below 20mph) as well as effective e-motorbikes that can go at speeds over 28mph and aren’t really intended to be pedalled.

Clearly, there’s a major difference between these kinds of vehicles, but the very generalized category name combined with low understanding of the specific differences between subcategories means it’s easy for folks to buy e-bikes with a lot more oomph than they realize (leading to accidents) and lawmakers being overly broad with their rules.

Unfortunately, there’s also clearly a need for tighter rules to be brought in, because e-bike injuries are on the rise (via Velo and the BBC). Part of this comes from increased ridership, but that alone doesn’t account for the increase – factors including increased power and weight for e-bikes are also seemingly playing a part.

Mycle Cargo e-bike — (Image credit: Mycle)

Is there a happy middle ground?

I’ve written about being a major fan of e-scooters, and e-bikes to some extent too, as they offer commuters a great alternative to driving that’s more environmentally friendly, fairly convenient, and not as physically taxing as riding a non-e-bike. So I’m not keen on overly restricting access.

Instead, I’d follow the example set by various countries’ drone rules. E-bike manufacturers would start by getting their bikes approved for use on public roads, and they would be assigned a category based on their top engine speed and weight.

The lightest, slowest speed e-bikes would be accessible to younger riders given they’re comparable to a pedal bike with just a little assistance; medium-weight and speed bikes would be restricted to older riders and require some kind of online test that checks their understanding of road and rider safety; and then those heaviest and fastest ‘e-motorbikes,’ as they’re called, would require some kind of license akin to a driver’s license.

Hamsh Hector riding the Xyber e-bikes — (Image credit: Segway)

All bike types, however, would need to be registered by an adult and be tagged with their registration number. Registration could be done online, and would be similar to how UK drones need to have your Operator ID on them.

This is just one idea, and I’m keen to hear your alternative suggestions in the comments, but I feel it strikes a better balance between improved safety and keeping e-bikes relatively accessible than do the New Jersey rules.

We’ll have to wait and see how other jurisdictions tackle e-bike rules, but as I sit here hoping for looser e-scooter laws in the UK, I have a suspicion we might, in fact, see tighter restrictions in various places. Hopefully, my hunch is wrong.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/health-fitness/bikes/a-new-era-of-e-bike-use-or-the-dumbest-ebike-law-ever-e-bike-restrictions-are-hitting-new-jersey-and-im-worried-its-just-the-beginning jAoFqbsXhkeKBt84jgz6cg Tue, 20 Jan 2026 18:00:00 +0000 Tue, 20 Jan 2026 15:56:04 +0000 <![CDATA[ Threat hunters can’t waste time stumbling in the dark – they need real observability ]]> The worst security breaches rarely begin with fanfare. Rather, the most dangerous attacks are the ones that quietly slip past the perimeter and begin spreading while defenders remain unaware. It’s the time attackers spend under the radar inside the system that turns an intrusion into a disaster.

Yet, despite years of investment in preventative tools, we still see organizations consistently struggle to detect and contain attackers once they are inside.

Often, the shortfall isn’t a lack of visibility or security alerts, but a lack of clarity. In today’s hybrid environments, resilience depends less on blocking every threat and more on spotting the ones already within reach.

The growing gap between detection and real visibility

It’s no secret that security teams are feeling increasingly overwhelmed, and it’s easy to see why when you start drilling into the numbers.

Our research shows that a typical organization faces more than 2,000 alerts every day, and much of this is noise offering little real value.

Analysts spend more than 14 hours each week chasing false positives, and two-thirds of leaders admit their teams simply cannot keep up. Missed alerts quickly turn into missed opportunities to stop attackers early.

Tooling complexity adds to the problem. While most organizations now use multiple cloud detection and response platforms, we found that almost all (92%) still report significant capability gaps.

More data does not necessarily equal better detection, and overlapping systems create fragmented visibility and conflicting information. Without meaningful context to tie these signals together, defenders are left piecing together fragments of a story rather than seeing what truly matters.

Why lateral movement remains the attacker’s favorite blind spot

The widespread challenges of separating signal from noise are a massive boon to threat actors, who are increasingly favouring low-and-slow tactics. Once in an organization, rather than act immediately, they often creep through the network, escalating privileges, probing workloads, and searching for sensitive systems.

This lateral movement is where small breaches escalate into major operational crises, and it remains one of the most difficult stages of an attack to detect.

It's paying off for cyber attackers, with nearly 9 in 10 organizations telling us they experienced an incident involving lateral movement in the past year. On average, these breaches resulted in more than seven hours of downtime, with ongoing operational disruption and full recovery stretching on even longer.

These incidents persist because east–west traffic in modern hybrid environments remains poorly understood. Even when organizations believe they are monitoring internal communications effectively, almost 40% of that traffic lacks the context required for confident analysis.

Attackers thrive when defenders are overloaded, uncertain, or unable to distinguish legitimate activity from the early signs of an intrusion spreading through the network.

The importance of observability

Effective defense against these threats demands deep observability. There has been an ongoing push in this direction from industry bodies, with the UK’s National Cyber Security Centre (NCSC), for example, recently issuing guidance.

The NCSC stresses that organizations cannot hunt for threats they cannot see, and traditional indicators of compromise are no longer enough. Instead, defenders need visibility across behaviors, patterns, identities, workloads and east–west traffic to uncover the subtle signals that reveal an attacker already in motion.

This aligns closely with our own analysis on the lack of context for internal traffic, despite widespread confidence in monitoring capabilities.

Observability must move beyond collecting more logs. It requires understanding how systems relate, behave and change over time, and connecting those insights before an attacker does.

Why context, correlation and containment must replace alert-hunting

For years, we’ve seen security programs react to rising attack volumes by gathering more data. But far from keeping up, this approach often only serves to intensify alert fatigue.

Because large portions of network traffic still lack the context needed for meaningful investigation, analysts end up sifting through unprioritized alerts rather than focusing on attacker behavior.

What security teams need is a connected view of their environment, not isolated signals. Contextual models, such as security graphs, help map the relationships among workloads, identities, devices, and data flows.

They turn scattered indicators into a coherent picture. A low-level alert on one system can suddenly make sense when linked to suspicious behavior elsewhere, revealing attacker intent rather than isolated anomalies.

This shift from alert-hunting to understanding pathways is essential for breach containment. When defenders can see how systems interact and where the most sensitive assets sit, they can identify the routes an attacker is likely to take.

That clarity allows teams to act with confidence, slowing or stopping lateral movement before it spreads.

AI, automation, and scaling human judgment responsibly

As environments expand, the volume and complexity of security data have surpassed what human analysts can manage alone. This is where AI and automation play a critical role.

Many organizations are turning to AI and machine learning to improve detection accuracy and accelerate response times, seeing these capabilities as central to spotting lateral movement earlier and easing the burden of alert fatigue.

However, it’s a mistake to believe that investing in AI will solve everything on its own. AI is most effective when it augments, not replaces, human expertise. Automated systems can correlate signals across hybrid environments, enrich them with context and filter out noise, giving analysts a more precise starting point.

Combined with a security graph approach, for example, AI-powered analysis can continuously plot every workload and connection in real time, highlighting behavioral patterns that would be impossible to surface manually.

This creates a force multiplier that enables faster, more confident decisions and supports the rapid containment that modern resilience demands.

AI-powered security graphs can connect the dots of seemingly disparate network events, establishing a connected narrative where busy human analysts might see isolated alerts.

For example, a workload accessing a database it has never accessed before could be traced back to a misconfigured identity, revealing an attack path that is already being exploited.

What this means for business and IT leaders

For leaders, the path forward begins with recognizing that resilience depends on what happens after an attacker gets in. That means investing in observability across the entire hybrid estate, not only perimeter logs, but the identities, workloads, cloud services and east–west traffic that reveal how attacks unfold.

It also requires shifting detection strategies toward behaviors and relationships, supported by threat hunting and hypothesis-driven investigation. Automation and AI can help, but only when grounded in high-quality, contextualized data.

Finally, success should be measured not by the number of blocked threats, but by how quickly organizations can detect, contain and recover from an intrusion.

Breaches are now an inevitable part of operating in fast-moving, hybrid environments. What matters is how quickly an organization can recognize when something is wrong and limit the impact.

We've featured the best encryption software.

]]> https://www.techradar.com/pro/threat-hunters-cant-waste-time-stumbling-in-the-dark-they-need-real-observability tmCY8Yni5fsvCGbWnbsJTh Tue, 20 Jan 2026 15:16:39 +0000 Tue, 20 Jan 2026 15:16:39 +0000 <![CDATA[ Simplifying your business response to worst-case scenarios ]]> Over the course of the early 2020s, businesses have faced a barrage of challenges that few could’ve predicted. From global supply chain disruptions and pandemic-driven shutdowns to on-and-off tariffs, these unexpected events have tested the limits of traditional planning strategies.

The recent United States government shutdown provided yet another reminder that well-established systems can come grinding to a halt, lasting a record-breaking 43 days and gridlocking everything from local project funding to air traffic control and food assistance programs.

These scenarios demonstrate that rare and complex disasters are difficult to anticipate, and even harder to model and prepare for. While worst-case scenario planning has long relied on historical data and contingency buffers, what happens when history offers no precedent?

This is where mathematical optimization plays a key role in simplifying and streamlining planning efforts.

The worst-case scenario

If you’re familiar with optimization techniques, you may identify the process of optimizing under the worst-case as “robust optimization.” While it’s a noteworthy application of optimization, that’s not precisely what this article is about.

Instead, imagine that you run a mid-sized manufacturing company based in the United States. Your business relies on timely regulatory approvals in order to ship products across state lines.

During a prolonged government shutdown, these approvals stall, leaving your team with a steadily growing number of finished products and nowhere to send them. Inventory piles up, storage space disappears, employees idle with no work to do, and contractual delivery deadlines loom.

This stalled scenario is complex and full of interdependencies. Storage limits, budget caps, and fixed workforce hours can’t be adjusted, and missed deadlines could trigger financial penalties and damage customer relationships.

There are certainly factors that you can control—production scheduling, resource and labor allocation, and alternative sourcing, for example—but none can be adjusted without having their own downstream impacts on your bottom line.

How might you effectively plan for this kind of situation?

The traditional response

Businesses have traditionally approached worst-case scenario preparation with manual planning methods. In our shutdown example, this might include your manufacturing company’s leadership team getting together to review spreadsheets, examine historical data, and brainstorm scenarios to estimate their potential impact.

In itself, this is an entirely plausible method. Your team could review a range of past data, talk through what the scenarios may entail in the future, and implement relevant contingency buffers as a means of proactive response. The success of this approach, however, is largely dependent on predictability.

It’s not unlikely that historical production data could help a manufacturer prepare for spikes in demand or the loss of a strategic supplier. But if there’s anything we’ve learned from the first half of this decade, it's to expect the unexpected.

Volatile supply chains, fluctuating international markets, global conflicts, and widespread health crises are not as easy to predict and effectively prepare for.

And when they’re layered atop dozens of preexisting variables—storage limits, labor hours, contracts, and more—the traditional response quickly becomes slow, reactive, and generally inadequate.

Responding with optimization

Mathematical optimization offers today’s businesses a fundamentally different approach to worst-case scenario prepping. Rather than relying on gut instinct and static data, optimization leverages advanced algorithms to assess multifaceted challenges and provide the best possible solution.

Each mathematical optimization problem includes:

An objective function, or the goal you’re trying to achieve.
Decision variables, or the factors that can be changed.
Constraints, or the factors that cannot be changed.

In our manufacturing scenario, these factors would be:

An objective function of minimizing total cost while still meeting delivery obligations.
Decision variables such as production levels, labor allocation, and outsourcing decisions.
Constraints that include storage capacity, budget, workforce availability, and regulatory restrictions.

Your team could use these variables to create a mathematical model that represents your problem. Then, using an optimization solver to run the relevant algorithms, you could determine the optimal balance between competing priorities in minutes, rather than days.

The model might recommend reducing production by a certain amount, reallocating workforce resources to preventative maintenance projects, and/or outsourcing a portion of production to avoid storage overflow.

Whatever the recommendation, you can rest assured knowing it was informed by all relevant variables and determined to be the best possible solution.

Using optimization drastically streamlines the preparation process, making worst-case scenario considerations and decisions faster and more reliable.

Human decision-makers are still empowered to share key input and have the final say on how the team moves forward, all while outsourcing the time-consuming computational work to the power of steady, trustworthy, and easily updatable mathematical models and the advanced algorithms that solve them.

Towards a less frantic future

Worst-case scenarios are no longer as rare as they used to be. From pandemics and tariffs to lengthy government shutdowns, disruptions are more frequent, more complex, and more difficult to predict.

Mathematical optimization gives today’s businesses a smarter way forward. By taking on the legwork of scenario modeling and review, optimization can help transform lengthy and uncertain manual planning into proactive business strategies.

While an optimization model cannot completely eliminate unpredictability and risk—as no solution truly can—it can act as a metaphorical “rock in the storm,” helping teams react to their changing environment with fast and reliable problem-solving capabilities.

When the next unprecedented disaster strikes, optimization-driven businesses can rest assured in the results of their proactive efforts, rather than scrambling to make up for a lack of preparedness.

We've featured the best business intelligence platform.

]]> https://www.techradar.com/pro/simplifying-your-business-response-to-worst-case-scenarios gsxjJX5FELjFGKqdwFiWyR Tue, 20 Jan 2026 14:59:19 +0000 Tue, 20 Jan 2026 14:59:18 +0000 <![CDATA[ ‘ChatGPT Go might be worth the downgrade’ — OpenAI's new money-saving tier costs 60% less than Plus ]]> As a ChatGPT Plus subscriber, I’ve been eyeing OpenAI’s new ChatGPT Go subscription plan with interest. In fact, I think it might even be worth the downgrade. Let me explain…

ChatGPT Go is a lightweight version of ChatGPT that sits somewhere between the Free tier and the Plus tier in terms of what it offers. While it was already available in some territories, last week ChatGPT Go went global, arriving in the US for the first time.

So, if you’re currently subscribed to ChatGPT Plus and don’t really push its limits, downgrading to Go could give you everything you need. There is, however, one notable drawback, which I’ll come to shortly.

Here are the current ChatGPT US subscription prices:

ChatGPT Go: $8/month
ChatGPT Plus: $20/month
ChatGPT Pro: $200/month

(In the UK, it’s £8 for Go and £20 for Plus.)

That means ChatGPT Go is 60% cheaper than Plus. A $12 saving each month adds up to $144 a year, so it’s worth taking a closer look to see whether switching could save you some serious money.

ChatGPT Go vs ChatGPT Plus

A free ChatGPT account is usable, but very restricted. You get limited access to ChatGPT-5.2 before it drops down to a less powerful language model, tightly capped image generation, limited Deep Research access, fewer file uploads, and restricted memory.

With a Plus account, those limits are dramatically higher. You can generate far more images than most people will manage in a day, benefit from much stronger memory, choose between more language models (including the popular ChatGPT-4o), and get increased access to Deep Research. On top of that, Plus subscribers get access to Codex for coding and Sora for video generation.

ChatGPT Go sits neatly between these two tiers. You get increased access to ChatGPT-5.2, improved memory, more image generation, more messages and uploads, and longer memory retention. What you don’t get is access to Codex or Sora.

Here’s the key point: if you’re a casual ChatGPT user currently paying for Plus, and you’re not generating large numbers of images or creating videos in Sora, then downgrading to ChatGPT Go could make a lot of sense. The Plus limits are so generous that many users are probably nowhere near hitting them anyway, and a Go account could be all you need.

The sting in the tail

While downgrading to Go may be tempting, there’s one final thing to consider. OpenAI has announced that it will start showing ads in ChatGPT, but only on the Free and Go tiers.

For now, ads are rolling out in the US only, but OpenAI has confirmed they will eventually expand to other territories.

We’ve already seen a preview of how these ads will appear. While they’re not ideal, they seem relatively unobtrusive and are contextually relevant to the conversation you’re having.

Ultimately, it comes down to whether the $12-a-month saving is worth having ads appear in your chat conversations. For some users, that trade-off will be an easy win. For others, staying ad-free with Plus may still be worth the extra cost.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/ai-platforms-assistants/chatgpt/chatgpt-go-might-be-worth-the-downgrade-openais-new-money-saving-tier-costs-60-percent-less-than-plus 63Qq3WFC9urHBs59jyz6yN Tue, 20 Jan 2026 09:16:24 +0000 Tue, 20 Jan 2026 09:16:25 +0000 <![CDATA[ Apple really went backwards when it ditched titanium on iPhones ]]> So there I was, thinking about how the iPhone 17 having a larger, and correct size, 6.3-inch display complete with a 120Hz refresh rate, would mean I wouldn't miss the big specs of my iPhone 16 Pro Max.

Heck, the new Fusion main camera even goes some way to making me forget my love of telephoto cameras. So all was looking rather rosy…

Until the iPhone 17 with its satin-finished back and sides slipped from my hands.

The first time this happened the phone came away with a solid dent in its lower frame. The second, a sort-of horizontal splat onto a sidewalk left shattered glass in the top and bottom corners diagonally to each other, with hairline fractures connecting the two.

an image of a cracked iPhone 17 — (Image credit: Future)

Of course this is down to my own dumb clumsiness, which seems to have got worse over the years. And as I’m not a fan of cases ruining the slick industrial designs of the best phones, I tend to raw-dog my iPhone experience, leaving the hardest I test out at the mercy of the elements and my fumblings.

But as I’ve written before, all this was fine with the past two generations of Pro iPhones as they made use of a titanium chassis, which proved pretty much immune to Roland foolery.

Yet that’s not the case with the iPhone 17, with its aluminum frame. Or indeed the iPhone 17 Pro and iPhone 17 Pro Max, both of which have ditched the titanium chassis for aluminum in order to save weight.

an image of a cracked iPhone 17 back — (Image credit: Future)

Given the above, I’m not a fan of this. And I also think that at its current price, the standard iPhone could make use of titanium. I get that it’s harder to work with. But Apple touts sustainability, so surely having stronger phones out of the box would save on energy, and indeed money, spent on repairs.

And while I’m talking durability, I still think more could be done, especially on flagship phones.

Sure, Gorilla Glass is tough and weathers scratches, and Apple’s own Ceramic Shield coating is very good on the display. But outside the use of titanium, I’ve yet to see a glass or coating that truly makes iPhones, Pixels phones and Galaxy S-series handsets truly durable.

Maybe I’m just trying to eschew blame for my own butterfingered nature, possibly aided by enjoying the holiday season festivities with a little too much gusto. But here’s to harder, stronger, tougher phones in 2026.

What’s do you reckon? Do phones need to toughen up? Or are you a case user? Let me know in the comments below.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/phones/iphone/apple-really-went-backwards-when-it-ditched-titanium-on-iphones 2YYNGiEcqjF9RvMjrQUyqC Mon, 19 Jan 2026 23:00:00 +0000 Mon, 19 Jan 2026 17:17:14 +0000 <![CDATA[ Why agentic AI pilots stall – and how to fix them ]]> Agentic AI is the latest buzz in boardrooms. Unlike generative AI tools, agentic AI act as autonomous agents that can reason, make decisions, and act across workflows to achieve goals. Done right, they promise to reduce manual work and unlock new levels of productivity.

But many early adopters of AI tools are struggling. Pilot projects stumble, costs escalate, and results fail to match expectations. The problem isn’t that agentic AI is overhyped, it is that businesses are moving too fast without the strategy, infrastructure, and data foundations required to make them work as intended.

And this isn’t surprising when you consider that 80% to 90% of all enterprise data is unstructured -- based on multiple analyst reports in recent years.

As someone who has built platforms through multiple waves of ‘intelligent automation,’ I’ve seen firsthand the same repeated patterns: technology alone doesn’t transform organizations -- alignment, governance, and cultural readiness do. The real breakthrough comes when innovation is grounded in trust and connected to business outcomes.

Where conventional AI might sort invoices, an agentic AI could approve payments, flag anomalies, and update compliance systems. That leap demands a contextual understanding of how data, processes, and rules fit together.

Too many organizations are treating agentic AI as bolt-on upgrade, as if they’re simply more advanced chatbots. The reality is more complex: agentic AI needs to be woven into the enterprise fabric, connected to the right data and workflows, and supported by governance. Without that foundation, autonomy quickly becomes chaos.

Infrastructure first

One of the biggest stumbling blocks is infrastructure. Many enterprises still run on siloed content repositories, legacy systems, and fragmented integrations. In these environments, agentic AI can’t access the full unstructured data they need to perform at their best.

In government, for example, content and processes are spread across different agencies, often using decades-old applications. Asking an AI agent to make decisions without integrating those systems is like asking it to assemble a puzzle with half the pieces missing.

Preparing for agentic AI requires investing in cloud-native foundations and interoperable content platforms that unify information and enable seamless connections across applications. Without this groundwork, agentic AI risk acting on partial or outdated information, and making flawed decisions as a result.

Bad data kills autonomy

Even with the right systems in place, poor data quality is a critical flaw. Agentic AI thrives on complete, accurate, and governed information. If datasets are inconsistent or scattered, agentic AI can’t make sound decisions.

Healthcare illustrates this challenge clearly. An agent supporting clinicians must pull from medical histories, lab results, and imaging data in real time. If one piece is missing or misaligned, the recommendations these agentic technologies produce could be flawed.

The lesson for early adopters is clear: start with a data audit and gain a firm understanding of where your unstructured data is. Know what you have, where it lives, and how it’s governed before handing decision-making power to AI.

Getting governance right

Another misconception is that agentic AI removes people from the loop. In reality, the most effective early use cases blend autonomy with oversight.

Take financial services. Agentic AI may verify documents and draft compliance reports, but humans still make the final call on high-risk cases, or how to proceed when a document is flagged by an agent. This balance accelerates workflows without eroding trust and accountability.

Strong governance must be embedded from the outset, covering regulation, ethics, and operational control. Without it, these agents risk amplifying bias, undermining trust, and exposing organizations to compliance failures.

Lessons from early adopters

The experiences of early adopters reveal three clear lessons.

The first, projects work best when they begin with a clear business outcome, not a fascination with the technology or jumping on a trend. Organizations that take time to define the processes they want to improve and the results they need to achieve are the ones seeing value.

Second, they invest early in the groundwork. Modern infrastructure and clean data may not grab headlines, but they are essential to making the headline-grabbing innovations possible.

And finally, they treat autonomy as something to scale gradually. The most effective implementations begin with human-in-the-loop models and only expand to greater autonomy once confidence and maturity grow. This approach builds trust in the technology while maintaining accountability.

These early lessons are already shaping a picture of maturity.

The shape of maturity

As agentic AI matures, it will move beyond isolated experiments and towards interconnected systems. The real breakthrough will come from agentic AI networks coordinating across workflows.

In a hospital, for example, one agent might surface patient histories, another manages scheduling, and a third flag billing issues; all contributing to a shared context that supports clinicians.

Proof-points will become a non-negotiable. Businesses will expect agents to show their work, like the data they used, the reasoning they followed, and the compliance checks they applied. Without this transparency, agentic AI won’t be trusted to handle sensitive or high-value work.

And the technology landscape itself will have to open up. Organizations will want the flexibility to integrate agentic AI powered by different models, switch providers as needs evolve, and scale across hybrid or multi-cloud environments. Flexibility and interoperability will be essential to protect long-term investments.

Beyond the hype

Far from failing, agentic AI is in its adolescence. Just as cloud computing went through a difficult transition phase before proving indispensable, agents too will require a period of adjustment.

The organizations that succeed will be those that prepare best, not adopt the fastest. By aligning strategy, modernizing infrastructure, cleaning data, and embedding governance, enterprises can move from experimentation to transformation.

With the right foundations, agentic AI can do far more than just automate tasks. It will enable genuinely intelligent systems that reshape how work gets done – and that could be the most significant shift in enterprise technology for a generation.

We've featured the best AI website builder.

]]> https://www.techradar.com/pro/why-agentic-ai-pilots-stall-and-how-to-fix-them Mpjic87QprpUhuWnQnhXg8 Mon, 19 Jan 2026 15:23:11 +0000 Mon, 19 Jan 2026 15:23:11 +0000 <![CDATA[ Solving AI's energy challenge: sustainable data centers for a competitive UK future ]]> The world is racing to harness AI's transformative potential, but this revolution comes with a stark reality: AI's energy appetite is enormous and growing rapidly.

Building AI infrastructure efficiently and meeting net-zero commitments is crucial. For the UK, this isn't just an environmental imperative; it's an economic one.

I’ve spent my career at the intersection of technology, energy, and government action for net zero converge. I’ve always been fascinated by this challenge and when it comes to unlocking game changing innovations often been inspired by the Silicon Valley ethos of “move fast and break things”.

But for data centers, I see a new mantra: move fast, but break nothing.

The UK faces a perfect storm. A "rush" to build data center capacity, driven by AI's insatiable computational demands and the critical need for data sovereignty. Simultaneously, the UK's high energy costs already pose a significant barrier to competitiveness.

Building energy-hungry infrastructure on an expensive, carbon-intensive grid is economically unsustainable and threatens net-zero targets. We cannot afford to build infrastructure that locks us into spiralling operational costs and carbon debt.

The three pillars: A virtuous circle

The solution lies in understanding that efficiency, affordability, and sustainability aren't competing priorities, they're inextricably linked in a virtuous circle.

Efficiency means designing data centers that extract maximum computational value from every watt consumed. This involves advanced cooling systems minimizing water and power use, predictive capacity management preventing energy waste from underutilized infrastructure, and optimized hardware.

Affordability flows directly from efficiency. Lower energy consumption means reduced operational costs - critical when UK energy prices remain among the highest in Europe.

Right-sizing infrastructure, rather than over-provisioning, prevents capital waste and ensures capacity matches demand. This economic logic makes a compelling business case for green infrastructure, benefiting both the environment and bottom-line competitiveness.

Sustainability becomes the natural outcome when efficiency and affordability converge. Reduced energy demand means less strain on the grid, lower carbon emissions, and infrastructure contributing to national net-zero targets.

The UK's vast offshore wind potential becomes a genuine competitive advantage when data centers are designed to operate efficiently on renewable power.

A change of mindset

Achieving this virtuous circle for data centers demands a change of mindset.

This calls for 'engineering thinking': a zero-rate-of-failure mindset where reliability and efficiency are designed in from the outset, not retrofitted later. Just as failure is catastrophic in a power grid or transportation network, the same rigor must apply to AI's data infrastructure.

This thinking manifests in concrete innovations. Consider, for instance, the emergence of advanced 800-volt direct current (VDC) power distribution systems as a significant step forward in optimizing power delivery for high-performance AI.

Historically, data centers relied on alternating current (AC), requiring multiple conversion steps from grid to racks, with each conversion introducing inherent energy losses that collectively waste substantial power.

However, pioneering efforts in high-density computing environments are now demonstrating the efficacy of 800 VDC systems. By delivering direct current power closer to or directly to the AI computing clusters, these systems dramatically reduce the number of conversions needed.

This translates to substantially improved efficiency, often cutting conversion losses significantly. Fewer conversions also mean consistent voltage, reduced heat generation, lower cooling needs, and a more compact power infrastructure footprint.

This isn't incremental; it's rethinking infrastructure from first principles to support extreme AI power densities sustainably.

The ecosystem approach: Partnerships and local leadership

Strategic partnerships accelerate this transformation. These collaborations, for example between technology providers and infrastructure specialists, combine frontier AI with deep infrastructure expertise, translating innovation into deployable solutions.

These relationships ensure evolving infrastructure that supports AI's rapid advancement, avoiding wasteful rip-and-replace cycles.

Yet technology alone isn't sufficient. Whilst central government sets high-level policy, local leadership plays a critical role in sustainable data center development. UK mayors and local authorities can proactively create "energy hubs" with existing water and energy access (e.g., ports) and reimagining them as strategic digital infrastructure.

Integrating data center development into sustainable urban plans ensures digital growth aligns with local environmental and economic objectives, creating jobs and driving prosperity while meeting national climate commitments.

Building infrastructure that delivers on climate commitments

For too long, data center climate impact was opaque, with inconsistent energy consumption and carbon accounting. As AI accelerates, it must align with global climate and net-zero commitments.

This means rigorous, transparent measurement of energy consumption and carbon emissions, with sustainability designed in from the outset, not retrofitted as an afterthought.

The UK possesses unique advantages - offshore wind resources, strong governance, and growing data sovereignty demands.

By prioritizing efficiency, fostering partnerships, empowering local leaders, and ensuring transparency around climate impact, the UK can build data centers that power AI responsibly whilst strengthening competitiveness and advancing net-zero commitments.

It is imperative that engineering rigor and collaborative thinking ensure AI's development enhances, rather than undermines, our climate future.

We've featured the best green web hosting.

]]> https://www.techradar.com/pro/solving-ais-energy-challenge-sustainable-data-centers-for-a-competitive-uk-future JHedwVcNR2ZoFbAbUp53Se Mon, 19 Jan 2026 10:52:03 +0000 Mon, 19 Jan 2026 10:52:03 +0000 <![CDATA[ Passwords are still a problem for UK businesses - what next? ]]> Despite years of public awareness campaigns, repeated high-profile breaches and ongoing regulatory scrutiny, the use of weak, guessable passwords such as “admin” and “123456” persists in the UK. This problem is emblematic of a deeper, systemic issue in the country’s cybersecurity posture.

However, this is not a failure of awareness or education, but a failure of credential management, cultural reinforcement and policy enforcement at an organizational level.

In early December, the National Cyber Security Centre (NCSC) published updated guidance on credential management. Specifically, the guidance “advocates a greater reliance on technical defenses and organizational processes, with passwords forming just one part of your wider access control and identity management approach.”

This represents a shift away from user-centric security models and puts the onus on organizations to secure user credentials and, ultimately, the systems the organization relies on to operate.

Among the suggestions outlined by the UK government are clear calls to reduce reliance on passwords, help users cope with password overload and manage shared access. The challenge for British security leaders is no longer whether this guidance is sound, but how to operationalize it effectively across their organizations.

Why the guidance is important and necessary

The NNCSC’s latest guidance is an important step forward because it reframes password management not as a user burden, but as a security control that should be automated, centralized and protected by design.

Recent research shows that nearly one in five organizations still operate without formal credential controls, relying on shared spreadsheets, hard-coded passwords or no management system at all. Against this backdrop, it is unsurprising that weak passwords remain widespread across both consumer and enterprise environments.

The password overload problem

Password overload is a symptom of our digital society. Research suggests the average person has about 250 accounts, with 168 passwords across personal accounts and 87 passwords across business accounts.

There’s a risk, as the NCSC points out, that the requirement for the creation of such a vast quantity of passwords could result in ‘password overload’ and force end users to devise their own coping mechanisms, like password re-use, writing passwords down and predictable passwords.

Another coping mechanism that some users may rely on is using browser-based password managers.

While browser-based password managers offer convenience, they were never designed to support enterprise-grade access controls or governance requirements.. They also introduce operational risk through limited visibility, inconsistent policy enforcement and vendor lock-in.

For individuals, reputable third-party password managers remain one of the simplest safeguards – supporting strong, unique credentials while reducing reliance on memory. At an organizational level, however, password management must be governed and enforced as part of a broader identity strategy.

This satisfies the government’s caution around ‘password overload’ and, ultimately, makes the entire organization safer, with minimal cost or disruption.

Reducing reliance on passwords

Passwords are unlikely to disappear overnight – but their role is steadily diminishing as attackers increasingly exploit them at scale.

The problems around passwords remain the same, but the way credentials are being compromised and exploited is rapidly changing. AI-accelerated cracking, credential stuffing and phishing continue to lower the barrier to compromise, while inconsistent organizational practices continue to provide easy entry points.

The rise of passkeys and passwordless authentication reflects the industry’s shift toward stronger, built-in controls that remove reliance on human behavior at a time when credentials remain the primary target for attackers.

Managed shared access

For UK organizations, managing shared access through Privileged Access Management (PAM) is critical to reducing risk in increasingly complex IT environments. For boards and executive teams, unmanaged privileged access represents both a security exposure and a governance failure.

In the UK, this risk is increasingly amplified by regulatory expectations around accountability, auditability and operational resilience. Shared and privileged accounts remain a prime target for attackers, particularly where credentials are reused, poorly monitored or manually managed across teams.

PAM helps address this by enforcing least-privilege access, securely storing and rotating shared credentials, and providing clear visibility and auditability over who accessed what, when and why.

In the event of a breach, PAM solutions can significantly limit lateral movement by restricting unnecessary privileges and isolating high-risk accounts, helping organizations contain incidents more quickly.

Beyond security, PAM also delivers tangible operational benefits, including reduced credential-related incidents, stronger protection of sensitive data and a lower IT helpdesk burden, making it a foundational control for UK organizations navigating regulatory pressure and an evolving threat landscape.

Identity as a new perimeter

For organizations, the top priority is to treat identity as the new perimeter and implement end-to-end credential lifecycle management.

This means securing every stage of a user’s digital identity, from onboarding and access provisioning to ongoing authentication, privilege changes, and timely deprovisioning when roles change or employees leave.

By managing credentials holistically rather than in silos, organizations can reduce attack surfaces, limit lateral movement, and ensure that access is continuously aligned with real business needs.

In an environment where users, devices, and applications operate far beyond traditional network boundaries, strong identity governance becomes the foundation of effective security.

Weak passwords are not an inevitability. They are the result of inadequate controls, inconsistent policy enforcement and outdated behaviors.

With robust password management, strong privileged access oversight and a zero-trust mindset, organizations can significantly reduce their exposure and, in doing so, weaken one of the most exploited attack vectors facing UK businesses today.

We've featured the best private browser.

]]> https://www.techradar.com/pro/passwords-are-still-a-problem-for-uk-businesses-what-next ppipyz2qJoeoSeNqA6756H Sun, 18 Jan 2026 15:00:00 +0000 Fri, 16 Jan 2026 14:59:48 +0000 <![CDATA[ Responsible innovation: what 2026 should look like for businesses harnessing AI ]]> B2B marketing is entering a new phase, defined not by speed or scale alone, but by responsibility, trust and more integrated ways of working.

As businesses face tighter budgets, longer sales cycles and rising customer expectations, marketers are under pressure to do more with less while maintaining impact and creativity.

With UK business investment in AI set to rise by an average of 40% over the next two years, the challenge is no longer whether to adopt AI, but how to ensure it is implemented responsibly and supports measurable, trusted outcomes.

In 2026, success will depend on how effectively organizations combine responsible AI adoption, strong data integrity and unified channel strategies to deliver measurable results.

Responsible, secure AI starts with people, process and governance

AI is now embedded in everyday marketing and operational workflows, but the focus is shifting from experimentation to responsible discipline.

The opportunity is significant: AI can streamline planning, accelerate content development, and support everything from search engine optimization (SEO) to lead scoring, enabling teams to spend more time on strategic thinking and creativity.

In addition to marketing, AI is transforming IT and device operations through predictive maintenance systems that prevent problems before they happen. This proactive approach is becoming the gold standard for operational performance, cost-efficiency, and overall experience.

Without clear oversight, however, AI can generate new risks as swiftly as it creates value. While organizations are adopting AI-powered security products and solutions, hackers are also getting smarter, adding AI-powered approaches to their arsenal.

Recent research shows that 78% of CISOs say AI-powered cyber-threats are already having a significant impact on their organizations. This clearly shows that - outside of technology - people, processes and governance are fundamental to maintaining a safe AI environment.

It is increasingly important for businesses to take a measured, people-first approach when it comes to harnessing AI processes.

This can be achieved by introducing structured training and establishing clear guidelines and compliance frameworks that empower employees with the data confidence and AI literacy to use it responsibly and securely.

In 2026, as leading organizations look to leverage AI to enhance human judgment rather than replace it, those that prioritize upskilling people, implementing clear processes and maintaining governance will be the ones able to scale AI confidently, safely and sustainably.

Trust and reputation will underpin performance

In the backdrop of continued economic uncertainty, buying behavior continues to shift, with budgets being scrutinized, decisions taking longer and expectations around transparency rising.

One theme has become increasingly clear this year: trust and reputation are critical performance drivers. In the UK alone, recent research shows that 67% of consumers consider brand trust essential to their purchase decisions.

As organizations adopt more automated and AI-supported processes, the real differentiator will be the quality, traceability and integrity of the data behind them.

In a landscape where misinformation, content saturation and duplicated AI outputs are rising, businesses can no longer rely on speed alone, they must demonstrate that what they deliver is accurate, authentic and accountable.

In practice, this means prioritizing data integrity and security, content accuracy and transparent guardrails across every stage of the marketing workflow. These capabilities directly influence how customers perceive credibility and risk, and ultimately whether they choose to engage.

A unified approach across channels will be essential for growth

The pressures of 2025, including tighter budgets, fewer events and slower deal cycles, have made siloed working unsustainable, amplified by the increasingly blurred line between B2B and B2C expectations, as businesses enjoy the same intuitive digital experiences they do as customers. Today, success depends on collaboration.

Marketing, sales and product teams must work together to build integrated customer experiences across both direct and indirect partner channels.

A growing shift toward a more balanced direct–indirect model mirrors a broader trend across the industry, where organizations rely on consistent messaging, shared data foundations and coordinated planning to stay competitive.

This shift is not only operational but cultural. Teams must take a 360-degree view of customer needs and prioritize what delivers commercial impact, rather than what fits legacy structures. Approaching channel alignment as a strategic opportunity will be crucial to navigating the realities of 2026.

Looking ahead

The future of B2B marketing will not be defined by a single technological breakthrough, but by a series of deliberate and responsible choices across all business sectors, including adopting AI with governance, strengthening data integrity, integrating channels and investing in people.

As the technology landscape becomes more complex, the ability to remain adaptable, transparent and customer-focused will set organizations apart.

2026 will demand a new kind of marketer: curious, commercially minded and confident in navigating data-driven decision-making and evolving governance standards. Those who embrace this shift will drive performance and help shape a more connected, trusted and resilient future for B2B marketing.

We've featured the best business intelligence platform.

]]> https://www.techradar.com/pro/responsible-innovation-what-2026-should-look-like-for-businesses-harnessing-ai DZM4mxFkh4bLUrw6NMRWpf Sun, 18 Jan 2026 09:00:00 +0000 Thu, 15 Jan 2026 15:13:22 +0000 <![CDATA[ The future of chips: how gate-all-around design is powering the AI era and the next node ]]> Artificial intelligence (AI) is transforming how the world builds and uses computer chips. From massive data centers to devices at the edge of the network, AI requires chips continuing to get faster, smaller and more energy-efficient.

For decades, Moore’s Law, which accurately predicted that the number of transistors on a computer chip would double around every two years, kept the industry moving forward.

Unfortunately, conventional scaling approaches through planar complementary metal-oxide-semiconductor (CMOS) devices, or FinFETs, which extended Moore’s Law, have reached their limits. The industry is now facing a critical challenge: trying to keep advancing chip technology when the old rule of thumb no longer applies.

The industry’s answer is gate-all-around (GAA). This design wraps the gate material completely around all sides, including the control part of the chip, known as the gate, that carries the electric current.

This gives engineers more precise control over how electricity flows through the chip, enabling GAA devices to perform better even as the industry carries onto the next node. This also allows more power without taking up more area.

The new bottlenecks

But GAA isn’t perfect. While it solves challenges via lower power consumption and more efficient use of space, it moves the bottleneck to other areas.

Specifically, older chips encounter resistance — anything that slows down electricity — from inside the channel. Most of the resistance comes from the contact points and areas where the current enters and exits.

To fix this, engineers have added materials called dopants to help electricity flow better. But during this doping process, the dopants can inadvertently spread to nearby places on the chip that should be undoped.

When this happens, it can not only affect performance but create additional issues in the chip such as increased leakage, change the threshold voltage or introduce variability.

Another challenge stems from the manufacturing process, specifically when silicon-germanium (SiGe) layers are removed to shape parts of the chip.

This can leave behind rough surfaces and interfere with how smoothly electricity flows through the devices. Later, when metal contacts are added on top, this creates further resistance at the point where the metal and silicon meet.

In short, GAA may address electrostatics challenges, but it also introduces new ones. This is where advanced materials come into play.

Atomic-scale materials, big fixes

To tackle these new challenges, chipmakers are turning to advanced materials and working at the atomic level to help realize the full potential of GAA.

Here’s how these new materials help:

Blocking unwanted dopant diffusion: Inserting an advanced barrier between heavily doped and undoped areas can prevent dopants from seeping into other areas of the chip. This containment is essential for boosting performance.
Smoothing surfaces: Rough surfaces at the atomic level can scatter electrons and slow them down. Advanced materials engineering can smooth out surfaces that may become uneven during the removal of sacrificial SiGe layers, reducing this scattering. This can boost carrier mobility under normal operating conditions, resulting in more current, faster switching and better performance, all without requiring more power.
Boosting power without compromising size: Advanced materials allow engineers to pack thinner performance structures into the same space. This change can boost current per footprint by about 10% without increasing chip size.
Reducing contact resistance: As device dimensions get smaller, the electrical contact resistance at the point where metal connects to silicon becomes a major limiting factor. By modifying the materials at these junctions, engineers can significantly lower resistance and unlock greater efficiency.

Looking ahead

The explosive growth of AI is driving a fundamental shift in how the industry thinks about computing efficiency. Engineers now face an increasingly complex trade-off between power, performance, area and cost (PPAC).

In the past, the industry has relied on incremental improvements to stay on track, but with the scale and intensity of AI pushing existing architectures to their limits, these small gains are no longer sufficient.

To unlock the next wave of progress, the industry needs a more transformational shift, one that resets the foundation and enables further improvements to continue delivering impact.

The next node, also referred to as the angstrom era, will accelerate innovations in advanced materials beyond what’s possible today to deliver advancements across the PPAC equation. GAA is just the beginning.

To keep shrinking and improving chips for massive AI systems, engineers are exploring new ways to deliver more with less. These breakthroughs in advanced materials are empowering the industry to achieve more performance with less space and energy, driving smarter and more sustainable computing across the board.

Beyond GAA, the industry is already doing work on a new structure called a CFET, or complementary FET, which may take us a few more generations before we start looking at 3D structures like stacked CFETs to keep Moore’s Law progressing.

One thing is certain is that new advanced materials will be required at each step of the way to unleash the performance that these new transistor structures are designed to enable.

We've featured the best green web hosting.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here : https://www.techradar.com/news/submit-your-story-to-techradar-pro

]]> https://www.techradar.com/pro/the-future-of-chips-how-gate-all-around-design-is-powering-the-ai-era-and-the-next-node 6NoZS44vZA3LT9oU2dTTEk Sat, 17 Jan 2026 14:00:00 +0000 Tue, 13 Jan 2026 15:18:50 +0000 <![CDATA[ Data sovereignty: an existential issue for nations and enterprises ]]> Data has long been recognized as an organization's most valuable asset, arguably more important than physical infrastructure or even brand. This is reflected by intangible corporate assets, primarily data including R&D and intellectual property, exceeding $60 trillion in value in 2024.

When used effectively, data unlocks competitive advantage, new markets, better decisions, and helps deliver transformative customer experiences.

Given how critical data is to the day-to-day operations of modern businesses, it needs to be managed, and safeguarded, more than ever. As global geopolitical uncertainty persists, the topic of data sovereignty has become top of mind for governments, regulators, and businesses.

Data residency, data sovereignty

Defined as the principle that data is subject to the laws and governance structures of the country in which it is collected or stored, data sovereignty concerns who has the authority to dictate how data is managed, accessed, and used, particularly in an increasingly interconnected and data-driven world.

For a long time, companies believed data sovereignty simply meant where their data resided, but amid geopolitical shifts and AI’s impacts, organizations now need to distinguish between data residency – where data is physically stored, and data sovereignty – who has legal jurisdiction over that data.

Data sovereignty risks; a perfect storm

Today, new risk factors are reshaping the data sovereignty landscape and pose new questions over access to and use of business-critical data.

Geopolitical conflicts, emerging regulations, international competition and the desire for tighter control of data to power innovation, are forcing company leaders to reconsider their business-critical data’s location, who has authority over it, and how this impacts operations.

Until recently, the idea that an organization's digital operations or services could be interrupted by a third-party ‘kill switch’ would have seemed impossible. However, conditions now exist for governments or global businesses’ core operations being interrupted or revoked without warning via foreign laws or regulations.

Examining three factors in particular shows that service disruption or outages are no longer just hypothetical.

Geopolitical tensions

As conflicts between countries and economic sanctions increase, nation-states are restricting the flow of goods, services and data, trade, collaboration and free information exchange. OECD/WTO research estimates that disruptions to cross-border data exchange alone could reduce global GDP by 4.5%.

Today’s uncertain geopolitical landscape has introduced a heightened risk of service disruption for organizations that depend on services from non-domestic providers—stressing the importance of considering where data is located and managed and where services originate

Regulatory pressure

Law-making bodies have in recent years sought to regulate data flows to strengthen their citizens’ rights – for example, the EU bolstering individual citizens’ privacy through the General Data Protection Regulation (GDPR). This kind of legislation has redefined companies’ scope for storing and processing personal data.

By raising the compliance bar, such measures are already reshaping C-level investment decisions around cloud strategy, AI adoption and third-party access to their corporate data.

Critical infrastructure

Changes in individual governments’ policies are causing uncertainty for cross-border data governance, cloud access and international regulatory harmonization.

Across all regions, organizations are seeking greater control, visibility, and jurisdictional alignment in their data infrastructure – not just for compliance, but for achieving business objectives, operational resilience, and maintaining trust.

Many enterprises are re-evaluating their supply chain and infrastructure locations, vendor jurisdiction, and legal risks, especially when operating in heavily regulated sectors such as healthcare.

Leaders rethink risk

New research commissioned from the University of Technology Sydney (UTS) examined enterprise leaders’ views of the changing landscape. It shows how data sovereignty has moved from a background compliance requirement to a board-level priority.

There was universal agreement (100% of respondents) that sovereignty concerns, such as service interruption, have forced their organization to review where data is located. More than nine out of ten (92%) said geopolitical changes have increased the risk of enterprises failing to fully address data sovereignty questions.

Company leaders fear their data sovereignty could be compromised: 92% fear reputational damage, and 85% fear they could ultimately lose customer trust.

Faced by anything from potential service outages to existential threats to their business, leaders have acted: 78% are embedding sovereignty in core processes, migrating from multiple service providers to investing in sovereign data centers, and putting governance clauses in contracts.

Containing data sovereignty risks

Faced with dynamic data sovereignty risks, enterprises have three main approaches ahead of them:

First, they can take an intentional risk assessment approach. They can define a data strategy addressing urgent priorities, determining what data should go where and how it should be managed - based on key metrics such as data sensitivity, the nature of personal data, downstream impacts, and the potential for identification.

Such a forward-looking approach will, however, require a clear vision and detailed planning.

Alternatively, the enterprise could be more reactive and detach entirely from its non-domestic public cloud service providers. This is riskier, given the likely loss of access to innovation and, worse, the financial fallout that could undermine their pursuit of key business objectives.

Lastly, leaders may choose to do nothing and hope that none of these risks directly affects them. This is the highest-risk option, leaving no protection from potentially devastating financial and reputational consequences of an ineffective data sovereignty strategy.

Ensuring data sovereignty

Given today’s converging geopolitical, regulatory and operational risk factors, company leaders have quickly grasped that data sovereignty no longer equates to data residency; it is a more complex principle, encompassing legal authority over data, how it is accessed or shared, and whose jurisdiction it falls under.

True data sovereignty goes beyond physical location to include operational control, governance, and an organization having full authority over its complete digital ecosystem.

Forward-looking companies can successfully navigate data sovereignty challenges by implementing data strategies that define what data should go where while managing all relevant infrastructure, partner, supply chain and regulatory risks.

We've featured the best data visualization tool.

]]> https://www.techradar.com/pro/data-sovereignty-an-existential-issue-for-nations-and-enterprises eKttzH2UvuXRDdJH63jJKo Sat, 17 Jan 2026 09:00:00 +0000 Mon, 12 Jan 2026 15:08:23 +0000 <![CDATA[ Goodbye cheap OLED TVs — you had a good run, but RGB mini-LED and ‘wallpaper’ OLEDs will soon make you irrelevant ]]> The recent CES 2026 was, as usual, a showcase for the best OLED TVs, with LG and Samsung trotting out new flagship models with brighter screens, sleeker designs, and all manner of AI-based features to enhance both picture quality and the overall user experience.

At this year’s show, however, there was an even greater sense of urgency to the OLED TV unveilings I attended. And that may be because OLED TVs, which consistently earn our highest praise year after year, are now under threat from a formidable new contender for the best TV crown: RGB mini-LED.

TechRadar provided extensive coverage of RGB mini-LED tech and its variants both before and during CES. A key takeaway from the show was that RGB mini-LED will be available in smaller sizes in 2026, with 55- and 65-inch screens featured alongside the monster-sized models slated to arrive from Samsung, Hisense, TCL, LG and potentially other TV brands.

The first RGB mini-LED TVs to be released in 2025, such as the Hisense 116UX and Samsung 115-inch Micro RGB, were extremely large and expensive. We expect the forthcoming smaller-size offerings to be significantly more affordable, though, as examples of first-generation tech, they will likely be more expensive than their regular mini-LED TV counterparts out of the gate.

From what I’ve seen so far of this tech, it’s capable of OLED-level contrast, shadow detail, black uniformity, and viewing angles – ie, all the things that OLED used to beat the competition on, year after year.

Many new models also have a claimed 100% BT.2020 color space coverage, which is a benchmark that OLED doesn’t match. (The LG G5 topped out at 81.3% when we measured it, and the Samsung S95F at 89.3%.)

The new RGB mini-LED TVs slated to arrive in 2025 are likely to be priced higher than the same-size OLED models, even the flagship ones such as the LG G6 and Samsung S95H. But if history is any indication, those prices will quickly fall, while OLED prices will keep stalling, and the real competition with OLED will then kick in, and I don't expect the tide to turn in OLED's favor.

A tale of two OLEDs

LG OLED evo G6 showing image of impressionist painting — The LG OLED evo G6 Wallpaper TV on display at CES 2026 (Image credit: Future)

Another CES TV takeaway was that the OLED TV category is not just facing an existential threat, but is starting to split into two distinct camps: luxury OLED and everyday OLED.

The main example of the luxury trend on the CES show floor was the new LG OLED evo W6 Wallpaper TV. This series features a 9mm, pencil-thin Primary RGB Tandem 2.0 display panel – the same one used in the company’s 2026 flagship G6 OLED TV – that’s wirelessly tethered to LG’s latest-gen Zero Connect Box, an external connection box that supports lossless 4K 165Hz transmission.

@techradar
♬ original sound - TechRadar

LG hasn’t released pricing information yet for the W6, but we expect it to arrive at a premium over the G6 OLED series. It was the star attraction of LG’s CES tech briefings, and appears to be a main focus for the company, which touted its innovative design, flush-mount installation features, and also its reflection-free screen and ability to display artworks from LG’s Gallery+, a subscription-based platform that lets viewers select from a library of over 4,500 still and motion images to display in ambient mode.

In Samsung’s off-site CES display, the new flagship Samsung S95H OLED TV also impressed with its bright QD-OLED display panel featuring a Glare Free OLED screen coating that thoroughly eliminated reflections. Samsung says that the S95H is 35% brighter than last year’s Samsung S95F and features refined tone mapping that lets it process HDR sources mastered at up to 4,000 nits with ease.

We always expect to see year-over-year improvements such as these in Samsung’s flagship OLED TVs, but the most notable changes were a new metal ‘art frame’ that gives it an elegant floating effect when wall-mounted, and support for the Art Mode found in the company’s The Frame TVs. The S95H features built-in connections, but can also optionally be paired with the company’s Wireless One Connect Box for a cable-free installation, even to have up to eight HDMI ports.

Samsung S95H on display at CES 2026 showing image of a black cat — The Samsung S95H features a new metal frame to give the TV a 'gallery' effect (Image credit: Future)

Compared to these very design-forward TVs, LG and Samsung’s more affordable, mid-range OLED offerings at CES looked comparatively bland. For 2026, LG has opted to split its trusty C-series lineup into two, with the larger 77- and 83-inch models getting the same Primary RGB Tandem 2.0 display panel found in the flagship G6 and W6 series, while the smaller screen sizes will carry over the same type of OLED panel used in last year’s LG C5.

With the ‘better’ C6 version (LG calls it the C6H) limited to large screen sizes – and potentially priced higher due to the upgraded OLED panel – the smaller, more mainstream models in the 2026 C-series don’t appear to have advanced much over last year’s LG C5, which in itself represented only an incremental improvement over the previous year’s LG C4.

And then there’s LG’s budget B-series OLED models. A new B6 series wasn’t revealed at the show, but LG confirmed it's coming – and given the direction the C6 series is taking, I don’t expect it to be much different than last year’s LG B5 series, which again was only an incremental improvement on the previous year’s LG B4 series.

As for Samsung, the company’s new S90H OLED series was announced at the show. Similar to last year’s Samsung S95F, we expect only the 65-inch model to feature a higher-performance QD-OLED display panel. The full S90H series does gain the Glare Free screen tech used in Samsung’s flagship TVs, however, which will make it a more enticing option for bright room viewing.

Again, we know a budget-friendly Samsung S85H OLED model is coming, but it looks set to be extremely similar to the Samsung S85F from 2025.

Cheaper OLED TVs: Winter is coming

LG B5 OLED TV with Battlefield V and game optimizer menu on screen — LG's B-series OLED TVs have significantly lower brightness than the company's flagship TVs, but offer a full range of gaming features (Image credit: Future)

So, the story for OLED TVs in 2026 is essentially that flagship models are brighter, more beautiful, and more packed with innovative features than ever.

The midrange and budget models, in contrast, will be basically the same versions we’ve seen over the past few years, with only minor enhancements – and yet the world of competitor TV tech is moving fast around them

If RGB mini-LED TVs (or other variants such as TCL’s SQD mini-LED) end up priced competitively with the more affordable OLED TVs, while offering higher brightness and otherwise equivalent or even better performance, the market for even mid-range OLED, let alone budget models like the LG B-series and Samsung S85-series, could quickly evaporate.

I’d fully expect that fancy, design-savvy, wireless OLED TVs with much higher than average brightness and effective anti-glare screen coatings will continue on well into the future.

As for cheaper, basic OLED TVs… it was nice knowing ya.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/televisions/goodbye-cheap-oled-tvs-you-had-a-good-run-but-rgb-mini-led-and-wallpaper-oleds-will-soon-make-you-irrelevant fueC9ULzRtBgJtbw9ESUgE Fri, 16 Jan 2026 18:00:00 +0000 Mon, 19 Jan 2026 06:53:10 +0000 <![CDATA[ The key to the UK's AI success lies in closing the skills gap ]]> The UK’s ambition to become a global AI superpower is being stifled by a widening skills gap, undermining its potential in research and innovation. Recent data revealed that 73% of UK workers have had no formal AI training, despite two-thirds of workers, using the technology daily.

This disconnect between training and use highlights an imbalance that is causing friction when it comes to deploying AI across businesses. Currently, a minimal 1% of businesses leaders believe their organizations have reached true AI maturity.

Closing the gap will require coordinated action from the UK government and businesses.

The planned £187 million investment in a national skills program to bring digital and AI learning into classrooms is a strong foundation, but lasting results will depend on pairing education initiatives with support for businesses to both recruit specialist AI talent and upskill their existing workforce.

Installing AI confidence in the future workforce

At present, the UK lacks the institutional structure within schools and universities for AI training and education, despite the soaring usage of the technology.

A recent study indicated that over half of UK students want more clarity from their schools and teachers on when and how they should be using AI tools for their schoolwork. Students risk entering the workforce without confidence or the correct understanding of AI tools if action to provide structured guidance is not taken.

The government must create a long-term plan to futureproof the workforce; this includes embedding AI education into school and university curriculums. In classrooms, AI tools like ChatGPT should be embraced, not feared, with students being taught how to effectively and appropriately use the tools.

Through early exposure, students will gain the skills that will be required by future employers, ensure the skills gap becomes narrower, and the talent pool grows.

Balancing hiring and training to avoid internal siloes

Demand for specialised AI talent has driven the most severe tech skills shortage in over 15 years. An immediate step businesses can take towards remedying this dilemma is to invest in upskilling their existing workforce, not just hiring external talent.

Hiring specialist AI talent is important for consultancy and the technical aspects of deployment. However, relying exclusively on external hires and neglecting the upskilling of current employees risks the creation of internal knowledge siloes that prevent successful usage of the technology on a business wide scale.

Businesses must create tailored and accessible training programs, that evolve along with its usage of AI and the different tools and systems it may adopt, if they want upskilling to be worthwhile. Ultimately, this will serve to ensure AI is used to its full potential, across business operations, delivering a maximum ROI.

AI won’t succeed without short, pilot phases

Many businesses are moving quickly to roll out multiple AI models and are often investing significant funds before properly evaluating how well they fit with existing operations.

Rushing into deployment is a reason why businesses struggle with AI adoption as it can lead to poor integration, misuse, and underwhelming outcomes. In fact, 95% of enterprise AI projects fail because large-scale deployments are pushed through without sufficient groundwork or pilot testing.

As organizations look to embed AI, whether it be large-scale agentic systems or LLMs, into existing workflows and business processes, leaders must take a considered, deliberate approach that ensures long-term success rather than chasing the latest tools on the market

This is why short, pilot phases are critical for successful AI integration. Trialing a new AI tool through a focused project gives businesses an opportunity to test performance, address issues, and prepare employees before committing to a full-scale roll out.

Short pilot phases also play a crucial role in building employee confidence, which is essential for adoption at scale. Reskilling is most effective when workers feel supported to learn, which pilot phases are well placed to enable.

With 1 in 4 workers worried that greater use of AI in the workplace will lead to job loss, often driven by lack of knowledge and support, businesses must work to address these concerns. Pilot phases help to do this by providing structured support and actively involving employees in AI implementation through feedback and participation.

Closing the gap benefits everyone

The UK government’s AI action plan is a strong foundation for advancing the UK’s global AI ambitions, however, strategic execution will be the deciding factor in whether the AI skills gap is narrowed and adoption accelerates.

By approaching AI education as a long-term commitment, the UK can develop lasting capability, realizing meaningful returns, and remain competitive in the global landscape.

We've featured the best online courses and online class sites.

]]> https://www.techradar.com/pro/the-key-to-the-uks-ai-success-lies-in-closing-the-skills-gap vYAEbBzrUW6DWC8wfXb2MH Fri, 16 Jan 2026 15:16:16 +0000 Fri, 16 Jan 2026 15:16:15 +0000 <![CDATA[ "It does look like a giant donut": IKEA's latest Varmblixt smart light is a viral sensation, but it's not the only show-stopping product coming in 2026 ]]> As the dust settles on the chaos that is always CES 2026 and we take our time to chew on the tech we saw from the 2026 iteration of the Las Vegas tech show, I can’t help but feel like IKEA stole the show.

We discussed the coolest tech we saw on the latest episode of our podcast, and while awesome tech that feels ripped from the future – like the brain-reading Neurable HyperX headset and over-100-inch Samsung Micro-RGB TV – were among our favorites, we all conceded that IKEA’s simple yet effective smart appliances were the true stars.

Watch our latest podcast below, and make sure to subscribe to our YouTube channel.

Speakers, lights, even air-fryer-microwave-combos, IKEA had a lot of fantastic home gadgetry to showcase – much of which came in dazzling colors, and of course, the brand’s iconic fun Swedish names.

The reason for IKEA’s excellence is a simple one: its tech hit the sweet spot of interesting, useful, and relatively affordable.

Its Varmblixt LED Lamps are beautiful, brightly colored donuts that boast a color representation that even some supposed smart lighting experts can't beat, and they’ll only cost $99 (prices will vary outside the US) when it ships in April.

Our US Editor-at-large saw the light in person, and said that while it "does look like a giant donut," there's "just something so beautiful about it." You can hear more about his first-hand experience with Varmblixt in the episode linked above.

IKEA Varmblixt smart lights — (Image credit: IKEA)

Or you’ve got the tiny, beautiful Kallsup speakers, which are properly affordable at just $9.99. You can connect to up to 100 at once, and there’s a fantastic opportunity here for a delightful audio setup if you’re brimming with creativity.

While it’s arriving a little after CES proper, we’ve also got the three-in-one Gåtebo microwave making its global rollout following its European launch last year. Perfect for smaller spaces, this microwave meets airfryer looks to be a simple yet effective appliance, adding to IKEA’s improved smart home image

We’ll have to see how well IKEA’s tech performs when it lands in front of our expert reviewers, but if it hits the mark with everything it has shown off so far this year, IKEA might have already won 2026.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/home/smart-home/it-does-look-like-a-giant-donut-ikeas-latest-varmblixt-smart-light-is-a-viral-sensation-but-its-not-the-only-show-stopping-product-coming-in-2026 Rk5hd9VyT2Rc3ssjsdjqRA Fri, 16 Jan 2026 15:15:00 +0000 Fri, 16 Jan 2026 13:38:30 +0000 <![CDATA[ This hidden vacuum feature could change the way you clean your home ]]> A vacuum cleaner is a pretty straightforward concept. You switch it on, it sucks up dust and dirt, and then you switch it off again. But after spending my days testing most of the best vacuum cleaners on the market, there's one useful feature I've found can make a big difference to the cleaning experience – and I'd wager that most people aren't aware of it.

Here it is: an air intake adjustor on the floorhead. This usually takes the form of a small manual sider, and it does exactly as the name suggests: it changes the amount of air that can get into the main suction area underneath.

Vacuums rely on there being a 'vacuum' (in the original sense of the word) underneath the floorhead. The head needs to connect to the floor around its entire perimeter in order to maximize the cleaning power – if it were open, the motor would be wasting energy sucking air from the surrounding area rather than being concentrated on its main task of pulling dust from the floor beneath it.

Dyson Digital Motorbar floorhead — Vacuum cleaner heads rely on creating somewhat of a seal around the roller, to maximize pickup power (Image credit: Future)

However, if there's too much of a vacuum created underneath the floorhead, it can become impossible to push. That can be a particular issue on thick carpet, where the long fibers help form more of a seal around the floorhead. That's where the air intake slider comes in.

The trick is to find the setting that makes maneuvering comfortable while still creating enough of a seal for efficient cleaning. So you'll probably want to open up the sider just a little bit on thick carpet, to can make the vacuum easier to push (when I was running my group test to find the best cordless vacuum, I found on the models that had the most power, I often had to open the slider when cleaning carpet in order to make maneuvering possible). Conversely, on a hard floor, you might want to close it down to maximize pickup power.

Vacuum cleaner floorhead with air intake slider, cleaning a hard floor — On a hard floor, it might be worth closing off the slider to create more of a seal (Image credit: Future)

The good news is that this isn't a particularly advanced feature – a large proportion of vacuums have it already, you just might not have noticed it. If you're struggling with maneuvering on carpet, or want to maximize your vacuum's cleaning power on hard flooring, it's well worth checking around on your floorhead to see if there's a slider you might have missed.

]]> https://www.techradar.com/home/vacuums/this-hidden-vacuum-feature-could-change-the-way-you-clean-your-home x9zoCLuA57Qu3ZrZAqhuqQ Fri, 16 Jan 2026 12:52:11 +0000 Fri, 16 Jan 2026 12:52:11 +0000 <![CDATA[ Matthew McConaughey trademarking himself saying 'alright, alright, alright' is a preview of Hollywood’s coming AI identity crisis ]]> Matthew McConaughey has never been shy about turning himself into a brand, and he's been happy to lean into the many memes based on his performances. His latest move is to trademark his own voice saying “alright, alright

alright” and video clips of himself standing on a porch, as first shared by The Wall Street Journal. That might seem odd until you realize it's not about a new product, it's a kind of legal preemptive defense against the onrushing AI celebrity cloning wars.

Filing a few trademarks now might scare off a few rogue YouTubers using AI to imitate his catchphrases. But McConaughey and other celebrities are uneasy for a reason. Movie stars, singers, and famous people of all stripes see the viral fake Drake songs that sound real enough to fool Spotify, or the videos showing Tom Cruise in movie roles he never booked. And they know those aren't made with approval or compensation of the humans they're based on.

What McConaughey is really doing is preemptively staking a legal claim on the parts of himself that AI could soon reproduce at scale without his involvement, permission, or benefit. That iconic three-word drawl and slow porch stare are now registered and protected. But putting up a fence around your digital likeness will likely prove more complicated than filing a few applications.

Fractured protection

Almost no legal framework has caught up to the age of easy deepfakes. Some state-level “right of publicity” laws offer celebrities protections from unauthorized commercial use of their name, face, or voice, but it's hardly universal, and the rules vary wildly internationally. Trademarks like McConaughey’s make sense at first glance, but are actually a poor fit. Trademarks are best at protecting logos, slogans, and distinctive commercial identifiers. They’re not built for filtering deepfakes from fan art or fair use satire from full-on digital impersonation.

McConaughey's lawyer admitted as much to the WSJ, saying outright that they don’t yet know how the courts would rule if someone pushed back. This is preemptive maneuvering, a legal scarecrow planted in a field full of fast-growing weeds. On the other hand, no one knows where the real lines are yet.

An actor might successfully stop a brand from deepfaking them into an ad. But what about the open-source model trained on hundreds of YouTube interviews? What about an indie filmmaker using AI to recreate McConaughey’s voice for parody? Or an influencer building a synthetic “romance sim” using the voiceprints of multiple celebrities? The law doesn’t have a clear answer, and that’s not an oversight. It’s a symptom of how quickly the tech is moving.

If you’re not a celebrity, it’s easy to assume this is a problem for richer, more recognizable people to worry about. But synthetic identity isn’t just coming for the famous. AI tools are already capable of scraping publicly available audio and video data to build custom voice models. You don’t need a famous catchphrase to become part of a training dataset.

But most of us don’t have lawyers on retainer or the resources to monitor the internet for unauthorized AI clones. So when AI begins generating synthetic people at scale, the power imbalance between those who can control their digital likeness and those who cannot is likely to deepen.

Hollywood arms race

For Hollywood, this isn’t theoretical. Actors fought hard in the 2023 SAG-AFTRA strike to win protections against AI replication. That included clauses about informed consent, compensation, and limits on how synthetic versions of themselves could be used.

That's why SAG-AFTRA President Sean Astin made a trip to CES this year to engage with AI companies and try to better understand and position his organization to best serve its members amid the chaotic changes.

"There's a moment at which, well, I see things all of civilization needs to be paying attention to it," Astin told TechRadar's Lance Ulanoff. "But my focus as a labor leader is to figure out the point at which it crosses over as an entertainment or as a business...and moves into our area where it affects our members and their name image and likeness and their – the things where we have jurisdiction over covering and so forth."

The negotiations haven't stopped some studios from experimenting with AI actors. There are digital doubles that perform in scenes the real actors never shot. There are “de-aging” projects using AI to make older stars appear decades younger. And then there are the AI tools capable of generating entirely synthetic performances without a camera ever rolling. Trying to keep up on the labor side when the technical aspect moves so quickly is tough.

“We knew, when we were achieving certain language in our last agreement, which was foundational, new, novel stuff about different guardrails and things, that…it would be out of date,” Astin said.

Fake faces and real consequences

There’s a deeper shift underway here, one that trademark law can’t resolve. As synthetic media becomes easier and more convincing, the question of ownership becomes not just legal but cultural. If people expect to be able to remix and regenerate anything, then the law alone won’t be enough to stop them. There will need to be new norms, new taboos, and new expectations around consent.

McConaughey’s line is clear: if you want to use his voice, ask him. That shouldn’t be controversial. Consent and attribution are low bars, and yet they’re absent in much of today’s AI landscape. Most AI voice tools don’t tell users where the source material came from. M

In some ways, McConaughey’s actions set a precedent. If a famous phrase or moment can be legally protected, maybe yours can too. If not through trademark, then through pressure on platforms to flag synthetic content, on lawmakers to draft modern regulations, and on AI developers to build with consent in mind. The truth is, most people won’t have the means to file lawsuits every time their face appears in an unauthorized AI video. But maybe they shouldn’t have to.

We need a broader shift in how synthetic identity is treated, including penalties for violating consent. We’re all walking toward an uncertain era where the most persuasive versions of ourselves might not even be ours.

This era demands new legal constructs, regulatory clarity, and international cooperation to govern how AI can use and reuse personal identity. Without that, celebrities might end up fighting a series of narrow battles without winning the larger war. And even someone who just likes to upload videos of themselves telling stories could find their voice snatched away without their permission to sell a product they've never heard of. And that's not alright, alright, alright.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/ai-platforms-assistants/matthew-mcconaughey-trademarking-himself-saying-alright-alright-alright-is-a-preview-of-hollywoods-coming-ai-identity-crisis i6dcXKq3PQYz4BVwsgQu2S Fri, 16 Jan 2026 12:46:05 +0000 Fri, 16 Jan 2026 12:47:34 +0000 <![CDATA[ Powering the AI data center boom: the infrastructure upgrades behind innovation ]]> As hyperscalers and other data center operators accelerate efforts to deliver the capacity required for generative AI and large-scale model training, modernizing data centers for the latest GPU technologies has become a defining challenge.

The shift to highly advanced accelerators demands radical improvements across power, cooling, and high-speed connectivity. What was once considered cutting-edge even a few years ago is no longer enough to support today’s AI workloads, forcing operators to rethink everything from rack design to thermal strategy.

The scale of global investment reflects this pivot. McKinsey estimates that data center spending will reach $6.7 trillion by 2030, with the majority funneled into facilities engineered specifically for AI.

Yet the industry’s rapid expansion is already confronting major constraints: supply chain bottlenecks for GPUs and interconnects, design limitations that restrict density, and shortages of skilled engineers capable of supporting complex builds.

These pressures have helped fuel the rise of “neocloud” providers whose business models revolve entirely around high-performance GPU compute.

A neocloud surge

Neoclouds have become one of the most dynamic forces reshaping data center infrastructure. Unlike traditional operators, which must balance AI capacity against broader cloud demands, these firms design everything around GPU acceleration.

With demand for generative AI growing faster than existing data centers can be upgraded, neoclouds are capturing momentum by deploying at extraordinary speed and offering high-performance compute at competitive rates.

The scale and ambition of these projects are unprecedented. CoreWeave, for instance, has rapidly grown from modest deployments to tens of thousands of GPUs per build, complemented by the rollout of NVIDIA’s GB300 NVL72 systems.

The performance gains are dramatic, up to ten times greater responsiveness and significant improvements in energy efficiency versus previous generations. Meanwhile, NScale’s 230-megawatt facility in Norway aims to deliver 100,000 GPUs by 2026, powered entirely by renewable energy.

Nebius underscored the scale of market appetite with a multi-billion-dollar GPU infrastructure agreement with Microsoft, a deal that instantly transformed its market position.

The motivation extends far beyond a technical race for capacity. Nations increasingly view AI infrastructure as a pillar of long-term competitiveness. Countries capable of deploying at speed stand to attract investment and talent. Those who move too slowly risk watching opportunities go elsewhere.

The engineering bottlenecks

Building AI-ready infrastructure exposes the limits of even recently constructed facilities. Power density requirements are rising sharply, while cooling and bandwidth constraints frequently demand wholesale redesigns.

Many operators face the uncomfortable reality that retrofits may be costlier or more disruptive than expected, resulting in delayed projects or cancelled expansions.

The most significant shift is the transition from traditional air-cooled systems to various forms of liquid cooling, particularly direct-to-chip.

These systems enable dense GPU clusters to operate within acceptable thermal limits, but require entirely new facility considerations such as fluid distribution and containment to power integration and safety protocols.

Connectivity presents another critical challenge. AI workloads depend on vast east–west traffic flows between GPUs, pushing interconnect technologies such as InfiniBand and advanced fiber optics to the limits.

Supplies of these components remain constrained globally, while the installation itself requires specialist skills and careful coordination. Dense GPU fabrics are only as strong as the cabling underpinning them; poorly designed or cheap deployments quickly become performance chokepoints.

This surge in complexity is echoed in workforce requirements. AI data center builds routinely require several times the manpower of conventional projects, often involving teams with specialized fiber, power, and cooling expertise.

Coordinating these disciplines while maintaining speed, quality, and safety has become a defining operational challenge of the AI era.

Why the right partners matter

That’s why data center operators are increasingly turning to trusted partners capable of bringing technical depth, global experience, and operational scale. No single operator, regardless of size, can shoulder the full burden of AI projects alone.

Strong partners help bridge gaps across engineering, logistics, compliance, and workforce mobilization, enabling operators to move quickly without sacrificing quality or resilience.

These partners contribute in several critical ways. Their familiarity with high-density cabling architectures, advanced cooling solutions, and GPU cluster integration allows them to design and execute upgrades that match the demands of next-generation AI systems.

They also help navigate local regulatory and permitting environments, mitigating risks that can halt or delay builds. On the operational side, they can mobilize large, skilled teams at pace.

That means sourcing, training, and coordinating engineers while ensuring health and safety, and quality control remain robust under accelerated timelines.

In short, the ability to draw on partners with deep technical capability and agile delivery models can be the difference between an ambitious design on paper and a functioning AI data center ready for commercial workloads.

Winning the infrastructure race

The race to build AI-ready data centers is no longer about deploying the latest GPUs. It is a test of coordination between technology, regulation, labor, and supply chains. Operators that combine strong internal leadership with the right external partnerships will be best positioned to bring capacity online quickly and reliably.

As global demand for GPU compute continues to outstrip supply, those able to deliver advanced infrastructure at speed will secure a decisive competitive advantage. In this new era of hyperscale AI, collaboration and capability will determine who leads.

We've featured the best data migration tool.

]]> https://www.techradar.com/pro/powering-the-ai-data-center-boom-the-infrastructure-upgrades-behind-innovation thARjF5fR3fPYesCQZSJSQ Fri, 16 Jan 2026 11:19:56 +0000 Fri, 16 Jan 2026 11:19:55 +0000 <![CDATA[ CIOs don’t need more AI—they need AI that actually understands their business ]]> Most enterprise AI today still operates in what can best be described as single-player mode.

It helps individuals write emails, summarize documents, or generate answers based on the information immediately available to them.

These tools can deliver real productivity gains, but they address only the simplest use cases.

Enterprises, however, do not operate on individual context alone. They run on institutional context. In other words, the interconnected relationships among customers, systems, metrics, processes, ownership models, policies, and historical decisions that determine how the business actually functions.

As AI moves from email help to business decisions, it starts to break because it doesn’t understand how the company actually works.

So the next phase isn’t adding more copilots. It’s building institutional-mode AI; agents that can reason across the company’s real operating context, not just whatever data happens to be in one place.

But that only works if the basics are in place: clear definitions, ownership, governance, and training so the AI is reliable and safe at scale.

Why AI fails in the enterprise

Single-player AI succeeds because the problems it addresses are narrow and self-contained. Drafting an email or summarizing a report does not require an understanding of how the enterprise operates as a system. Institutional AI does.

To answer meaningful business questions, AI must have access to structured business meaning that many organizations have never fully formalized, including:

Shared semantics that define what metrics actually represent
Data lineage that explains where numbers originate and how they change
Clear ownership models that establish accountability
Governance rules that encode trust, access, and approval
Business logic and exceptions that reflect real decision processes
Historical signals that distinguish normal variation from true anomalies

Without this context, AI does not become more intelligent; it becomes more articulate and less reliable. When asked to explain performance or recommend actions, it fills informational gaps with plausible narratives rather than grounded reasoning.

This dynamic explains why so many enterprise AI initiatives stall after promising pilots. Adoption is not the primary obstacle. Making the business legible to machines is.

The gap appears first in analytics and executive decision-making

The absence of institutional context becomes most apparent when AI is applied to analytics. Writing an email is relatively straightforward. Diagnosing churn, forecasting revenue, or explaining cost variances is not.

Analytics exposes the fragmentation that exists in most enterprises. Metrics reside in multiple systems. Definitions vary by team. Ownership is often implicit. Historical context is incomplete or undocumented. AI trained on isolated datasets can still generate answers—but those answers are fragile, inconsistent, and difficult to trust.

In practice, CIOs encounter familiar patterns:

The same metric carries different meanings across systems
Data ownership is assumed rather than explicitly defined
Historical context needed for interpretation is unavailable
Exceptions and edge cases exist only as tribal knowledge

This is where single-player AI breaks down. Faced with ambiguity, it produces explanations that sound reasonable but lack decision-grade reliability.

As a result, analytics has become the proving ground for enterprise AI, and the first area where its limitations are exposed without institutional context.

How context-aware agents unlock enterprise value

The first wave of enterprise AI has been largely additive, For example, chatbots that explain reports, copilots that summarize meetings, and tools that help individuals work faster. These capabilities are useful, but incremental.

Material enterprise value emerges when AI becomes proactive. This is when it can operate across systems and over time, rather than responding to isolated prompts. Context-aware agents can:

Identify anomalies before they escalate
Explain root causes instead of surface symptoms
Diagnose operational issues across functions
Recommend actions grounded in business reality

This class of AI depends on institutional context: without it, outputs may be interesting but they aren’t dependable enough for production.

And that’s the crux of getting ROI from AI—enterprise value won’t come from better summaries, but from agents that can reason with institutional knowledge and drive real operational decisions.

The future is AI that understands how the business operates

The future of enterprise AI will not be defined by larger models, faster inference, or additional vendors layered into the stack. It will be defined by AI systems that reflect how the business actually works.

That requires treating business semantics, governance, and operational context as first-class infrastructure. Relationships among data, systems, and processes must be explicit.

Governance must be continuous rather than static. AI systems must learn from operational signals over time—what changed, what failed, what was approved, and what succeeded.

With this foundation, AI can function as a credible decision-support partner. Without it, it remains a tool that is impressive in isolation but unreliable in practice.

CIOs do not need more AI. They need AI that understands their business well enough to help run it.

We've featured the best AI website builder.

]]> https://www.techradar.com/pro/cios-dont-need-more-ai-they-need-ai-that-actually-understands-their-business qWNAyDZrfFh7hRPwAKUnYd Thu, 15 Jan 2026 15:19:49 +0000 Thu, 15 Jan 2026 15:19:48 +0000 <![CDATA[ ChatGPT saved me from a roadside nightmare — this is exactly what we should be using AI for, not for replacing creatives ]]> Just imagine the scene. It’s cold, it’s dark, it’s raining, you’re driving in heavy traffic — and suddenly, your car stops accelerating. It will only crawl along at a painfully low speed.

There are no obvious indications that anything is wrong: no flashing warning lights, no strange noises. So you pull over to the side of the road, switch on your hazard lights, and feel your heart sink.

Great, you’re thinking — this means a call to roadside recovery, and who knows how long they’ll take to show up?

That was me this week. Luckily, ChatGPT came to the rescue.

I hadn’t owned my car for very long, and because it was second-hand it didn’t come with a manual. I wasn’t familiar with all the symbols and messages on the dashboard, but I did notice it was proudly displaying “19 mph limit” — with no explanation as to why.

Pushing the limit

So I took a photo of the dashboard, uploaded it to ChatGPT, and asked what it meant. Almost instantly, it replied that my car had a speed limiter feature, and that it was switched on, and set at 19 mph.

I had no idea my car even had this feature! “How had it been turned on?”, I asked. ChatGPT suggested there was probably a button on one of the stalks behind the steering wheel. I had a quick look around, and there it was, clearly labelled 'Limit'. I turned it off, and just like that, my problem was solved.

As for how it was activated in the first place? My thumb or fingers must have brushed against the button while I was driving, without me realizing.

If I’d waited an hour for roadside recovery to arrive, only for them to tell me to press a single button, I’d have been absolutely mortified. Now that I know this feature exists, I’ll recognize the issue instantly if it happens again, and I know how to change the settings properly.

On the road again

I probably could have reached the same conclusion by Googling it on my phone, but the ability to upload a photo of exactly what I was seeing and get a clear, tailored explanation made all the difference.

The response wasn’t just accurate, it was instant. More importantly, it got me safely back on the road without having to make what would have been a deeply embarrassing call for help.

And this, to me, is what AI should be for. Quietly solving small, real-world problems at exactly the moment you need it, not churning out endless content, flooding social feeds with low-effort AI slop, or being bolted onto every product whether it’s useful or not. Used thoughtfully, AI can genuinely make everyday life easier. Used carelessly, it risks becoming yet another noisy, intrusive technology we didn’t ask for.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/ai-platforms-assistants/chatgpt/chatgpt-saved-me-from-a-roadside-nightmare-this-is-exactly-what-we-should-be-using-ai-for-not-for-replacing-creatives 4LUD8tNDYQR9pvWr2ezff Thu, 15 Jan 2026 13:36:53 +0000 Thu, 15 Jan 2026 15:08:16 +0000 <![CDATA[ How people use ChatGPT – analysis of OpenAI’s first user study ]]> On September 15, 2025, OpenAI published its first in-depth research examining how individuals and organizations use ChatGPT.

This paper, How People Use ChatGPT, represents the company’s attempt to document patterns of adoption, evaluate user behaviors, and provide a foundation for policymakers, industry leaders, and educators to understand the role of generative AI in daily life and professional contexts.

The report, produced by a team of economists and social scientists, combines survey data, anonymized usage statistics, and interviews.

While many findings confirm what industry observers might expect—that ChatGPT is used across work, education, and creative fields—the research also uncovers important nuances in frequency, motivation, and impact.

Interpreting the findings

For the broader mobile and digital ecosystem, several implications stand out:

Shift toward professional augmentation – Generative AI is being embedded not as a replacement for workers, but as a complement to their productivity.
Education as a frontier – The educational sector emerges as a proving ground for responsible AI adoption. Teachers, institutions, and students are experimenting at scale, which may shape norms in other industries.
Programming as an anchor use case – The high share of technical use illustrates how AI is reshaping software development, potentially accelerating innovation cycles.
Early signs of healthcare potential – Although still modest, healthcare adoption signals that AI may expand into regulated and sensitive domains sooner than some expect.

The report ultimately underscores that generative AI’s trajectory is not about replacing human work, but about redefining its contours. Where the technology thrives is in taking friction out of routine processes such as drafting, summarizing, coding. It seems to let people use judgment, creativity, and human interaction.

The most profound change may not be in the tasks ChatGPT performs, but in how it reshapes expectations of efficiency and fluency in the digital age.

As adoption widens, the challenge will be to balance productivity gains with ethical and social safeguards. If the first chapter of ChatGPT’s story has been about curiosity and experimentation, the next will be about accountability and trust.

Key patterns of use

OpenAI’s research highlights that ChatGPT is no longer a niche tool. Instead, it has entered mainstream workflows across personal and professional settings. The study classifies usage into several broad categories, capturing how individuals interact with the system in both structured and ad hoc ways.

A third of all reported interactions are work-related. This includes drafting professional documents, summarizing materials, generating presentations, and supporting decision-making.

Education and learning represent the second largest category: students and self-learners use ChatGPT for explanations, practice problems, and language learning. Programming and coding account for a substantial share (21%), reflecting the model’s strength in technical problem-solving.

Interestingly, personal tasks—such as meal planning or travel advice—are a relatively small slice of the overall usage. This suggests that despite media portrayals of “AI assistants” in daily life, the technology’s uptake has been most significant where it complements productivity, not leisure.

Professional adoption by sector

The report breaks down professional use of ChatGPT by industry. Not surprisingly, adoption is strongest in education and technology, where experimentation and digital literacy are high.

Professional services—such as consulting, legal, and financial advisory—also feature prominently, reflecting the demand for tools that streamline analysis and client communication.

In education, nearly half of surveyed professionals report frequent use, often in tasks such as lesson preparation or student engagement. Technology workers use ChatGPT to accelerate coding, debugging, and documentation.

Healthcare, while lower, still shows meaningful uptake: clinicians and administrators experiment with AI for documentation, patient communication, and training.

Broader trends in adoption

Several broader trends emerge from the research:

Generational divides – Younger professionals and students are the most intensive users, but uptake among mid-career professionals is accelerating.
Productivity framing – Users increasingly describe ChatGPT as a “time saver” rather than a novelty. Its role is shifting from experimentation to embedded workflow.
Creative vs. technical balance – While much public attention has focused on AI’s creative abilities, data suggests its strongest pull remains in technical and task-oriented contexts.
Emerging institutional adoption – The report notes that organizations are beginning to formalize guidelines for AI use, marking a transition from individual experimentation to enterprise integration.

Criticism and limitations of the data

While the report is a welcome contribution, several limitations stand out:

Geographic skew – Most data comes from North America and Europe. Adoption patterns in Asia, Africa, and Latin America may differ significantly, yet are underrepresented.
Short time window – The data reflects a relatively early stage of generative AI adoption. Longitudinal studies will be needed to capture sustained usage and long-term effects.
Lack of counterfactuals – The report highlights how people use ChatGPT, but not necessarily what they would do without it. The productivity benefits are assumed, not measured.
Enterprise opacity – Although organizational use is acknowledged, the data is far more robust for individual adoption. The extent of corporate integration remains somewhat speculative.

These caveats are not unusual for a first study, but they underscore the need for more robust work in the following years.

We've featured the best AI website builder.

]]> https://www.techradar.com/pro/how-people-use-chatgpt-analysis-of-openais-first-user-study ocC5xK56jRvUKDdpFDA8JN Thu, 15 Jan 2026 13:27:53 +0000 Thu, 15 Jan 2026 13:27:52 +0000 <![CDATA[ Euphoria season 3 was the last thing I thought I’d be excited for in 2026, but its action-packed trailer has hypnotized me like a moth to a flame ]]> This year’s movie and TV calendar is already packed with must-see titles, but there’s one HBO Max show that I never thought I’d be excited to see return: Of course, I’m talking about Euphoria season 3.

The teen drama series led by the fabulous Zendaya, who plays troubled high schooler Rue, is set to make its highly anticipated return on April 12, which means we’ll finally get to see where her character and array of problematic friends have found themselves five-years on from the events of season two.

Given the third season's delayed rollout and various productions hiccups, my bitter anger towards the show and its creator for making me wait so long between seasons grew substantially, leading me to lose all interest in its upcoming final chapter. But lo and behold, the Euphoria season three trailer drops, and I hate to admit defeat, but it looks bloody great.

I blame you Sam Levinson

Sydney Sweeney as Cassie in the new trailer for Euphoria season three — Sydney Sweeney returns as Cassie in Euphoria season 3. (Image credit: YouTube / HBO Max )

It’s been a wild ride for the HBO Max show to say the least, but its multiple production delays and the seemingly flippant attitude of its creator Sam Levinson have been main driving forces for the grave disappointment in the fanbase – as I mentioned, it’s been difficult to put my own anger aside.

Though I can’t fault some of its production delays – the LA wildfires, writers' strikes, and untimely deaths of Angus Cloud and executive producer Kevin Turen have all had a significant impact – my frustration has been projected mainly towards Levinson. While unavoidable circumstances delayed production, we would’ve had Euphoria season 3 much sooner if Levinson wasn’t allegedly picking fights with the stars and production team.

One of the actors we won’t be seeing in season three is Barbie Ferreira, who departed following alleged conflict with Levinson over her beloved character’s storyline. There are some rumors that link her exit from the show to a toxic work environment while working on Euphoria season 2.

But that’s not all, as even Zendaya reportedly had her share of hostility with Levinson over his rumored undivided focus on producing the limited series The Idol, which turned out to be one of the most god awful pieces of television ever made. In hindsight, this was wasted time that could’ve been better spent on Euphoria season 3.

Euphoria season 3's dramatic trailer has me locked in

Zendaya as Rue in the trailer for Euphoria season three — Rue will face drug dealer Laurie one last time in Euphoria season 3. (Image credit: A24)

I wanted to hate Euphoria season 3 so badly, but since the trailer dropped it’s all I can think about. It confirmed a lot of fan theories, but also spilled a handful of unexpected surprises: Nate (Jacob Elordi) and Cassie’s (Sydney Sweeney) wedding being the shining star given their dramatic breakup in season two. This also raises questions about the whereabouts of Nate’s ex and Cassie’s former best friend Maddie (Alexa Demie), who’s by far the most iconic of the bunch.

But what’s gripped me the most is Rue’s run-in with ruthless drug lord Laurie to whom she’s $10,000 in debt to (a plotline that I’m relieved wasn’t totally forgotten about). Laurie’s five year-long hunt for Rue appears to hint at a violent and life-threatening conclusion that won’t go down smoothly. Rue has found herself in her most vulnerable position yet, but we won’t know if she’ll make it out alive until the final season drops in April.

You can watch the trailer for Euphoria season 3 here:

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/streaming/hbo-max/euphoria-season-3-was-the-last-thing-i-thought-id-be-excited-for-in-2026-but-its-action-packed-trailer-has-hypnotized-me-like-a-moth-to-a-flame WwcsRhwVDbAKS98Y4xciFE Thu, 15 Jan 2026 13:07:28 +0000 Thu, 15 Jan 2026 13:07:28 +0000 <![CDATA[ 'Society cannot function if no one is accountable for AI' — Jaron Lanier, the godfather of virtual reality, discusses how far our empathy should extend to AI in episode two of new podcast, The Ten Reckonings ]]> Whether we like it or not, we can’t ignore AI. What started as a fun, gimmicky chatbot on our desktops, albeit one that could talk a bit like a human, is already taking jobs, accessing medical records, and reshaping workplaces. We are rapidly approaching the point where the practical realities of building and governing advanced AI systems must be confronted.

As the recent furor over indecent Grok-generated images on X, and the use of Meta AI smart glasses to record women without their permission for social media clicks has shown, the guardrails meant to help society cope with the deluge of AI devices and new technologies seem seriously lacking.

Even before the latest controversies around AI-generated images, one of the biggest shocks to me was the way some AI companies decided it was perfectly acceptable to train their models on copyrighted material from authors and artists without permission – and the fact that, despite a few lingering lawsuits, they appear to have faced few consequences so far.

Zero accountability

Society cannot function if no one is accountable for AI
Technologist Jaron Lanier

All of this makes me wonder whether we’re really ready for a world in which AI runs everything with zero accountability. Two people who have been grappling with similar questions are technologist Jaron Lanier and Dr Ben Goertzel, CEO of SingularityNET and founder of the ASI Alliance, in the next upcoming episode of The Ten Reckonings podcast.

“Society cannot function if no one is accountable for AI”, says Lanier, who is often described as the 'godfather of virtual reality'.

This new episode forms part of a series where these issues are explored in depth. According to Goertzel, “The ASI Alliance’s purpose is not to present a unified position, but to create space for the world’s leading thinkers to openly debate and, in doing so, help society reckon with the profound choices ahead.”

Lanier discusses the idea of AI sentience and its implications. He argues: “I don’t care how autonomous your AI is – some human has to be responsible for what it does, or we cannot have a functioning society. All of human society, human experience, and law is based on people being real. If you assign this responsibility to technology, you undo civilization. That is immoral – you absolutely can’t do it.”

Shaping the future

I agree with him. While accelerating toward more autonomous, decentralized AGI could ultimately prove safer and more beneficial than today’s fragmented landscape of proprietary systems with weak guardrails, Lanier’s point about human accountability is exactly right. Right now, AI companies seem to be operating on the assumption that it’s better to beg forgiveness later than ask for permission now, and that approach cannot continue.

And while there appears to be little hope of meaningful AI regulation coming from the US at the moment, the rest of the world may be prepared to step in. The UK regulator Ofcom is launching an investigation into X over Grok, and Indonesia and Malaysia have banned Grok altogether.

At this point we all know that AI is going to shape our future, but the question of responsibility still lingers. Governments are going to have to be willing to step up because if they hesitate then the current lack of accountability edges us into even more dangerous territory. Whether that’s through images, or medical advice, or the protection of our rights. Progress without accountability isn’t innovation, it's recklessness.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/ai-platforms-assistants/society-cannot-function-if-no-one-is-accountable-for-ai-jaron-lanier-the-godfather-of-virtual-reality-discusses-how-far-our-empathy-should-extend-to-ai-in-episode-two-of-new-podcast-the-ten-reckonings SjbhVAyft6NaP7ZeYLzZkk Thu, 15 Jan 2026 10:48:34 +0000 Thu, 15 Jan 2026 10:48:35 +0000 <![CDATA[ Forget the remaster — 2000's Deus Ex is still a masterpiece that you should play in its original form ]]> Towards the end of last year, Deus Ex Remastered got delayed, with no new release date offered and refunds for preorders issued. Developer and remaster specialist Aspyr noted that it had taken on board negative feedback about the remastered graphics, and noted that the game would not be released on February 5, 2026.

Now, given I still consider the original Deus Ex a masterpiece and one of my favourite games of all time, this was a bit of disappointment. I’d rather like to play the game again, but with a coating of modern graphical paint and some quality-of-life updates.

But screenshots and video from Aspyr’s remaster didn't showcase a slick rework of developer Ion Storm’s 2000 original game, but rather something that looked like it was from the mid-2000s, with visuals that would look at home on an original Xbox or the early days of the Xbox 360, rather than for modern hardware in the year 2026.

I’m sure that taking a 26-year-old game and remastering it is no easy feat. But reworking textures that look out of place and in keeping with the dark and stark look of Deus Ex’s original graphics isn’t the way to do it. I’d rather see the game remade in the Unreal 5 engine or the Dawn Engine Deus Ex: Mankind Divided used; getting access to the latter might be a bit of a pipedream given its under the ownership of Square Enix.

But the negative reaction to Deus Ex Remastered, and Aspy’s arguably wobbly record when it comes to remastering games, got me thinking: Do we actually need a remaster of Deus Ex?

Art and ambition over graphics

an screenshot from Deus Ex — (Image credit: Eidos Interactive / Valve)

Even when it launched, Deus Ex was hardly a good-looking game; textures were blocky, and the game was rather dark in a lot of places. But this was part of the art style of the game that was capturing the conspiracy theories vibe of the late 90s and early 2000s, with a sprinkle of Matrix-esque and cyberpunk feel.

Moreover, a lack of cutting-edge fidelity for the time meant that Deus Ex didn't demand very powerful hardware and could have large, expansive maps and hub areas to explore. Even playing it years later, I’d uncovered a secret place, with some solid story implications, that I’d missed the first time around.

And by not prioritising graphical flair, Deus Ex had the mass of immersive sim elements and systems that could be manipulated and exploited to get around all manner of obstacles. The amount of variation and immersion here outpaces many more modern immersive sims.

So part of me thinks that to experience Deus Ex properly, one should play it in as much of its original guise as possible. The Game of the Year Edition makes the most sense, as that has support for modern controllers – though I advocate for playing Deus Ex on mouse and keyboard – and should run on a lot of hardware.

Heck, I managed to get it running reasonably well on my Steam Deck, though there’s a need to use the trackpads to handle some of the mouse cursor bits.

For people who might want early Xbox-era graphics, there's Deus Ex: Invisible War; it’s a lot more streamlined than its predecessor, but I still think it has plenty of systems to exploit and is perhaps more underappreciated than it deserves.

More fidelity, and a simply amazing art style in my opinion, can be had with Deus Ex: Human Revolution. It’s a prequel to Deus Ex but captures a lot of the game’s essence, with some neat systems.

Then Mankind Divided builds upon that, especially with a great hub area in the form of near-future Prague. Mankind Divided ended somewhat abruptly, which didn’t win it masses of fans. But its areas felt more like playgrounds of opportunity with a myriad of routes to take and systems to explore, whereas Human Revolution felt occasionally like you could only take a couple of different routes through a mission.

So with that all in mind, I don’t really think we need Deus Ex Remastered, but rather a brand-new Deus Ex game, one that builds on the story of the prequel or explores an alternate sequel to the direction Invisible War took.

Those are my musings; were you looking forward to Deus Ex Remastered, or would you rather have a different Deus Ex game? Let me know in the comments below.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/gaming/forget-the-remaster-2000s-deus-ex-is-still-a-masterpiece-that-you-should-play-in-its-original-form xgNCL2aPm5s88fQVQugsni Thu, 15 Jan 2026 03:00:00 +0000 Wed, 14 Jan 2026 12:16:42 +0000 <![CDATA[ Meta just killed some of its best Quest 3 game studios — and convinced me to buy a Steam Frame instead of a Quest 4 ]]> Meta is turning its back on VR gaming. That’s the feeling many, myself included, are feeling after the reports that Meta has shuttered Twisted Pixel, Sanzaru, and Armature, the game studios behind Marvel's Deadpool VR, Asgard’s Wrath, and Resident Evil 4 VR, respectively (among other VR projects).

In addition, Supernatural – the excellent VR fitness app that was the linchpin of my VR exercise routine – will no longer be receiving updates or new content (via the Supernatural Facebook group and emails sent to users). It will, however, “remain active”, though questions remain about certain specifics – for example, will Supernatural’s licensed music catalog eventually dwindle to almost nothing, or will partnerships still be renewed?

This also all comes in the wake of the news from late last year that Meta had “paused” its Horizon OS partnerships with Asus and Lenovo.

In a statement to Engadget, a Meta spokesperson said, “We said last month that we were shifting some of our investment from Metaverse toward Wearables. This is part of that effort, and we plan to reinvest the savings to support the growth of wearables this year."

When I followed up with Meta to address claims it was turning its back on VR, I was told by a representative that “Meta doesn’t have anything to share on this.”

Deadpool fighting Hand ninjas with kunai — Deadpool VR was a blast (Image credit: Meta / Marvel)

While Meta might not have much more to share on this matter, I do: it’s pretty clear Meta isn’t keen to push VR in quite the same way it once was.

The Oculus Quest 2 felt like the moment VR was waiting for. Aided, in part, by people being trapped at home due to COVID-19-era lockdowns, and its truly affordable price for a solid VR machine, the Quest 2 sold phenomenally well.

In less than half a year, it had sold more units than all previous total sales of all other Oculus headsets, and while it's hard to track precise figures for every machine, it seems that for a while the Quest 2 was seemingly even selling better than the Xbox Series consoles (via Android Central).

However, Meta’s golden goose didn’t last forever, and the Meta Quest 3 didn’t land with anything like the same splash, despite being such an awesome device in its own right.

And now Meta has a new favorite child: AI and its AI wearables like the (also hugely popular) Ray-Ban Meta glasses, and as it has said, it is shifting some of its investment and priorities towards this new venture.

Ray-Ban Meta Smart Glasses Meta Connect updates — Ray-Bans are in, VR is out (Image credit: Future / Lance Ulanoff)

Is VR gaming dead?

VR gaming isn’t dead yet, if not least because Valve has its Steam Frame headset due to land later in 2026, and while it won’t be as affordable as a Quest 3 (most likely) the fact it’s a Steam Deck for your face – so it kind of doubles as a Steam Deck and a pair of XR glasses to enjoy your games on a large virtual screen – as well as having all of its VR gaming potential makes me at least a little confident it will have a strong landing.

Plus, if Valve or Google are looking to bolster their XR gaming efforts – Google has its Android XR platform – I’ve heard a lot of talented folks are currently free agents.

Marvel’s Deadpool VR is a fantastic game for capturing the true promise of what it would be like to be Deadpool, and Asgard’s Wrath 2 is a superb, more open-world action adventure that VR doesn’t have enough of. Seeing the teams behind those games work on something for a VR headset with some PC power behind it could be incredible.

The Steam Frame could save the day (maybe) (Image credit: Valve)

Or if Android XR wants a dedicated XR fitness platform, the Supernatural team knows what it’s doing, and I think there’s so much more to explore with fitness in XR – especially in AR. Plus, from a music licensing side of things, Google might have an even easier time, as it already has services like YouTube Music, which could open up an even broader suite of workout options.

We’ll have to wait and see how 2026 unfolds, but while this announcement is frustrating for Quest fans, I believe there’s still some potential for VR to have a stronger year than it did in 2025 – even if it is elsewhere.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.

]]> https://www.techradar.com/computing/virtual-reality-augmented-reality/meta-just-killed-some-of-its-best-quest-3-game-studios-and-convinced-me-to-buy-a-steam-frame-instead-of-a-quest-4 mHKptg2x3oZwBabFPU3zzG Thu, 15 Jan 2026 00:00:00 +0000 Wed, 14 Jan 2026 20:08:25 +0000 <![CDATA[ I’m a backup and recovery provider, but here’s why you shouldn’t just trust me ]]> How much do you trust your backups? It’s an important question, and one that few businesses think to ask themselves until it’s too late. There’s a persistent belief in operational technology (OT) environments that a completed backup equates to a recoverable system.

A green flag on a dashboard may indicate a successful backup, but unless that backup is continuously tested and validated against current OT conditions, the “recovery” element – the most critical part of a backup and recovery strategy – is left to chance. And the more complex the environment, the more those chances dwindle.

That’s especially true in critical infrastructure such as factories, hospitals, labs, and transport networks, where the underlying architecture is usually far more fragile and diverse than mainstream enterprise IT. Many of the systems that underpin production or safety are built on legacy systems that can’t be easily virtualized or replaced.

A backup taken from these environments may appear intact, but without validation there’s no way of knowing if the data is corrupted, if drivers are missing, or if images are incomplete.

Those issues rarely reveal themselves until an incident occurs and what should have been a “backup and recovery” process turns into a “disaster recovery” process.

A lot of organizations treat a completed backup as the final word on resilience. They see the green light, assume the process has worked, and trust that if anything goes wrong everything will behave as expected.

That’s a lot of trust to place in a basic backup process at a time when the threat surface is expanding faster than legacy-heavy OT environments can keep up. Last year, almost one-third of global ransomware attacks exploited unpatched vulnerabilities.

Cybercriminals are also four times more likely to target end-of-life systems – a list which, as of October 2025, now includes Windows 10. For organizations without a continuously validated backup and recovery process in place, the risks are mounting.

OT environments face pressures that traditional IT rarely encounters. Any interruption has immediate financial or safety consequences, which makes them prime targets for ransomware groups who know manufacturers, hospitals, and logistics providers can’t afford extended downtime.

The convergence of OT and IT only widens this attack surface, creating a landscape where even minor configuration drift or unspotted corruption can carry outsized consequences. In this context, treating a green tick as proof of resilience simply doesn’t hold up.

Why OT recovery is never as simple as it seems

The reality is that a company’s technology stack is rarely as modern as it might outwardly seem. Critical processes still rely on unsupported operating systems like Windows XP or Windows 7, bespoke embedded editions, or equipment controlled by aging Programmable Logic Controllers (PLCs).

Windows XP support ended in 2014, yet many organizations continue to operate XP-dependent devices. These systems often sit behind brittle chains of custom drivers and proprietary interfaces that may not have been manufactured in years.

Documentation is often missing, and the engineers who originally configured them have long since moved on. What’s left are inconsistent system states that can’t easily be lifted onto new or even slightly different hardware during a crisis.

Some OT environments limit change by necessity. Hospitals must avoid patching certain devices to maintain certification; manufacturing lines depend on chipsets that can’t be virtualized; air-gapped or remote sites rely on images that may not reflect current conditions.

In these cases, a backup that “succeeds” is often just one that didn’t encounter an obvious error – not one that can actually be restored.

Production lines, clinical systems, logistics hubs, and industrial control networks aren’t built with pause buttons. Even brief outages ripple outward into missed quotas, stalled deliveries, spoiled batches, safety risks, or overtime recovery costs.

It’s why ransomware campaigns increasingly target OT systems: they know the business impact is so severe that many organizations will pay simply to resume operations.

The Jaguar Land Rover incident, dubbed by some as “the most costly cyberattack in UK history”, is a case in point. When production was disrupted by issues linked to unprepared OT processes, delays cascaded across supply chains and dealer networks for weeks.

It demonstrated a truth the OT sector knows all too well – once operations stop, the financial and operational damage continues long after systems come back online.

Without proof that systems can be restored reliably, organizations are effectively gambling their production schedules, reputation, and revenue on the hope that the restore will work when they need it most.

How to validate your backups

So how do you actually validate? It's not a single test - it's a systematic process that moves from quick checks to full-scale recovery drills. Here's how:

Start with integrity checks Run hash verification or checksum comparisons to confirm that backup data matches the source and hasn't been corrupted. This catches silent data degradation - file corruption, partial overwrites, and unexpected changes that sit undetected for months.

Move to virtual test restores Boot a backup in an isolated virtual environment to confirm that operating systems, drivers, and applications load as expected. This reveals missing dependencies, configuration issues, and service initialization failures that integrity checks can't detect.

Test on actual hardware Restore to the same type of production hardware you'd use in a real recovery. This exposes physical dependencies that virtualization masks: driver compatibility issues, firmware mismatches, and hardware-specific configurations. A backup that boots in a VM might fail entirely on real hardware.

Run full recovery drills Restoring one system is different from restoring 20 or 200. Test scenario-based drills that mirror real incidents - ransomware, site failures, supply chain disruptions - and document how long recovery actually takes versus your RTO targets.

Build it into incident response Train teams on which backups to use in different scenarios, how to isolate compromised systems, and how to restore in the right order. Make recovery muscle memory, not something you frantically figure out during a crisis.

Document and refine After every test, record what worked and what didn't. Update your runbooks, feed lessons back into your backup schedule and storage choices, and create a cycle of continuous improvement. The 3-2-1-1-0 model captures this in its final digit: zero errors.

When organizations rehearse these restores systematically and refine their processes based on results, they turn backup and recovery from a box-ticking exercise into a resilient operational function. Validation gives you certainty, not hope, that recovery will work when it really counts.

The green light means nothing

I'm a backup and recovery expert, and this is why you shouldn't just trust me—or anyone who says your backups will simply work when you need them.

When it comes to operational resilience, organizations should operate with zero trust until they can prove to themselves, and demonstrate to others, that they can recover exactly as needed. Trust is what you place in a green light on a dashboard. Proof is what you earn through testing and validation.

In OT environments where downtime is detrimental, where legacy systems can't be easily rebuilt, and where attackers target the most vulnerable points - proof isn't optional. A completed backup offers reassurance. A validated backup offers certainty. And in critical infrastructure, only certainty keeps operations running.

We've featured the best cloud storage.

]]> https://www.techradar.com/pro/im-a-backup-and-recovery-provider-but-heres-why-you-shouldnt-just-trust-me gborDxNTzGraU9kTtXoaBA Wed, 14 Jan 2026 15:12:25 +0000 Wed, 14 Jan 2026 15:12:26 +0000 <![CDATA[ AI is no SKU—and what that means for the enterprise ]]> Nothing grabs readers' attention like bad news. For example, technology leaders everywhere paid close attention to the recent MIT study, which found nearly all generative AI pilots fail to deliver real financial results – 95%, to be exact.

Research from RAND has been similarly bleak, showing that four out of five AI projects stall out. And S&P Global discovered that organizations are abandoning AI initiatives at double the rate of the previous year.

We’ve all read think pieces that diagnose these failures as a result of weak models, immature tools or a lack of internal expertise and capability. However, the truth is simpler and perhaps even more uncomfortable.

The reality is that most AI projects fail not because of the technology, but instead because of the strategy—or lack thereof—behind them.

The problem: Treating AI like a product you can buy

Anyone walking the floor of an enterprise software conference will be greeted by a sea of vendors hawking shelf-ready “AI solutions”. Implicitly, the message they send is that AI is a product that someone can purchase, plug in, and then watch the efficiency benefits roll in.

And it is this highly commercialized view of AI, treating it like a stock-keeping unit (SKU) that you can order from a catalogue, that leads to AI projects unravelling shortly after purchase.

This is because AI is not a pre-boxed answer in search of a use case. Instead, it is a set of techniques that only create value when they are applied to a specific, well-defined business issue. So without a clear path to ROI from the start, it is highly likely that AI projects will fail.

When organizations forget this, they fall into what many leaders sometimes refer to as the “science-experiment” trap. It appears in two common forms. In the first, we may see internal excitement about a new model or tool, and the pilot may even secure generous funding from senior leadership, followed by an impressive demo.

But within a few months, it will fade away because no one can link it to measurable outcomes.

The MIT research confirms this, finding that purchased AI tools succeed 67% of the time while internal builds succeed only one-third as often, largely because vendor solutions come with clearer use cases and success metrics tied to specific business outcomes.

The second version starts at the top. This is when leadership puts out a call for an “AI strategy,” and in the rush to demonstrate action, an organization buys hardware or software that looks impressive on paper.

Only once it arrives does the team realize they don’t have a clear use for it, or the AI fails to deliver. In both cases, we see the same story: when the business problem comes last, there’s no space for value.

Reframing AI around business reality

The good news is that there is a better way to evaluate and prioritize AI investments. At its core, it comes down to how well teams and leaders distinguish between AI ideas that are and aren’t worth pursuing.

The starting point is to identify the biggest challenges facing an organization and what matters to its core performance. For most, this will be revenue growth, efficiency, customer satisfaction, risk exposure or perhaps productivity. The conversation must start with a business need, not a technical curiosity.

The next step is to assess the data required to solve those challenges. In many organisations, data is scattered and duplicated across multiple sources or trapped in legacy systems.

Until this data is consolidated and cleaned up, there is no model on earth that will be able to work with it effectively. Once the legwork is done to get the data estate in order, the potential for AI to solve this problem becomes tangible.

Then it's important to consider measurability, and how the organization already measures things like revenue growth, efficiency or customer satisfaction. Having this baseline in place means any improvements from AI can be demonstrated clearly.

Metrics such as cycle time, accuracy, cost per transaction or customer scores give you the ability to show impact.

Allowing AI to earn its place in the business

By applying these considerations, you’ll likely see the number of potential AI projects shrink. That’s a good thing—because the projects that remain will be directly to business value and supported by data that’s fit for purpose. Teams can then set success criteria that everyone understands and can measure over time.

In turn, this approach can shift AI from being a costly experiment to adding tangible value. It focuses investment on what is solvable now rather than what is theoretically exciting. And most importantly, it enables AI to truly deliver on its promises of transformation.

We've featured the best AI website builder.

]]> https://www.techradar.com/pro/ai-is-no-sku-and-what-that-means-for-the-enterprise UhmPYsiKaA6T3kmg8kEjPn Wed, 14 Jan 2026 15:08:36 +0000 Wed, 14 Jan 2026 15:08:36 +0000