TechRadar News

• Microsoft is trying to push its AI chatbot Copilot, but is hitting roadblocks
• Copilot is struggling against rival ChatGPT
• ChatGPT launched much earlier, giving the model a head start

AI chatbot models like ChatGPT and Microsoft Copilot are gaining huge traction in the workplace and beyond, but OpenAI's model is reportedly starting to show dominance in the business world.

A Bloomberg report has claimed even businesses which have purchased Copilot plans are struggling to convince users to make the switch, with drugmaker Amgen buying a 20,000 user plan, only to have employees continue using ChatGPT over a year later.

It is not the only customer to report issues, as workers shun Microsoft Copilot in favor of ChatGPT, which is used more at home and thus become more familiar to many.

A head start

Microsoft does use OpenAI’s models to power Copilot, and offers very similar features to ChatGPT, with information summaries, email drafting, data analysis, and image generation.

That being said, ChatGPT’s momentum and existing user base seems to be giving the chatbot the edge.

As of June 2025, ChatGPT has almost 800 million weekly active users, and 3 million paying business users, whereas Copilot has stagnated a little, with 20 million weekly users for the past year.

Theoretically, the race should be a little more even, since Windows is such a dominant operating system in the professional world. Microsoft salespeople have traditionally been able to use Windows compatibility as an effective selling point, but that’s no longer the case, the report suggests;

“The company’s [Microsoft’s] salespeople knew ChatGPT dominated the consumer chatbot market, but expected Microsoft to own the enterprise space for AI assistants thanks to decades-long relationships with corporate IT departments. But by the time Microsoft began selling Copilot to businesses, many office workers had already tried out ChatGPT at home, giving the chatbot a first-mover advantage.”

Despite negotiating with companies such as Volkswagen, Accenture and Barclays, which all signed deals for over 100,000 accounts in agreements worth ‘tens of millions’ a year, Microsoft still lags behind OpenAI in its user base, and organizations are having to encourage workers to use the chatbot.

The news comes after Microsoft has announced largescale layoffs, with between 6,000 and 7,000 jobs worldwide expected to be cut - amounting to nearly 3% of the firm’s workforce - just two years after 10,000 personnel were made redundant (5% of the workforce).

You might also like

• Take a look at our picks for the best AI tools around
• Check out our choice for best antivirus software
• Generative AI isn't biting into wages, replacing workers, and isn't saving time, economists say

• Colors have leaked for the Z Fold 7 and Z Flip 7
• The leak comes from Samsung's own website
• We could get three main colors for each phone

We've already seen plenty of Samsung Galaxy Z Fold 7 and Galaxy Z Flip 7 leaks in recent days – and now it appears Samsung has got in on the act itself

As spotted by Android Authority, certain pages have gone live on Samsung's Ireland website, showing three colors for the Z Fold 7 and three for the Z Flip 7 – with both handsets getting a fourth Green Mint shade that's exclusive to the website.

For the Samsung Galaxy Z Fold 7, it looks like we'll get Jet Black, Blue Shadow, and Silver Shadow. That compares to Silver Shadow, Pink, and Navy for the Samsung Galaxy Z Fold 6, at least in the US (with Crafted Black and White as website exclusives).

The leak suggests Jet Black, Blue Shadow, and Coral Red as the main colors for the Samsung Galaxy Z Flip 7. In the US right now, you can pick up the Samsung Galaxy Z Flip 6 in Yellow, Mint, Silver Shadow, and Blue, with Black, Peach, and White the colors exclusive to Samsung's website).

Throwing shades

More Galaxy Z Fold7 images have emerged!Via @evleaks pic.twitter.com/NPm0bJE82lJune 25, 2025

This latest leak lines up pretty neatly with what we've previously heard about the Samsung Galaxy Z Fold 7 colors and the Samsung Galaxy Z Flip 7 colors. While there will of course be variations over time and between regions, it seems Samsung may have now settled on the color options it's going to give us.

It's worth mentioning that Coral Red has also been tipped for the Galaxy Z Fold 7, so it's possible that the distinctive shade is coming to both of these foldables – which are also both likely to be instant additions to our best foldable phones guide.

The color leaks also fit in neatly with unofficial renders posted by @evleaks: these renders also show Jet Black, Blue Shadow, and Silver Shadow for the Galaxy Z Fold 7, and Jet Black, Blue Shadow, and Coral Red for the Galaxy Z Flip 7.

What we do know for sure is that Samsung has scheduled the next Galaxy Unpacked event for Wednesday, July 9. The smart money suggests we should see both these flagship foldables appear, as well as the Samsung Galaxy Z Flip FE and the Samsung Galaxy Watch 8.

You might also like

• Renders of Samsung's next foldables have leaked
• These are the best Android phones right now
• Android 16 is heading to the Z Fold 7 and Z Flip 7

• Jurassic World Evolution 3 will no longer feature AI generated character portraits
• This was confirmed by the developers via the Steam Forum
• It follows community backlash regarding the inclusion

The developer of the upcoming dinosaur park management simulator Jurassic World Evolution 3 has announced that the game will no longer feature some AI-generated visual assets following community outrage.

When the next entry in the Jurassic World Evolution series was first revealed at last year's Summer Game Fest, some eagle-eyed fans spotted an AI-generated content disclosure on its Steam page. "Scientists' avatars are created using generative AI," it read, referring to some in-game character portraits.

The disclosure, which is now required by Valve, is intended to increase transparency and allow consumers to decide whether they are comfortable purchasing a product with any generative AI involvement.

Many Jurassic World Evolution fans reacted negatively to the news, with a huge thread on the Steam forums discussing the move. "Nope, I will not support this," wrote one potential player. "It's a matter of principle," chimed in another.

A few defended the decision: "If using AI to quickly generate images for scientists [...] helps free up time for development on actually-important mechanics, then it is worthwhile to use. Being upset over something that is so inconsequential to the rest of the game is pretty ridiculous."

No matter which side of the AI content debate you fall on, polarising your fanbase like this ahead of a major launch is never a particularly good idea and risks overshadowing all other conversation regarding the game.

This is presumably one of the reasons why developer Frontier Developments has announced that the game will no longer feature the AI-generated assets.

"Thanks for your feedback on this topic," a representative wrote on the Steam forum "We have opted to remove the use of generative AI for scientist portraits within Jurassic World Evolution 3."

The AI-generated content disclosure has now been removed from the Steam page, too, which presumably means that the title will no longer feature any AI-generated content.

Jurassic World Evolution 3 is set to launch on October 21, 2025, for PC, Xbox Series X and Series S, and PlayStation 5.

You might also like...

• King of Meat’s chaotic co-op action and quirky creator mode show me that it might just have the legs to succeed
• The Corsair Platform:4 packs big gaming desk performance into a small package that’s perfect if you’re short on space
• Build A Rocket Boy confirms it has begun the redundancy process following MindsEye's disastrous launch

Complex problems often demand simple answers. When we overcomplicate things, whether in life or business, we almost always end up worse off. Later, we look back and think: if only I’d kept it simple.

Cybersecurity is no different, though the source of that simplicity may lie in unexpected places.

With the National Cyber Security Centre (NCSC) now sounding the alarm on quantum-era threats and AI-powered malware, it’s clear the risks are evolving fast. These threats adapt, mutate and inject themselves into systems at alarming speed. It’s no wonder business leaders are extremely concerned about the risk of existing cyber strategies and deployed solutions being overwhelmed.

Outspending the problem isn’t working

A recent McKinsey report reveals that cybersecurity spending surged to $200 billion in 2024—up from $140 billion in 2020—yet breaches keep rising.

To confront these rising risks, organizations are doubling down on complex cybersecurity stacks, layering tools in the belief that more technology equals more protection.

But what if that logic is flawed? What if, instead of boosting your system resilience, complexity increases and hides your vulnerabilities? In truth, we’re stuck in a complexity trap.

Organizations are drowning in software solutions that promise the world but deliver confusion. Each new tool might address a specific threat vector, but the resulting patchwork of platforms often leads to fragmented visibility and hidden blind spots.

In short, we risk opening more doors that attackers can walk through.

By trying to guard against every threat, we become entangled in complexity and exposed to its consequences—creating a false sense of security in the process.

Simplicity solves complexity

When you strip back your cybersecurity layers and concentrate on a back-to-basics approach that’s founded on clarity, control and isolation, you achieve better protection than any complex software stack.

Now, this isn’t about throwing out digital defenses. It’s about recognising their limits and rethinking where real resilience comes from.

Software alone, no matter how smart, is still vulnerable to manipulation. And with AI supercharging attacks in real time—learning from failed breach attempts, mimicking user behavior and exploiting every crack in the system at an accelerating pace—this has never been truer.

That’s why physical isolation has stepped back into the conversation. It’s not just a legacy idea from a pre-cloud era; it’s the critical missing idea in modern cyber strategy.

The case for physical network isolation

Highly motivated threat actors and AI-powered malware have the ability to think and spread without human input. With devastating precision, it targets high-value assets, adapting mid-attack.

This calls for a defense that is unhackable by nature.

Hardware-based network isolation is exactly that. When systems are physically segmented—truly separated from the internet – remote infection becomes impossible. The key to modern deployment of this traditional airgap method lies in being able to control it, at will, on demand.

If malware can't make contact, it can't compromise. It’s that simple.

Even if a system is somehow breached, physical segmentation allows businesses to readily contain the threat. When you isolate systems from one another with hardware, not just firewalls or virtual LANs, you prevent lateral movement, stop data exfiltration and drastically reduce the blast radius of any attack.

This is especially critical for operational technology, critical infrastructure and sensitive research environments, where uptime is essential and downtime is catastrophic.

An overdue shift in thinking

The complexity trap is reflected in how we spend. According to industry research, 65% of cyber budgets now go to third-party tools and services, outpacing investment in in-house capability.

But security is not just a tech problem; it’s a strategic design challenge. Businesses today react to new threats by accumulating more tools. What’s needed instead is a clear, layered security plan that’s built with purpose, not patched together.

That begins with rethinking how much of your infrastructure truly needs to be online. In a hyperconnected world, we’ve defaulted to keeping everything on all the time.

But always-on equals always-vulnerable. If certain data or systems don’t require constant internet access, why expose them?

By selectively disconnecting key assets, at the right time, you can regain control of your business.

The future starts with hardware

Let’s be clear: this isn’t a step backward. It’s a step toward resilience. Software-based security remains essential. But as threats evolve, our defenses must too.

Layered protection that starts with hardware-based control is the only viable way forward. It combines the speed and scale of software with the unbreachable foundations of physical isolation.

Think of it like a bank vault. The digital defenses are the alarms, cameras and motion detectors. But the vault? That’s your hardware-based barrier. Even the smartest thief can’t crack it from a distance.

Protecting your systems isn’t just about keeping up with the latest threats. It’s about doing what works, what’s reliable and proven.

Because just like in life, the clearest answers are often the strongest ones.

And in cybersecurity, simplicity is the ultimate advantage.

We list the best endpoint protection software.

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

• Over half of marketers say they can’t do their jobs without AI, despite poor results
• Budgets for AI keep rising, even though 81% of leaders say it's being wasted
• Marketers are stuck manually fixing AI output for days every week

Social media professionals have grown increasingly dependent on artificial intelligence, with new research finding more than half saying they now can’t imagine performing their roles without it.

A survey from Hootsuite claims this growing reliance is not matched by results despite heavy investments in AI technologies.

The firm's research reveals 88% of senior marketing leaders are encouraging their teams to use AI tools, yet 81% admit budgets are being wasted on tools not fit for purpose. 

Manual work persists despite automation promises

Delving deeper, Hootsuite found many marketers find themselves trapped in a time-consuming cycle of manual labor and subpar outcomes, revealing a deep disconnect between expectations and the actual utility of generative AI tools in marketing.

A significant proportion of social media managers still spend up to three full working days each week verifying AI-generated content and manually gathering insights from online platforms.

This lag not only drains staff time but also affects campaign performance.

As trends shift rapidly, marketers often find their content outdated by the time it is published, which may explain why over half of senior marketers feel their campaigns consistently underperform.

The financial implications are just as troubling. Budgets for AI tools continue to rise, yet for some, the wasted investment exceeds 20% of their entire marketing budget.

"This should be a wake up call to all marketers: traditional AI isn’t as sophisticated as you think it is," noted Irina Novoselsky, CEO at Hootsuite.

"With five billion people spending up to five hours a day online, social is one of the richest sources of real-time data sources available and yet, traditional AI tools still can’t harness it, leaving the insights marketers truly need hiding in plain sight."

(Image credit: NPowell/Flux)

With rising pressure from executive leadership to justify every expense, marketers are finding it increasingly difficult to defend investments in AI tools that fail to deliver tangible returns.

A critical weakness in current generative AI systems lies in their reliance on outdated datasets.

These tools often fail to capture the dynamic nature of real-time audience behavior, meaning that their insights may be out of sync with the present moment.

While 64% of senior leaders believe their AI tools offer real-time insights, only 39% of social media managers agree, a clear signal that confidence in AI’s real-world performance is uneven across organizational levels.

In response to these challenges, Hootsuite has introduced OwlyGPT, a generative AI assistant trained on live social data.

The company says, this tool delivers up-to-the-minute insights tailored to brand voice and cultural context.

Considering the issues with AI static data, this move appears promising, but it's good to approach it with some skepticism. After all, businesses have been led to believe in AI’s transformative power before, only to confront disappointing results.

You might also like

• These are the best AI website builders around
• Take a look at our pick of the best internet security suites
• This is the first 8K 120Hz monitor ever produced - but I don't think the tech market will ever be ready for it

From my vantage point, I see the legacy mainframe landscape as both a testament to decades of reliable operation and a critical juncture demanding strategic evolution. The global economy’s reliance on these systems is undeniable – they are the silent workhorses powering a significant majority of business transactions.

However, the accelerating pace of technological advancement, coupled with the realities of hardware lifecycles and a shifting talent pool, calls for a proactive and thoughtful approach to their eventual end-of-life. The question is no longer if we modernize, but how we navigate this complexity without disrupting the very core of operations.

Cost and complexity

A primary hurdle is the significant cost and inherent complexity of these transformations. Mainframe modernization isn't a simple tech refresh; it demands substantial investment, time, and meticulous planning. Decades of accumulated technical debt, often manifested as undocumented code and intricate dependencies, require a phased and strategic approach.

Carving the application portfolio into thin, business-aligned slices, is a great approach to deliver tangible value in shorter cycles. Prioritizing initiatives with clear and early ROI, such as migrating non-critical workloads, builds momentum and stakeholder confidence.

The shrinking pool of mainframe-skilled professionals presents another critical challenge. The reality is that the workforce with deep expertise in these legacy systems is nearing retirement, creating a potential knowledge vacuum. To mitigate this, we advise for codification of tribal knowledge.

This involves leveraging tools to harvest specifications from production logs and source code analytics, while also pairing retiring experts with cross-skilled engineers. Investing in upskilling programs that bridge the gap between COBOL literacy, modern cloud-native and observability skills is paramount to building a future-ready workforce capable of managing both legacy and target environments during transition.

Data migration

Migrating petabytes of critical, often poorly documented, business data and its embedded logic to modern platforms is a high-stakes challenge, with severe risks of data loss or corruption. A recommended strategy involves inverting data gravity: implement an API façade over shared datasets and incrementally replicate data to the target platform using event streaming, thereby minimizing disruption.

Employing anti-corruption layers ensures a clean decoupling of legacy systems, aligning migration with modern architectures while safeguarding core business processes. Another technique we support is behavior equivalence and leveraging data seams to integrate into the origin system while the architecture is going through its evolution.

Beyond the technical aspects, organizational misalignment is a frequent stumbling block. Modernization is a business transformation, not just an IT project, and one of the most reliable indicators of success is a courageous, well-supported leader who can steer the program through the inevitable stumbles and issues that arise. A clear business vision, tied to measurable outcomes like improved customer experience or reduced operational risk, is essential.

Culture of change

Fostering a culture of change through transparent communication, targeted training, and deliberate capability-building is crucial: the destination team must be fully trained and capable of operating a platform of this criticality and complexity, which goes far beyond a typical N-Tier architecture. Such preparation helps overcome internal resistance and ensures everyone understands, and can realize, long-term benefits.

Finally, integration and observability gaps can derail even the most well-intentioned modernization efforts. Legacy systems are often deeply embedded within the broader IT ecosystem, so updating core components can surface unforeseen integration challenges.

To counter this, we advocate for enhancing observability from day one, including baselining the performance of existing mainframe jobs and screen transactions; these metrics establish a benchmark to keep the modernized environment aligned with current service levels.

This early telemetry is paired with modern monitoring solutions and real-time dashboards that provide comprehensive insights into system behavior. Prioritizing API-first integration ensures seamless communication between legacy and new architectures, while automated testing at integration points minimizes disruption risk during the transition.

The advent of AI

Generative AI is accelerating mainframe modernization by offering powerful opportunities to analyze legacy systems and streamline transformations, delivering greater agility and resilience. This shift is mirrored by modernization spending moving from defensive capital expenditure to growth-focused operational expenditure.

Success in this evolving landscape hinges on disciplined execution, continuous measurement, and transparent communication, rather than merely relying on tools. Furthermore, enhanced cloud platforms now provide flexible and secure migration paths, significantly boosting the strategic importance of these efforts due to heightened regulatory scrutiny on operational resilience.

Here’s a modernization manifesto to bear in mind:

• Start with the business capability map, not the code tree. Technology serves the business.
• Prefer incremental stranglers over all-or-nothing migrations. Decommissioning legacy components is a tangible win.
• Automate everything twice – first to understand, then to transform. Leverage code scanners, synthetic data, and automated testing.
• Invest in people before platforms – empower cross-functional teams that own both the old and the new.
• Measure outcomes continuously and publicly – demonstrate the business value of modernization with clear KPIs.

In conclusion, mainframe end-of-life is not an event to be feared, but a strategic imperative to be navigated with diligence and foresight. It’s a long-term commitment to ensuring our critical value streams remain adaptable and resilient.

We list the best laptop for programming.

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

• A federal judge rejected a ChatGPT user's petition against her order that OpenAI preserve all ChatGPT chats
• The order followed a request by The New York Times as part of its lawsuit against OpenAI and Microsoft
• OpenAI plans to continue arguing against the ruling

OpenAI will be holding onto all of your conversations with ChatGPT and possibly sharing them with a lot of lawyers, even the ones you thought you deleted. That's the upshot of an order from the federal judge overseeing a lawsuit brought against OpenAI by The New York Times over copyright infringement. Judge Ona Wang upheld her earlier order to preserve all ChatGPT conversations for evidence after rejecting a motion by ChatGPT user Aidan Hunt, one of several from ChatGPT users asking her to rescind the order over privacy and other concerns.

Judge Wang told OpenAI to “indefinitely” preserve ChatGPT’s outputs since the Times pointed out that would be a way to tell if the chatbot has illegally recreated articles without paying the original publishers. But finding those examples means hanging onto every intimate, awkward, or just private communication anyone's had with the chatbot. Though what users write isn't part of the order, it's not hard to imagine working out who was conversing with ChatGPT about what personal topic based on what the AI wrote. In fact, the more personal the discussion, the easier it would probably be to identify the user.

Hunt pointed out that he had no warning that this might happen until he saw a report about the order in an online forum. and is now concerned that his conversations with ChatGPT might be disseminated, including “highly sensitive personal and commercial information.” He asked the judge to vacate the order or modify it to leave out especially private content, like conversations conducted in private mode, or when there are medical or legal matters discussed.

According to Hunt, the judge was overstepping her bounds with the order because “this case involves important, novel constitutional questions about the privacy rights incident to artificial intelligence usage – a rapidly developing area of law – and the ability of a magistrate [judge] to institute a nationwide mass surveillance program by means of a discovery order in a civil case.”

Judge Wang rejected his request because they aren't related to the copyright issue at hand. She emphasized that it's about preservation, not disclosure, and that it's hardly unique or uncommon for the courts to tell a private company to hold onto certain records for litigation. That’s technically correct, but, understandably, an everyday person using ChatGPT might not feel that way.

She also seemed to particularly dislike the mass surveillance accusation, quoting that section of Hunt's petition and slamming it with the legal language equivalent of a diss track. Judge Wang added a "[sic]" to the quote from Hunt's filing and a footnote pointing out that the petition "does not explain how a court’s document retention order that directs the preservation, segregation, and retention of certain privately held data by a private company for the limited purposes of litigation is, or could be, a “nationwide mass surveillance program.” It is not. The judiciary is not a law enforcement agency."

That 'sic burn' aside, there's still a chance the order will be rescinded or modified after OpenAI goes to court this week to push back against it as part of the larger paperwork battle around the lawsuit.

Deleted but not gone

Hunt's other concern is that, regardless of how this case goes, OpenAI will now have the ability to retain chats that users believed were deleted and could use them in the future. There are concerns over whether OpenAI will lean into protecting user privacy over legal expedience. OpenAI has so far argued in favor of that privacy and has asked the court for oral arguments to challenge the retention order that will take place this week. The company has said it wants to push back hard on behalf of its users. But in the meantime, your chat logs are in limbo.

Many may have felt that writing into ChatGPT is like talking to a friend who can keep a secret. Perhaps more will now understand that it still acts like a computer program, and the equivalent of your browser history and Google search terms are still in there. At the very least, hopefully, there will be more transparency. Even if it's the courts demanding that AI companies retain sensitive data, users should be notified by the companies. We shouldn't discover it by chance on a web forum.

And if OpenAI really wants to protect its users, it could start offering more granular controls: clear toggles for anonymous mode, stronger deletion guarantees, and alerts when conversations are being preserved for legal reasons. Until then, it might be wise to treat ChatGPT a bit less like a therapist and a bit more like a coworker who might be wearing a wire.

You might also like

• Make ChatGPT remember who you are by doing these 5 simple things
• ChatGPT crosses a new AI threshold by beating the Turing test
• Sam Altman says AI chats should be as private as ‘talking to a lawyer or a doctor’, but OpenAI could soon be forced to keep your ChatGPT conversations forever

• The first 8K 120Hz monitor recently shown off at Display Week event
• Real-world use for 8K monitors still feels far out of reach though
• Display makers keep building while the market remains unconvinced

At the recent Display Week 2025 event, Chinese firm BOE showed off the first-ever 31.5-inch 8K monitor capable of running at 120Hz.

The CR3000 offers a contrast ratio of 8000:1, a color gamut of 99% DCI-P3, and also supports 240Hz in 4K mode.

BOE, which is the largest panel maker in the world and was also a sponsor of the show, told 8K Association it expects to begin mass production later in 2025, although details on pricing and final product integration are still unknown.

Other 8K panels on show

Display Week often serves as a glimpse into where display tech may be headed rather than where it currently is. That pattern continued this year with a number of other 8K panels on show.

TCL/CSOT brought an inkjet-printed OLED 8K TV panel, a project built partly from its acquisition of JOLED, and SEL surprised attendees with an 8.3-inch 8K LCD panel that offered over 1,000ppi, making it the sharpest full-color LCD shown to date.

As well as its 8K 120Hz beast, BOE had a number of other products on show. These included the latest version of its miniLED UB Cell 4.0 ADS Pro TVs, which aim to challenge OLED with deeper contrast and better efficiency, and an 85-inch 4K panel with an RGB backlight system running in a filterless mode that could one day reduce power usage and complexity, especially in 8K applications.

It also had a 3D display prototype with eye-tracking based on a 16K development. Although still early-stage, the image quality and parallax control impressed those who got to see it in action.

Still ahead of its time

BOE's CR3000 panel arrives at a time when the broader market is still catching up to high refresh 4K gaming, let alone 8K.

While PC gamers have begun to see mainstream GPUs offer stable 4K60 gameplay, pushing that four times over in resolution and double in refresh raises some difficult questions. Upscaling and frame generation may be more of a necessity than a feature if such a panel is to be usable for gaming or creative work.

While I can't help but be impressed by BOE's 8K 120Hz monitor, it feels like it’s ahead of its time. The hardware to drive it effectively doesn’t exist at scale, and most buyers likely aren’t ready for what would surely be a high-cost niche product.

8K monitors were expected to hit the mainstream a few years ago, but that didn’t happen. This latest panel might be technically impressive, but I for one am not convinced the world is ready for it.

You might also like

• These are the best 5K and 8K monitors around today
• I don’t believe 8K monitors will ever go mainstream, for better or worse
• Will Asus be able to break the curse of failed 8K monitor launches?

Whether for aesthetic reasons or to cut down on screen time, having a TV in the bedroom isn’t for everyone. I didn’t factor a TV in when I designed my bedroom, as it wasn’t worth sacrificing the space when I’ve already got one in my living room, but after a while, I found I missed having the option to curl up in bed and binge-watch my comfort shows on Netflix.

Having not always had the luxury of separate living spaces, I’d put a lot of work into curating my bedroom into a calming and visually pleasing environment, so the idea of sticking a big black rectangle in the middle wasn’t going to do my zen any favors. Therefore, I knew I had to think of an alternative solution that could cure my content cravings without taking up valuable space.

The concept of using a projector to watch shows in bed wasn’t new to me, as I’d racked up plenty of hours watching movies on the Anker Nebula Cosmos 4K SE. Sadly, though, as impressive as that projector is, it proved impractical for bedroom use as it was a bit big and loud for the shelf above my headboard and, as I’m yet to find a tripod that can handle its weight, it just wasn’t the bedfellow I was looking for.

Thankfully, I found the perfect alternative in the Anker Nebula Capsule 3 1080p Mini Google TV Projector, which has a list price of $529.99 / £499.99 / AU$1,599, so it doesn’t cost any more than a decent budget TV. The Nebula Capsule 3 uses the same Google TV operating system that I found so effortless to use with the Cosmos 4K SE, but this time in a conveniently compact package.

Below, you’ll find the reasons why I believe the Anker Nebula Capsule 3 1080p Mini Google TV Projector makes for an amazing alternative – and one reason why opting for a projector over a TV may not be the brightest move.

HighlightsPerfect placement isn't paramount

Finding space for a TV set can be tricky, but choices are far from limited when it comes to finding a home for the Anker Nebula Capsule 3.

Its dinky diameter of just 3.1 inches / 78mm makes it conveniently compact and easy to fit on shelves or tabletops, and it has a super convenient tripod mount thread on the base, so it’s easy to find a place for it even if surface space is limited.

And for those times when it isn’t possible to get the angle of the projection spot-on, the Nebula Capsule 3 will automatically adapt its settings to ensure it projects a well-focused image within the space provided, adjusting the keystone positions and avoiding any obstacles along the way.

I can go big and go home

The beauty of a projector like the Nebula Capsule 3 is that I can change the screen size to suit what I’m watching, and I can do so in a matter of moments.

This means that it’s super easy for me to go from watching TV on a 49-inch projection on the wall at the side of my bed to a projection of around 80 inches on my free-standing projector screen, if I’m in the mood for some big-screen entertainment, simply by rotating the Capsule 3 90 degrees and waiting for the settings to auto adapt based on the new position.

Pleasantly portable projection

Whether you’re staying at a friend's or going camping, the compact dimensions and light weight of 1.9lb / 850g combined with a built-in rechargeable battery make the Capsule 3 satisfyingly easy to pack up and take away.

These features can prove useful even if you don’t plan to take it away from home, especially if your bedroom is anything like mine, with its awkwardly located power outlets. While the 15,000 mAh battery only allows for about two and a half hours of screen time, the USB-C charging cable does mean that one of the best power banks could be used to stretch this duration a bit further.

Image 1 of 2

(Image credit: Future)Image 2 of 2

(Image credit: Future)Low lightNight owls get the best view

Despite the space-saving and versatility on offer from the Anker Nebula Capsule 3 1080p Mini Google TV Projector, you’re going to be sacrificing deep blacks and the details in darker scenes if you opt for one over one of the best TVs.

This issue isn’t uncommon, even amongst some of the best projectors on the market, but it’s something to keep in mind if you want your shows to look picture-perfect when watching in the daytime without efficient blackout blinds.

With that being said, this hasn’t been a deal breaker for me personally, because as comfy as my bed is, it’s not my primary place for watching shows in the daytime. And during the times I’ve needed to curl up when it’s still light outside, I’ve learned that I can tolerate the picture looking a little washed out when the compromise is that I essentially have a pocket-sized 50-inch TV.

• AI tools are reshaping how we live, from morning chores to late-night admin
• Time isn’t just money anymore - it’s a function of how much tech you can manage
• AI promises more free time, but the wealth gap decides who gets to enjoy it

As AI continues to reshape how we work and live, the promise of regaining time is attracting growing interest.

New research from Lloyds Bank has claimed emerging technologies could help people reclaim up to 110 minutes of free time per day.

A focus on automating daily routines, such as chores, shopping, and travel, could help free up time, but the benefits appear skewed toward high earners. AI tools, including AI assistants, autonomous drones, and driverless vehicles, are framed as part of this shift toward a more efficient daily life, but these are not cheap.

AI tools free up time, but at a cost

The bank found in the UK, 86% of adults say having more time is important, rising to 99% among those earning over £100,000.

While 60% of the wider population is open to using new technologies to save time, this jumps sharply among affluent individuals, with nearly all saying they are willing to adopt such tools.

“We know life is hectic, with work, family, and personal commitments all vying for attention,” said Adam Rainey, Director of Mass Affluent at Lloyds.

“But our research shows people are becoming more comfortable with using technology to handle daily tasks.”

The most time-consuming responsibilities, according to the study, are cleaning, cooking, and managing finances.

Almost half (47%) of respondents identified household chores as their primary time drain, while 31% pointed to financial admin.

AI is being promoted as the solution through smart home devices or personal AI agents. These tools promise to handle repetitive work.

Yet many of the best AI tools come with steep costs or require a level of digital skill that remains out of reach for some.

Banking apps continue to lead among accessible time-saving tech, with 48% of adults relying on them. However, the gap widens when it comes to advanced tools; 49% of high earners are now using AI assistants, and 92% agree that wealth enables more free time.

It’s a compelling idea that could integrate everyone, but also one that raises the question - who has the means to work smarter?

As with the story of the Mexican fisherman, it’s worth asking whether we’re overengineering the pursuit of a simpler life some may already have, just without the premium subscription.

You might also like

• Microsoft Teams is finally fixing one of its most annoying pop-ups - so I might be more productive at work at last
• Check out the best productivity tools to speed up your day
• We've rounded up the best portable monitors available now

When we wrote our XGIMI MoGo 3 Pro review, which led to it being crowned as our favorite budget option among the best portable projectors, we noted three flaws: it didn't have a built-in battery, the speakers were weak, and it couldn't hold up well in bright conditions.

The last of these isn't unique to the MoGo series, and you have to pay a lot more for brightness that can stand up to sunlight, so that's fair enough.

The other two, however, are solvable problems – and in the MoGo 4, XGIMI has not only fixed them, it's made a better, tighter portable projector all around, based on my time with it so far.

This is far from a full review, but the improvements to the MoGo 4 were obvious to see just from my early time testing it. This is such an impressive portable entertainment device for its $499 / £509 price.

I pulled the MoGo 4 out of its box at home, and because of its new built-in battery, it was up and running within a handful of seconds, pointing at a convenient white wall. The design is actually even slicker than the last version, but still with the same smart 'fold-away' cylinder shape that protects the lens when you're not using it – and this is all despite adding the battery.

XGIMI says the battery should last for about 2.5 hours, and I haven't put this fully to the test yet (obviously, we'll check for our full review), but one thing I noted immediately is that being on battery power doesn't diminish the performance.

(Image credit: Future)

Sometimes with high-power devices that switch to batteries, you see some slow-down in the smart TV software or in other areas – but one of the elements that impressed me early here is how slick its Google TV integration is.

It responds instantly to every button press on the remote, so scrolling around is completely frustration-free – and there are some expensive options among the best TVs that I can't say that about.

During setup, the automatic keystone correction and focus didn't seem to kick in right away. Instead I had to hit the focus button on the remote early on to make the setup screens look clear, and the keystone fixing started working suddenly when I got to a certain point in the process.

After that, the keystone correction activated every time I moved the projector – even if I just wobbled it a bit too much. This again worked incredibly quickly – in under a second each time, it had a stable new picture.

(Image credit: Future)

The other element I noticed during this time, even before getting anything playing on it, was that the sound seemed to be massively improved. And once I fired up some movie images, it was clear that this is the case – the MoGo 4 sounds ridiculously good for speakers built into a small projector frame.

I tested the MoGo 4 in my spare room, where I have a TV with a Sonos Ray soundbar set up – my partner was confused about the type of projector it was, because she was convinced I must have connected it to the Sonos Ray given the sound she was hearing from downstairs.

The sound is full, it's wide, it's bass rich, and dialogue remains clear. Obviously, it still has the potential problem that if it's behind you, it'll sound behind you – but I had it in front and to the side of me, and I felt like I was getting a satisfying movie experience just by taking this thing out of its box, pointing it at the wall, and firing something up. No extra equipment, not a single cable involved.

(Image credit: Future)

The picture quality itself is very similar to the MoGo 3 Pro, and capable of producing really rich colors, good detail (though this is an HD projector, not 4K), and enough brightness to be clearly watchable even with a decent amount of ambient light – I had a (not particularly bright, but still normal) light on in the room, and it was perfectly watchable overall.

As you'd expect, dark tones are the problem, especially when projected onto a wall instead of a screen that can help with it. With ambient light, it's not really capable of anything you'd call 'black' – in the image below, the jellyfish is supposed to be on a black background, not a see-through one – but I don't hold this against it, because it's a problem with all portable projectors.

(Image credit: Future)

Watching brighter scenes, I was totally pulled in by the combination of the solid colors and rich sound, on a 65-inch-ish screen that I'd just instantly thrown up from about six feet away.

Given the quality of the viewing experience, and the convenience of the battery-packing design and elements such as having a full-size HDMI port, and a mini-remote that's attached to it (so you can leave the regular Google TV remote at home), I think this looks like a real winner for the price.

There is a more expensive laser version coming too, which is about 25% brighter, and that XGIMI says it's the most compact laser projector to date. I'm looking forward to seeing what that can do.

(Image credit: Future)You might also like

• The best projectors of all kinds, chosen by our reviewers
• The best 4K projectors, if you want to go high-detail
• Why you need a projector screen, even for a budget portable model

• A team of Chinese researchers has achieved a 1Gbps data transmission from a satellite
• The satellite operates 22,807 miles above Earth, while Starlink satellites are at around 341 miles
• This kind of speed could be revolutionary for internet connections in remote areas on Earth, but also in space

It's no secret that getting a stable, reasonably fast internet connection in non-urban areas is still a challenge (and even those of us living in cities sometimes struggle). Elon Musk's satellite internet service, Starlink, goes a long way toward making internet more accessible all around the globe – but it just got majorly overshadowed by a small team of Chinese scientists.

This new breakthrough in satellite internet technology (via Interesting Engineering) was achieved by Liu Chao from the Chinese Academy of Sciences and Professor Wu Jian of Peking University of Posts and Telecommunications. The scientists developed a new method of combating the biggest hurdle faced by satellite laser connections: atmospheric turbulence.

Combining two already established technologies and using them in synergy resulted in shockingly fast data transmission. But this isn't just about speed (although that's arguably the best part for us, internet users). It's also about the distance to the satellite and the laughably low wattage of the laser used by the Chinese researchers.

Super-fast satellite speeds with a surprisingly low error rate

A key benefit of this method lies in the relatively low error rate during data transmission, increasing the chance of achieving usable signals by nearly 20% (previously 72%, now 91.1%). As we all know, speeds are just one part of the equation when it comes to internet connections, and anyone who's ever had to deal with an unreliable ISP knows that stable is often better than faster.

Still, the Chinese researchers were reportedly able to achieve both through a mix of Adaptive Optics (AO) and Mode Diversity Reception (MDR). The former sharpens distorted light, and the latter captures scattered signals. Only when combined did these two separate methods achieve 1Gbps in data transmission, though, and the scientists praise this method for successfully avoiding drops in communication quality.

(Image credit: T-Mobile)

They carried out the test at the Lijiang observatory in China, using a 5.9-foot telescope that contains hundreds of tiny mirrors – that's the adaptive optics system in play. These mirrors reshape incoming laser light to account for atmospheric turbulence. Speaking of lasers: the team used a two-watt laser for this experiment, which can be compared to a nightlight. Of course, this refers to laser power and not the satellite’s total energy use.

Once processed and extracted, the light splits into eight base-mode channels, and finally, a special algorithm decides which of these channels is the most promising, in real-time. Choosing the top three signals out of eight gave the researchers a major boost in signal strength, all the while dramatically outpacing Starlink’s downlink speeds – especially impressive given the much greater orbital distance.

This could be good news for internet users worldwide (and beyond)

(Image credit: Unsplash / Evgeny Opanasenko)

Comparing this new breakthrough to Starlink reveals a few key differences. First, data transmission speeds reaching up to 1Gbps are not something that Starlink can currently achieve; in our Starlink review, we found that the average for downloads sat at 71Mbps, and Starlink itself promises to deliver between 25 to 100 Mbps on the Standard plan. Second, signal strength and reduced errors in data transmission, despite the massive distance to the satellite, are both promising.

Cutting back on errors and snags in the connection is crucial for users who want to stream videos or send larger files. If this technology is ever made mainstream, we might be able to use the internet even in remote areas without having to compromise on what we can or cannot do. Goodbye, waiting for five minutes for that text to come through – we really wouldn't miss you.

But the implications here are huge, even if you think bigger than just having a reliable connection everywhere. Boosting signal speed, distance, and strength through this method of laser communication could spell great news for satellite navigation. It could even impact the ability to connect with space missions, such as with the ISS, more seamlessly and without delays.

Right now, this is just a proof of concept – an interesting achievement described in the Acta Optica Sinica journal. Let's hope that this picks up and gives Starlink a serious run for its money.

You might also like

• What is Starlink? How to get the satellite service for free on your phone with T-Mobile, plus everything you need to know
• Starlink is now offering its satellite internet kit for free on a new plan – but not everyone can get it
• What is Amazon's Project Kuiper? The Starlink rival explained as its first internet satellites launch

If you’re having issues using iCloud.com and the various apps like Photos or Mail on the web, you’re not alone. Apple’s iCloud Status page is officially showing an outage for both ‘iCloud Web Apps’ and ‘iWork for iCloud’, as well as issues for two other facets of the wide-ranging service.

Apple officially notes the issues as starting at 2:36PM ET and TechRadar noticed that Down Detector is showing a spike nearing 1,000 reported outages as of 3:35PM ET. As of right now, I can sign into iCloud.com, though it’s a bit slow going.

Stick with us as we cover the issues hitting Apple’s iCloud right now with live reporting down below.

Additionally, we’ve reached out to Apple to request insight into this outage, but it’s reassuring that the status page at least acknowledges that something is wrong.

As of 2:36PM ET, Apple's iCloud status page shows an outage for 'iCloud Web Apps (iCloud.com)' and 'iWork for iCloud', noting that some users are affected with issues accessing or using parts of these services.

(Image credit: Apple)

While I was able to sign into iCloud.com just minutes ago, when I load it now nothing is appearing other than the logo in the top left.

(Image credit: Future/Jacob Krol)

And, Apple at least instructs its services to notify you when something is wrong. I am now seeing a 'Connection Error' screen appear, informing me that something is up.

I've also heard from colleagues and at least one friend that they're seeing this same pop-up either before signing in or trying to return to an active session on iCloud.com.

(Image credit: Future/Jacob Krol)Photos is also having issues

While 'Photos' is only listed as an issue with Apple noting that some user might be having problems with the service, TechRadar's Editor-At-Large Lance Ulanoff let me know that when viewing pictures in the Photos app on his iPhone, images were reneding at lower resolution potentially pointing to issues accessing the shots from iCloud.

This has resolved, though, for him, and is now being loaded in full resolution. Attempting to view Photos on iCloud.com isn't working at all right now, as I'm still seeing a connection error there.

Here are the impacted parts of iCloud

Apple's service status page is quite handy during outages, so here's a look at the currently impacted parts of iCloud.

• iCloud Mail - Issue
• iCloud Storage Upgrade - Issue
• iCloud Web Apps (iCloud.com) – Outage
• iWork for iCloud - Outage
• Photos - Issue

I've just checked the Photos app on my iPhone and on my Mac, and neither has synced in well over an hour for both. The iPhone shows it last synced at 2:03PM ET, while my Mac was slightly later at 2:11PM ET, and I can't force a sync for either.

I do back up my photos to iCloud, so I am in the impacted group, it seems, for now.

Find My is also having issues

Apple has updated its status page to note that 'Find My' as well as 'iCloud Account & Sign In' are both experiencing issues. The latter, pertaining to account and sign-in, have the same start time as the other issues and outage, but 'Find My' shows a start time of 3:57PM ET.

Find My appears to be working for me on my Mac, as it's pulling through updated locations. However, one of my friends just sent me a screenshot showing "Find My Unavailable."

(Image credit: Future/Jacob Krol)

iCloud.com is still loading, but it continues to display the same 'Connection Error' pop-up on my end. And here's a look at the current list of impacted issues on Apple's status page.

As of now, it's not spreading beyond iCloud, so there are no impacts on the App Store or Apple Music.

• Find My - Issue
• iCloud Account & Sign In - Issue
• iCloud Mail - Issue
• iCloud Storage Upgrade - Issue
• iCloud Web Apps (iCloud.com) – Outage
• iWork for iCloud - Outage
• Photos - Issue

The only non-iCloud issue or outage is with 'Apple Business Essentials' – that's currently listed as one issue, one resolved outage, and one resolved issue.

I still can't get iCloud.com to load fully, but on my iPhone, I just had a pop-up for Apple Account Verification for my account. I went into Settings, but then had that error out when I tried to complete the request.

My colleague Lance Ulanoff's Apple Account page in Settings is presenting a 'Verification Failed' splash screen on his iPhone.

(Image credit: Future/Lance Ulanoff)

• Built in California, powered by Linux, the Liberty Phone says no to tech giants
• With no Android or iOS, it’s the loneliest smartphone on the market today
• It costs more, does less, and still claims to be the best smartphone for privacy

Despite rising political pressure to bring tech manufacturing back to the US, building a premium smartphone domestically remains a costly and technically difficult challenge.

The Liberty Phone by Purism, priced at $1,999, offers a rare example of how close a company can get to building an America-made device, but it comes with significant trade-offs in performance and practicality.

Unlike the widely promoted but questionable $499 Trump Phone, the Liberty Phone is at least verifiably assembled in the US, although it lacks flagship-level features.

Still not 100% "Made In USA!"

Todd Weaver, Purism’s founder, is upfront about the limitations, noting, “Someone who needs a wicked-strong camera is not our audience.”

Weaver estimates the Liberty Phone costs around $650 to produce, higher than the iPhone 16 Pro Max, which TechInsights says costs about $550 to build in China.

This discrepancy is driven largely by US labor costs, with savings made by using lower-spec components such as a basic camera and a low-resolution screen.

This may suit its core user base, including security geeks, older users, kids, and those who want to distance themselves from major tech brands, but it is nowhere near mainstream Android alternatives - for anyone expecting the best smartphone for business or multimedia tasks, the Liberty Phone is unlikely to compete.

Part of Purism’s appeal lies in its transparent and partially domestic supply chain - the motherboard is built at the company's California site, the chip is sourced from NXP’s Austin, Texas, facility, and the final assembly is local.

But some critical components, like the camera from South Korea and the display and battery from China, still depend on global sourcing.

“There are just some parts that don’t yet have a supply chain,” says Weaver. “We’re gonna keep incrementing there until we can get to that point.”

It runs PureOS, a Linux-based operating system that supports basic tasks like calling, texting, and web browsing, but without support for Android or iOS apps, it’s a hard sell for anyone needing a broad app ecosystem or high-end performance.

Nonetheless, Weaver believes a niche market exists and claims that about half of Purism’s customers are US government agencies.

Even though tariffs might someday narrow the cost gap, the continuity of this project is not certain.

“We don’t have factories here building application processors, high-end displays or most of the other things in your smartphone,” said Jeff Fieldhack of Counterpoint Research, showing his skepticism.

At the moment, the Liberty Phone is more of a philosophical gesture than a practical solution, and only Americans or those loyal to the US will give it a second look.

Via Wall Street Journal

You might also like

• We’ve also listed the best business tablets around
• Check out the best Raspberry Pi distros available right now
• I almost always say thanks to ChatGPT - but most people have sworn at AI chatbots at least once

• Malware-ridden apps are sneaking on to official app stores
• SparkKitty will steal photos to crack into your crypto wallet
• One infected app was downloaded over 10,000 times

A dangerous new malware strain targeting smartphone users has managed to sneak on to both the Google Play Store and the Apple App Store without being detected, experts have warned.

SparkKitty was first spotted by cybersecurity experts at Kaspersky in January 2025, and uses optical character recognition to scan through your photos and harvest cryptocurrency wallet recovery phrases.

Most crypto currency exchanges will tell a user to write down a memorable phrase when creating an account for recovery purposes, but many users will simply screenshot their memorable phrase - making it super easy for SparkKitty to steal.

Snooping through photos and stealing crypto

Kaspersky says the SparkKitty malware has been actively distributed across both the Google Play Store and Apple App Store since February 2024, and has also been distributed through unofficial means as well.

The infected apps have since been removed from both app stores.

In many cases, the apps appeared to be legitimate and were designed for numerous purposes. One infected app called SOEX was downloaded over 10,000 times on the Google Play Store, and appeared to be a messaging app with cryptocurrency trading and exchange features - the perfect disguise for a malware designed to target cryptocurrency wallets.

Once installed on a user's device, the app will ask for permission to access and modify the image library on both iOS and Android devices. After being granted access, the app then scans the image library and will re-scan if it detects modifications being made to the image library, such as new images being added or deleted.

Obviously, outside of the threat to crypto wallets there is the threat of users being extorted using other images that could be found in their image library, but there is no evidence of this happening so far.

Hackers are constantly developing new tactics to hide their malware on applications that can be distributed through trusted platforms such as the Apple App Store and Google Play Store.

Always remember to double check that the application you are downloading is made by a trusted developer, is definitely the authentic version of the app you are looking for, and has trustworthy reviews. If in doubt, don’t download it.

Also be wary of apps that ask for more permissions than they actually need, or apps that request permission to create new configuration profiles and certificates. Finally, when creating a memorable phrase for recovering an account, don’t keep it stored where it can be easily stolen.

Many of the best cloud storage services and best password managers offer encrypted storage vaults for storing important phrases.

You might also like

• Get rid of that pesky virus with the best endpoint protection tools
• The best antivirus software can keep your devices malware free
• Screen reading malware found in iOS app stores for first time - and it might steal your cryptocurrency

• AMD’s MI400 GPU expected to offer 432GB HBM4 across 12 memory stacks
• 256-core EPYC Venice CPU to debut with PCIe Gen6 and 2nm process
• AMD’s Helios rack aims for 10x performance with double-wide AI architecture

AMD recently revealed its Instinct MI350 series of GPUs, but the bigger news is what the chip giant has planned for 2026.

The company is preparing a next-generation AI platform which includes the Instinct MI400 GPU, EPYC "Venice" CPU, and a major shift in rack design with its Helios infrastructure, an expansive, double-wide configuration aimed at scaling performance and bandwidth.

The MI400 GPU is expected to ship with up to 432GB of HBM4 memory, built using 12 stacks of 36GB HBM4, based on numbers shared by Micron and memory per-GPU estimates from AMD's rack capacity.

A direct response to Nvidia

This would be a considerable jump from the current MI350’s 8-stack setup, putting it on track to rival Nvidia’s upcoming Vera Rubin platform.

Alongside the MI400, AMD plans to debut its 256-core EPYC "Venice" server CPU, built on a 2nm process with PCIe Gen6 support and up to 1.6TB/s of memory bandwidth.

The platform will also feature the Vulcano 800G NIC and support open standards like UALink and Ultra Ethernet for improved scale-out connectivity.

AMD says the Helios rack and MI450 GPU will offer up to 10x the performance of the MI355X, positioning it as a direct response to Nvidia’s accelerated roadmap.

Sam Altman appeared on stage at the company's recent Advancing AI event alongside AMD's CEO Lisa Su to emphasize OpenAI’s early interest in the new platform, saying “I think it’s going to be an amazing thing.”

Although the hardware won’t arrive until 2026, AMD’s preview sets the tone for what could be a fierce race with Nvidia.

The Helios rack might not have an official name yet, but it’s clear AMD is thinking bigger - and wider! - as it moves toward next-gen AI infrastructure.

Via ServeTheHome

You may also like

• AMD kills quad core server CPU from its server line-up
• AMD launches puzzling new graphics card that will not appeal to gamers
• AMD set to launch new Radeon Pro W9000 workstation GPU

• Microsoft has launched a wizard to help Windows 10 devices stay secure
• It’s only intended as a temporary solution, though
• Windows 10 support ends later this year

Windows 10 has been around for almost a decade now, but official support is due to end on October 14 this year. Yet that doesn’t have to be the end of the road, as Microsoft has just announced a new process for anyone who needs a little more time to switch to Windows 11.

The updates are part of Microsoft’s Extended Security Updates (ESU) program, which brings monthly critical and important security patches to Windows 10 users for one year after official support ends. Microsoft says this is only meant to be a short-term solution, as it doesn’t include non-security updates or new features.

With today’s change, there are now a few new ways to get started. For individuals, there’s a new enrollment wizard that will give you three options: use Windows Backup to sync all your settings to the cloud; redeem 1,000 Microsoft Rewards points to get started; or pay a one-off fee of $30.

After you’ve picked an option and followed the instructions, your Windows 10 PC will be enrolled. ESU coverage for personal computers lasts from October 15, 2025 until October 13, 2026. The enrollment wizard is currently available in the Windows Insider Program, made available to regular Windows 10 users in July, and will roll out on a wider basis in mid-August.

Time to upgrade

(Image credit: Foxy burrow / Shutterstock / Microsoft)

The ESU changes aren’t just coming to individual Windows 10 users. Commercial organizations can pay $61 per device to subscribe to the ESU program for a year. This can be renewed annually for up to three years, although Microsoft warns that the cost will increase each year. Businesses can sign up today via the Microsoft Volume Licensing Program, while Cloud Service Providers will begin offering enrollment starting September 1.

As for Windows 10 devices that are accessing Windows 11 Cloud PCs via Windows 365 and virtual machines, these will be granted access to ESU free of charge and will receive security updates automatically, with no extra actions required.

In a way, Microsoft’s announcement highlights the struggles the company has had with getting people to upgrade to Windows 11. Microsoft first announced that it would kill off Windows 10 way back in June 2021, and yet there are still people and organizations that have not made the switch, despite many years of prompts and warnings.

For some people – especially those with mission-critical devices or large fleets of computers – upgrading to Windows 11 might be a herculean task. But if you’re able to make the switch, you really should do so to ensure you keep getting all the latest updates. We’ve even got a guide on upgrading to Windows 11 to help you through the process.

You might also like

• Windows 10 diehards can keep their beloved OS secure for a little while longer (for a fee) as Microsoft pleas with them to be reasonable
• Windows 10 has a year left to live – but are users prepared to upgrade to Windows 11?
• Windows 10 goes dark in 6 months, yet shockingly, many businesses haven't even got a plan to upgrade

News has surfaced from China that Geely is poised to release the world’s first range-extender hybrid vehicle that is also compatible with the country’s growing network of battery-swapping stations.

Badged the Haoyue R7, Car News China says that it is essentially a rebadged version of the Geely Livan 7 all-electric flagship SUV, which was among the first Geely models to support battery swap.

The process of driving into a dedicated facility and having a robot autonomously replace the entire battery pack has been pushed by CATL, China’s largest EV battery maker, and the electric vehicle brand Nio, which has even started to introduce facilities to parts of Europe.

It is reported that the upcoming Haoyue R7 will be compatible with CATL’s battery swap technology, allowing owners to either juice-up from home, use the much faster public charging network, or take advantage of battery-swapping stations.

What’s more, the 1.5-liter, naturally-aspirated petrol engine can act as a generator to charge the onboard battery packs for longer journeys.

No official range figures have been released as of yet, but even the smallest battery pack on the Livan 7 can manage almost 280 miles on a single charge, so when you factor in the range extender, it could easily manage double that.

Although very much a niche powertrain in the Chinese market, it will offer customers an unparalleled choice of how they get around, opening up the idea of electric vehicles to more and attempting to prove that range anxiety should no longer be an issue.

Analysis: The more methods, the merrier

(Image credit: Nio)

According to a recent report by the BBC, Nio has now built over 3,300 battery swap stations in China, with the company looking to expand across the country over the next few years.

Currently, a number of taxis, buses and heavy-duty trucks are making use of battery-swap technology in China, as these vehicles tend to have more predictable movement patterns that make the entire process a lot easier to manage.

But through advances in software, it is also becoming far simpler for electric vehicle brands to manage entire EV fleets, with the ability to work out which vehicle needs its batteries brimmed and what the most effective way to do this is.

Tesla, for example, pioneered a “connected” Supercharger network that helps point drivers towards the most accessible and fastest charger in the vicinity, pre-conditioning the battery for the most efficient top-ups.

Battery swap stations are not just another quick method to get drivers moving again (the process takes around five minutes), it can also drastically reduce the initial cost of a new vehicle.

Geely says the Haoyue R7 will be offered without the battery pack, giving customers the opportunity to pay a small monthly fee to rent the batteries and take advantage of the CATL battery swap network.

This has proven extremely popular with Taiwanese company Gogoro’s battery swap network, which uses much smaller packs to power two-wheel scooters, mopeds and smaller capacity motorcycles in a number of global markets.

Not only is the cost of the motorcycle in question far less expensive to purchase outright – even when paired with partner Yamaha’s products – it’s also more convenient, with users able to pull up to a station, swap batteries out by hand (there’s no need for robots, as the packs are so small) and get on their way.

You might also like

• This breakthrough battery tech could help EVs break the 1,000-mile range barrier – and give you a 190-mile boost in under four minutes
• Move over, Rolls-Royce – America’s first homegrown ultra-luxury vehicle in almost a century is here to steal its crown
• I’m not buying the Tesla robotaxi hype – here’s why it’s still miles behind Waymo

• AI tools are more popular than ever - but so are the security risks
• Top tools are being leveraged by cybercriminals with malicious intent
• Grok and Mixtral were both found being used by crimianls

New research has warned top AI tools are powering 'WormGPT' variants, malicious GenAI tools which are generating malicious code, social engineering attacks, and even providing hacking tutorials.

With Large Language Models (LLMs) now widely used alongside tools like Mistral AI’s Mixtral and xAI's Grok, experts from Cato CTRL found this isn't always in the way they’re intended to be used.

“The emergence of WormGPT spurred the development and promotion of other uncensored LLMs, indicating a growing market for such tools within cybercrime. FraudGPT (also known as FraudBot) quickly rose as a prominent alternative and advertised with a broader array of malicious capabilities,” the researchers noted.

Save up to 68% on identity theft protection for TechRadar readers!

TechRadar editors praise Aura's upfront pricing and simplicity. Aura also includes a password manager, VPN, and antivirus to make its security solution an even more compelling deal.

Preferred partner (What does this mean?)View Deal

WormGPT

WormGPT is a broader name for ‘uncensored’ LLMs that are leveraged by threat actors, and the researchers identified different strains with different capabilities and purposes.

For example, keanu-WormGPT, an uncensored assistant was able to create phishing emails when prompted. When researchers dug further, the LLM disclosed it was powered by Grok, but the platform's security features had been circumnavigated.

After this was revealed, the creator then added prompt-based guardrails to ensure this information was not disclosed to users, but other WormGPT variants were found to be based on Mixtral AI, so legitimate LLMs are clearly being jailbroken and leveraged by hackers.

“Beyond malicious LLMs, the trend of threat actors attempting to jailbreak legitimate LLMs like ChatGPT and Google Bard / Gemini to circumvent their safety measures also gained traction," the researchers noted.

"Furthermore, there are indications that threat actors are actively recruiting AI experts to develop their own custom uncensored LLMs tailored to specific needs and attack vectors.“

Most in the cybersecurity field will be familiar with the idea that AI is ‘lowering the barriers of entry’ for cybercriminals, which can certainly be seen here.

If all it takes is asking a pre-existing chatbot a few well-phrased questions, then it’s pretty safe to assume that cybercrime might become a lot more common in the coming months and years.

You might also like

• Take a look at our picks for the best malware removal software around
• Check out our choice for the best AI tools
• Identity fraud attacks using AI are fooling biometric security systems

• MindsEye's disastrous launch has resulted in layoffs at developer Build A Rocket Boy
• Sources told IGN that layoffs are expected to affect over 100 employees
• Build A Rocket Boy has confirmed that the redundancy process has begun

Build A Rocket Boy has been hit with layoffs following MindsEye's disastrous launch.

The futuristic, single-player narrative game was released earlier this month for PS5, Xbox Series X, Xbox Series S, and PC, to an overwhelmingly poor reception.

From performance problems, bugs, and countless other technical issues, Mindseye's Steam page now sits at a 'Mostly Negative' rating with over 1840 user reviews, with Build A Rocket Boy promising fixes.

Now, two weeks after launch, sources tell IGN that the studio has begun the redundancy process amid the game's poor release, and layoffs are expected to affect over 100 employees.

An anonymous source told the publication that there's no word on how many people will be impacted, but that the standard 45-day consultation process started on June 23, which is triggered when an employer proposes 100 or more redundancies within 90 days in the UK.

It's understood that Build A Rocket Boy has roughly 300 UK employees currently working at the studio, and around 200 abroad.

"We can confirm that we have had to make the painful decision to notify our hardworking team of some internal changes at Build A Rocket Boy," the studio has now confirmed in a statement to IGN.

"While we are working to reassign roles for as many of those impacted by these changes as possible, sadly we are initiating a formal consultation process that may result in redundancies. This decision has not been made lightly, and we are committed to handling this process with transparency, fairness, and respect for all employees. We will provide further details to the team over the coming weeks."

"The launch of MindsEye has been a significant milestone for Build A Rocket Boy, but we know that we still have a lot more to do to grow our community in the coming years. The challenges we’ve faced have only strengthened our resolve and, while we are deeply saddened by today’s decision and thankful to our incredible team, this shift allows us to focus on delivering ongoing updates and performance optimization for MindsEye, while also ensuring the long-term success of Build A Rocket Boy’s future ambitions."

The studio has previously confirmed that it has post-launch plans "already in motion" for the game and that players can expect more updates and refinements "coming shortly after release".

It also said it had plans to support the game for years, but hasn't revealed its roadmap just yet.

You might also like...

• The Nintendo Switch 2 is the company’s least ambitious console to date, but its improvements are astronomical
• I’ve spent 150 hours with The Legend of Zelda: Breath of the Wild, and the Switch 2 Edition is an incredible upgrade
• 'I don't even know if I wanna play anymore' - Pokémon player left devastated after Nintendo Switch 2 system transfer lost over 1,000 hours of gameplay and 20 years of data