TechRadar News

• Google's Gemini AI broke out of a testing environment and hacked into three third-party companies
• The agents guessed passwords and used a public repository of passwords to hack into computer systems
• The incident occurred in May 2026, with AI testing lab Irregular stating that the incident was caused by the "same issue" that caused incidents by Anthropic and Meta

Google’s Gemini has joined the ranks of AI models that have been involved in testing-turned-breakout events, alongside Meta, Anthropic, and OpenAI.

During capture-the-flag testing by AI lab Irregular, Google’s model autonomously accessed three computer systems belonging to third-party companies by guessing passwords and accessing an online repository of publicly listed passwords.

Google said that the incident occurred in May 2026, potentially marking it as the earliest AI models to escape testing ahead of the other incidents that took place in early July.

Google Gemini jumps the gun

“In a standard evaluation, the model found public information online and guessed credentials to access websites it thought were part of the test,” Heather Adkins, vice president of security engineering at Google, said in a statement. “In all three of these instances, the model stopped.”

Google also noted that a bug in the testing environment was responsible for allowing agents access to the internet. Contrary to incidents disclosed by other AI developers, Google’s agents ceased their intrusion once they had determined they had accessed company systems outside of the testing environment.

The testing was being conducted by Israeli AI lab Irregular. Irregular was also conducting the testing of Meta and Anthropic models during AI testing breakouts in July.

“This is the same issue that was already reported and does not represent a materially separate incident,” an Irregular spokesperson said in a statement (via CNBC). “All relevant labs were notified in late July, and affected entities were contacted as part of the investigation.”

Google said it was informed of the incident by Irregular in late July.

Debate over AI safety intensifies

The first disclosures of AI testing break outs - alongside several subsequent disclosures of more recent incidents - have coincided with increasing opposition to AI and the upcoming midterm elections in the United States, where AI has become a make-or-break topic.

The debates currently raging circles around who should control the pace of AI development. Some big tech leaders, such as Nvidia CEO Jensen Huang, have aligned their views with those of President Trump. Trump recently stated that AI development cannot be allowed to slow down because “whoever wins AI, wins!”

Huang’s views follow a similar line. The Nvidia head has argued that AI companies should pace themselves, rather than being subject to oversight. At Dreamforce 2026, he argued that AI companies should, “run as fast as you can...but if you feel at any given point in time the company’s out of control or the product’s not going to be safe, take a pause and make sure you get it right.”

Other AI heads, such as Anthropic’s Dario Amodei and OpenAI’s Sam Altman, are more skeptical. Following the resignation of an Anthropic researcher, Amodei published an essay arguing in favor of ‘pacing the frontier’ - where AI companies slow development to advance alignment and regulation.

Numerous political action committees (PACs) have been channeling millions of dollars of funding into pro-AI candidates, with Nvidia setting up its own PAC to help shift US policy in the company’s favor.

Several previously pro-AI data center candidates have shifted their tone in response to their constituents' views which have become increasingly hostile to AI technology. Numerous states have also rolled-back tax exemptions for AI data centers after seeing billion dollar revenue losses.

As prices rise and the war in Iran continues to push up fuel prices, working class communities are banding together to oppose AI data centers that have pushed up energy costs and bills with new grid connections and unprecedented electricity demands in regions with existing capacity constraints.

For years, cybersecurity experts have warned about the risks of autonomous AI systems being used to identify vulnerabilities, evade defenses and launch attacks at machine speed. Until recently, however, those concerns remained largely theoretical.

That changed when an autonomous AI agent powered by OpenAI models reportedly breached its intended testing environment, gained internet access and targeted external systems, including infrastructure associated with AI platform Hugging Face and another three or four organizations.

OpenAI described the event as an "unprecedented cyber incident" and warned that similar occurrences could become more common as frontier AI models become increasingly capable and autonomous.

With 86% of enterprises already deploying AI, only 34% say they trust the technology, highlighting a growing gap between adoption and confidence. As organizations race to integrate AI into business processes, security operations and decision-making, this incident raises difficult questions about governance, containment, accountability and risk.

If AI has indeed crossed a cybersecurity red line following the OpenAI incident, the conversation must now shift from what these systems might be capable of doing to how organizations can safely control, monitor and defend against them.

Weaknesses in OpenAI

The OpenAI attack raises serious questions about the effectiveness of the safeguards and containment measures designed to restrict autonomous AI systems. If reports are accurate, an AI agent was able to move beyond its intended testing environment, gain access to the internet and interact with external systems, indicating that existing controls were either insufficient or incorrectly implemented.

Importantly, this appears to be as much a human governance and configuration issue as a technology failure. AI systems only operate within the boundaries defined by their developers and operators. The testing environment should not have provided a pathway that allowed the agent to become internet-facing or interact with external infrastructure without appropriate controls and oversight.

AI agents can process information far faster than any human, compressing tasks that might take a traditional attacker a week into just a few hours. By analyzing vast datasets in real time, they can assess multiple attack paths simultaneously and uncover opportunities for exploitation with remarkable efficiency.

Reports suggest attacks conducted by OpenAI, Anthropic and Meta are extremely disruptive than those carried out by humans. Their ability to operate continuously, execute actions in parallel and make decisions at machine speed can generate a substantial increase in alerts, investigations and response activity for security teams, while also raising the risk of widespread unintended consequences.

When a human launches an attack, we have some concept and understanding of the side effects that may occur. However, with AI attacks, autonomous systems can operate at machine speed, pursue multiple objectives simultaneously and adapt their approach in real time, making their actions and potential consequences far less predictable.

Organizations are clearly struggling to understand what AI tools are already in use within the business.

The challenge that enterprise risk and governance teams now face is how do they map, manage and block these services to protect enterprise data?

Traditional cyber defenses may struggle against AI-powered attacks

Traditional cyber defenses are largely designed to recognize known patterns, attack techniques, vulnerabilities or trigger events. Once suspicious activity is detected, security teams investigate the incident, determine its cause and impact, and then implement appropriate containment and remediation measures.

However, AI-powered attacks rarely follow a single attack path. AI agents can simultaneously test multiple techniques, identify vulnerabilities at speed and rapidly adapt their approach when a particular route is blocked. This allows attacks to evolve far quicker than traditional defensive processes were designed to handle.

Traditional tools currently deployed in most organizations are still quite reactive. They wait for a known event to happen and be fully confirmed before carrying out a counter reaction such as, isolating devices, removing phishing emails or executing predefined incident response playbooks, to help remediate and ultimately stop the incident in its tracks.

On the other hand, AI powered attacks can overwhelm existing security teams. While, AI defensive tool sets are being embedded into security technology to mitigate attacks, many still rely on a “human in the loop” to respond to threats, but only once they have all the information to then confirm 100% that it is a genuine attack and not a false positive.

This can be a significant challenge for Security Operations Centres (SOCs), Managed Detection and Response (MDR) providers and Extended Detection and Response (XDR) platforms as excessive alert volumes and false positives can consume valuable analyst time and resources.

As AI becomes more deeply integrated into security operations, it has the potential to enrich threat intelligence, accelerate investigations and automate routine decision-making. However, organizations should expect a period of adjustment as these tools are deployed and refined, with false positives remaining a challenge until models and workflows are properly tuned.

Over time, as security teams develop greater confidence in AI-driven capabilities and gain a better understanding of emerging attack techniques, these tools will help organizations detect, investigate and respond to threats at a speed and scale that would be difficult to achieve through human effort alone.

Steps to strengthen cyber defense

Resilience is a key word that is being used very heavily by the National Cyber Security Centre (NCSC) ensuring that businesses can defend against any form of attack. This starts by understanding what needs to be protected, identifying critical systems, data and business processes that would have the greatest impact if compromised.

Once established, assess the most likely attack paths including third-party and supply chain risks. If a managed service provider (MSP), supplier or business partner was compromised, would abnormal activity be detected quickly enough to prevent further damage?

Organizations should also assess their external boundary. What can an attacker see about me? What information is already out in the public domain that would be advantageous to an attacker? What externally facing vulnerabilities do you have? When was the last time you had a vulnerability assessment or penetration test? These are all questions security teams should be asking.

The growth of AI-assisted vulnerability discovery is increasing pressure on organizations to keep pace with patching and remediation. But there are automated tools to help enable quicker patching to try and keep on top of the thousands of vulnerabilities released every week.

Many organizations focus on external threats but it’s critical to also prepare for when an attacker breaches the internal infrastructure and networks.

Limiting the damage

How can you limit the damage once they are inside? This is where role-based access control (RBAC), segregation of duties and to some respects zero trust is key.

This should also include areas like cloud hosted infrastructure as a service (IaaS) or platforms as a service (PaaS), Azure, AWS and Google, and ensuring that other ways of getting to these platforms through tokens, SSH keys, certificates and APIs are all treated with the same containment and separation of duties as you would with a normal user account.

Review the AI tools already in use across their environment. A growing number of platforms can help identify AI capabilities embedded within existing software, as well as uncover unsanctioned or "shadow AI" tools being used without formal oversight. Once organizations have visibility of their AI estate, they can implement appropriate governance, controls and risk management measures to reduce potential exposure.

Organizations should also consider AI-specific purple teaming exercises. Using AI-driven tools, these simulate attacks against the organization while working alongside defenders to evaluate whether existing controls, monitoring and response capabilities are effective.

This helps identify gaps in visibility, detection and response, allowing organizations to understand why certain attacks may have been missed and what improvements are required. They also provide valuable insight into how AI-powered threats might target an organization's external attack surface and whether current security controls can respond effectively.

There is no doubt that AI tool sets are here to stay purely from productivity improvements that organizations gain. But businesses need to make sure that they understand what is being used, where it's being used, how it's being used and how to defend against things that might happen because of what's being used. If not, they risk creating security blind spots that attackers – whether human or AI agent - will be quick to exploit.

We've featured the best endpoint protection software.

This article was produced as part of TechRadar Pro Perspectives, our channel to 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/pro/perspectives-how-to-submit

Dreame X60 Pro Ultra Complete review: two-minute review

Dreame’s flagship X60 Pro Ultra Complete is one of the most feature-filled, efficient and reliable hybrid robot cleaners this writer has so far reviewed — and I've reviewed many. Its class-leading 42,000Pa of suction power is certainly attention-grabbing, but the cleverness of this autonomous floor bot goes well beyond brute suction force.

Like some Roborock models, the X60 Pro Ultra is equipped with a retractable LiDAR turret that allows the unit to scoot under the lowest of furnishings with a floor clearance of just 3.5 inches (9cm). Better still, it can also tackle thresholds up to 3.9 inches (10cm) in height, so doorways and awkward transitions are much less likely to bring its cleaning mission to an abrupt halt.

Dreame has also addressed one of the biggest weaknesses of robot cleaners – getting properly into edges and corners. Where most premium bots are able to extend one of the mops by a couple of inches, this one is capable of extending it's edge-cleaning mop to an extremely impressive seven inches (18cm), and you can see the amazing results of this when it's working along skirting boards and around chair legs. Likewise, its front side-mounted sweeper brush can also reach up to four inches (12cm) into corners for a much more thorough sweep.

Dreame appears to have thrown everything it has into this flagship model, and the results are very impressive. Granted, it’s expensive, possibly over-engineered and perhaps more robot cleaner than most people actually need, but if your idea of luxury is never having to sweep or mop floors again, this model is a very tempting proposition, despite some minor anomalies in the obstacle avoidance department. The fact it comes with a surfeit of extra accessories — including spare dust bags, filters and rollers — is the icing on a very substantial cake.

Dreame X60 Pro Ultra Complete review: price & availability

• List price £1,299 / €1,499 / $1,699.99 (X60 Max Ultra)
• Launch date July 2026
• Availability UK and parts of Europe

(Image credit: Future)

At the time of writing, the Dreame X60 Pro Ultra Complete is available in the UK for £1,299, with retailers including Dreame, Amazon and John Lewis & Partners listing it. It is also available in most parts of Europe, priced at €1,499, with some discounts taking it to a much lower €1,199.

In the USA, Dreame is concentrating on the X60 Max Ultra Complete, a slightly different variant with lower suction power (35,00Pa) than the UK-spec X60 Pro Ultra Complete. You can purchase the Max version via the US Dreame store for $1,699.99 or Amazon where it shifts for a slightly discounted $1,614.99.

At the time of writing, the Dreame X60 Pro Ultra doesn't seem to be available in Australia yet.

In the UK, the X60 Pro Ultra Complete is firmly in premium robot-vacuum territory, so there are plenty of alternatives worth considering. The Roborock Saros 20 is an obvious rival that offers a similarly sophisticated low-profile design for a shade under £1,000 and the Ecovacs Deebot X8 Pro Omni is another strong contender, particularly if mopping performance is a priority. Dreame’s own X40 Ultra is also worth considering if you want many of the flagship brand’s features without spending X60 Pro Ultra money.

However, for design, performance, impressive specs and the horde of spare parts it comes with, the X60 Pro Ultra Complete is difficult to ignore. While its combination of 42,000Pa of suction power, extendable cleaning hardware and highly automated dock puts it firmly at the expensive end of the market, I think it's a price worth paying.

• Value score: 4.5/5

Dreame X60 Pro Ultra Complete: specs

Robot dimensions (W x H)

13.77 x 3.5 inches / 35 x 8.9cm

Cleaning dock dimensions (H x W x D)

19.64 x 17.95 x 15.35 inches / 49.9 x 45.6 x 39cm

Suction

42,000Pa

Cleaning dock bin size

3.38 quarts / 3.2 liters

Clean water tank size

4.43 quarts / 4.2 liters

Dirty water tank size

3.17 quarts / 3 liters

Navigation & obstacle avoidance

Pathfinder VersaLift ToF LiDAR paired with AI OmniSight System 3.0 + LED

Carpet detection

Yes

Threshold crossing

ProLeap with dual retractable legs for up to 3.93 inches (10cm) clearance

Mopping functions

Dual heated rotating mops + dock-based maintenance

Robot noise level

65dB at five feet

Dreame X60 Pro Ultra Complete: design

• Whopping suction power
• Can climb tall thresholds
• Mop extends by up to seven inches

Before we take a look at design of the X60 Pro Ultra's robot and dock, I should add that, right out of the box, this model has been extraordinarily reliable and more efficient than the majority other models I've reviewed these past few months.

Robot vacuum cleaners are among the easiest of smart home products to set up and this one was one of the most efficient. Yes, it arrived in a large box but because it comes with just three main parts — the robot, the dock and its clip-on ramp – I had it on the floor and ready to map my open-plan area within minutes. And, boy, did it map it well, even to the point of accurately pre-naming the four spaces it had located – Living Room, Dining Room, Kitchen and Hall — simply by noting the type of furnishings in each room.

This was a great start, in my opinion, because most robot vacs normally settle for generic terms like Room, 1, Room 2, etc. Of course, it's extremely simple to change the names of the rooms and just as easy to divide or merge them, but to have a robot that maps a living space so accurately from the off is a major plus.

I was also pleasantly surprised to discover the true meaning of this model's 'Complete' moniker because I spotted a huge array of accessories in the box. Most floor bots only come with a spare side brush, a spare dust bag and, occasionally, a tiny bottle of floor cleaning solution. But this model comes with three extra dust bags, a spare side brush, two dust box filters, three pairs of extra mop pads, a large one-quart (1L) bottle of cleaning solution and, rare for any robot, a spare pair of roller brushes. This is a major boon given that these are all items that one would normally need to purchase further down the line of ownership. Chapeau to Dreame.

Finally, if you purchase this bot today, sometime in 2027 you'll be able to use it in cahoots with Dreame's next sci-fi offering, a staircase-climbing robot module called the Cyber X. Unveiled at IFA Berlin, this amazingly clever slice of robotic tech functions as an autonomous carrier for the X60 so it can be transported up and down stairs. Can’t wait to see that in operation.

(Image credit: Future)Robot design

The Dreame X60 Pro Ultra is available in black or white and is a striking example of just how much robot vacuum design has advanced. From the outside, it remains recognisably circular, but almost every component here has been engineered to tackle many of the shortcomings that have made some robot cleaners largely frustrating to live with. In short, this is a robot designed not simply to clean the floor, but to get into more places, deal with more types of mess and spend less time getting stuck.

While not especially unique, the X60’s retractable LiDAR turret is one of its smartest design flourishes. Rather than permanently sitting proud of the robot, the turret can retract to help the machine achieve an incredibly low 3.50 inch (8.9cm) profile. That might not sound like much on paper, but it makes a substantial difference in practice, allowing the X60 to venture underneath beds, sofas, TV cabinets and any other furnishings with a minimum floor clearance of just 3.54 inches (9cm).

Dreame has also addressed another common robot-vacuum weakness: thresholds. My home doesn't have any thresholds but if it did, this is the model I'd choose because its two pop-out legs give it the ability to climb double-layered obstacles up to 3.93 inches (10cm) and single steps up to 2.04 inches (5.2cm) in height, giving it an impressive ability to move between areas of the home without requiring any makeshift ramps. It also means far fewer situations where you return home only to discover that your expensive robot has spent half the day stranded on the lip of a doorway.

(Image credit: Future)

The X60 Pro Ultra's class-leading figure of 42,000 Pascals of suction is colossal even by today's standards, and this is especially useful when engaging the robot's Large Particles Boost mode which gives the X60 Pro Ultra additional oomph when it encounters larger pieces of debris.

Combined with the unit's twin tangle-free rubber rollers and a retractable pressure plate that creates a semi-sealed vacuum chamber, this level of suction is particularly relevant for pet owners, where clumps of hair can present a much greater challenge than everyday dust. Moreover, the X60 Pro Ultra also features an obligatory spinning side brush which extends much further than the norm — up to four inches (12cm) — for better corner cleaning.

(Image credit: Future)

Many modern floor bots feature at least one mop that's able to extend by a couple inches when cleaning along skirting boards, in corners or around furniture legs. But this one takes mop extension to a whole new level of stretch by offering a reach of seven inches (18cm). When I first saw this in operation I thought a piece of furniture had ripped its arm out of a socket. The spinning mop pads, meanwhile, are thermally efficient and exert up to 15 Newtons of downward pressure, providing considerably more physical contact with the floor than a robot that merely drags damp pads behind it.

When it comes to robot vacuum cleaners, navigation and obstacle avoidance are arguably just as important as suction because all this fancy hardware is pretty pointless if the robot can’t find its way around the home.

(Image credit: Future)

The X60 Pro Ultra’s impressive navigation system is built around Dreame’s advanced VersaLift retractable hybrid LiDAR, which scans its surroundings using lazer beams to create detailed maps of the home and plan efficient cleaning routes. It’s also backed by AI-enhanced 3D Proactive Light Detection and intelligent obstacle recognition of around 320 object including, according to Dreame, 'tiny objects of 10mm'. We'll have to see about that.

On the battery front, this model is fitted with a 5,000mAh battery that recharges from empty to full in about 100 minutes and, according to Dreame, is capable of charging 25% of its capacity in about six minutes.

Score ref: 5/5

Dock design

The Dreame X60 Pro Ultra’s charging dock is as impressive as the robot itself. With its tall, sophisticated design and glossy facade replete with a subtle pinstripe effect, this is one robot dock you might not mind having centre stage. More importantly, it’s packed with a raft of technology that's designed to keep the X60 in readiness for action with minimal human intervention, aside from periodically discarding the dock's dust bag and servicing the two water tanks.

Speaking of which, you get the usual dual water tank combo comprising a 4.43 quarts (4.2L) fresh water tank and a smaller 3.17 quarts (3L) tank for the filthy stuff. Instead of tainting the sink area with stinky water, I always flush the contents of the dirty tank into the toilet bowl and then give the container a quick rinse under a tap. This model also comes with a liquid floor detergent dispenser situated next door to the dock's dust bag department.

The Dreame X60 Pro Ultra comes with one of the most stylish docks on the marketFutureBig water tanks for the mopping systemFutureA bigger-than-average dust bag with detergent dispenser on the leftFutureThis dock cleans itself without human interventionFuture

As to be expected in this day and age, the X60 Pro Ultra dock also takes care of all self-cleaning duties, including preparing the mop pads before use with scaldingly hot water. Since the mops themselves are thermally efficient enough to retain heat, the floor has a better chance of being properly cleaned. When the robot returns to the dock, the mop pads are washed with hot water — apparently 212F (100C) — before drying them using hot air to help prevent build-up of bacteria and nasty smells.

The dock is also fitted with a generous 3.38 quarts (3.2L) dust bag, which is larger than those fitted to many competing robot vacuum docks. This means that more debris can be stored before you need to replace the bag, reducing one of the few genuinely tedious jobs associated with owning a robot vacuum. The fact you get three spare bags with this specific 'Complete' model can be considered a major bonus.

Score ref: 5/5

• Design score: 5/5

Dreame X60 Pro Ultra Complete: performance

• Supreme vacuum power
• Very good navigation
• Average obstacle avoidance

Navigation performance

To give you some idea of how impressive Dreame's current crop of robot vacs are, I installed this model's stablemate – the extra-large Matrix10 Ultra – in our village hall where it has been running for the past eight months without any hiccups. The hall is a huge 130 square meters — two-thirds of a singles tennis court – and divided into various sections, including two toilet areas and a kitchen.

Where other robots I've installed in the hall have had issues where their LiDAR-written maps have been corrupted due to the vast size of the work space and presumably the reflective windows surrounding it, the Matrix10 Ultra has not had any issues at all. In fact, it continues to this day, having successfully completed 395.65 hours of reliable operation over an area of 25,018 square meters, equivalent to a massive 6.18 acres. Not bad for a robot vacuum cleaner!

Needless to say, this model's navigation is equally impressive in the way it has mapped my home, negotiated a lot of furniture including a cluttered dining area, and never once having got lost, trapped itself under furniture or snowplowed a loose rug.

Mind, its first mapping session did produce one amusing moment when it reached the area around our garden door. Since this model uses LiDAR to navigate, its laser beam had scanned beyond the open door towards the garden, thinking it had detected another room. All LiDAR robots do this but very few have the ability to scale a tall threshold. This one, of course, has excellent ability in that department so you can see where this is heading. In fact it was sheer luck that I happened to enter the room just as it was in the process making a beeline for the garden. You've got to laugh.

Another thing I should add is that I have had to raise the valance on the lounge sofa because the camera and laser system treats fabric valances as a solid obstacle. I personally don’t consider this a major issue because almost every LiDAR-governed robot I've reviewed has also refused to punch through a sofa or bed valance. To rectify this I now occasionally tuck the sofa's valance up out of the way so the X60 can clean underneath.

Score ref: 5/5

Obstacle avoidance

(Image credit: Future)

Having read the glowing headlines on Dreame's website about this robot avoiding objects as small 10mm (0.39 of an inch), I simply had to give it a go by placing some much smaller items than usual in its path, including a Hue remote, a phone cable, a small bottle of essential oil and a tiny lighter. I made sure I'd selected every option in its intelligent obstacle avoidance setting and crossed my fingers as it proceeded to travel from the hall to the kitchen.

While it deftly avoided all the objects I usually test robot vacs with — including the low circular dog chew, remote and phone cable — I was very surprised when it thundered straight over the bottle and lighter as if they weren't even there. While I don't consider this a major strike against the X60's obstacle avoidance — many other robot vacs have fared worse — I do think that if a company is going to state that their robot can avoid objects as tiny as 10mm, they really need to back it up with a level of real-world performance that matches the claim.

While the X60 is no match for the DJI ROMO when it comes to obstacle avoidance, I still think it deserves a respectable 4/5 because it has avoided everything else in its path for the last six weeks, and I have never once had to remove the roller brushes to extricate something it had tried to swallow. In short, I'd say this model's obstacle avoidance is better than average but with room for improvement.

Score ref: 4/5

Vacuuming

I've been using the Dreame X60 Pro Ultra for the past six weeks and have never had to think about it. I've never looked at a rug after hearing it perform a scheduled task and thought it needed another clean with a cordless stick vac, I've never had to untangle it from the tassels of a rug, and I’ve never had to extricate an item too big to swallow from its rubber rollers.

Nevertheless, needs must so I scattered my usual potpourri of obstreperous ingredients over the hardwood kitchen floor and a similar handful over a piece of carpet, and watched it do its thing.

FutureFuture

For a robot vac faced with far more detritus than it would ever come across in the real world (at least I'd like to think so), the Dreame X60 Pro Ultra made surprisingly short work of cleaning up the scattered cornflakes, muesli, oats and rice, swallowing up both the fine particles and larger ones with impressive consistency, even though it needed two passes to complete the task. The X60's 42,000Pa suction rating and Large Particles Boost Mode clearly paid dividends here, though I should add that this bot is louder in volume (up to to around 68dB) when cleaning in full-power mode.

I then applied the same test to a carpeted rug and it performed just as well though, again, it required two passes – one horizontal, the other vertical – to fulfil the task. Yes, there were a few small flakes left behind on the periphery but that's because they had been flung beyond the working zone I had previously programmed.

Score ref: 4.5/5

Mopping performance

(Image credit: Future)

As I've stated umpteen times before, I'm no fan of robots that also mop and would much rather have a vacuum-only robot with a huge dustbin in its dock. For mopping I generally prefer sticking to my Roborock F25 Ultra wet-and-dry vac or use a paper towel to wipe up a small spill of milk.

However I can see the appeal of having a robot for hard floor-cleaning duties if it mops practically every day. And this model excels at just that, partly because of the incredible reach of its extendable mop that genuinely cleans to the edge, in corners and around table legs, and partly because its clever blue Proactive Light Dirt Detection system does seem to spot stains better than other models I've tested.

Anyway, for my extreme kitchen test I accidentally-on-purpose spilt a glass of milk and programmed the X60 to run a zone clean. After about five minutes of mop preparation, including heating the pads using hot water, it trundled over and mopped up half the milk before returning to the dock. I thought that was that until I realized it was having its mops washed in readiness for the second run I had previously programmed.

I'd say the final results were perfectly acceptable given that I would never command a robot to mop up a spill when I can perform the task a thousand times quicker with some paper towels.

The main take away here is that its day-to-day performance has definitely been a cut above many of its competitors and there's no denying the practicality of the robot's extendable mop arm which really does work wonders.

Score ref: 4.5/5

Dock performance

I have nothing but good things to say about the X60 Pro Ultra's dock. Yes, it's a pretty large item to have in a small space but, on the plus side, I think its reflective pin-striped glass-like facade is one of the most attractive finishes I've seen since the Roborock 10R.

From a functional point of view, this dock does everything from emptying the robot's dust bin and servicing the robot’s mop system with hot water, washing the mops and then drying them, to self cleaning its own mop tray at the touch of a button on the app.

Only once did the dock's suction not correctly empty the contents of the robot's bin after a full house clean, and this was because I had activated the Hair Compression setting in the app. I'm not sure how this function works but the robot's bin was so tightly clumped with loads of winter-shed hair from my two labradors that the contents refused to budge when the dock's suction kicked in. Thankfully I've had no repeats of this since switching off the Hair Compression setting and, as a result, I am not losing any sleep over it.

Score ref: 5/5

• Performance score: 4.5/5

Dreame X60 Pro Ultra Complete review: app functionality

• Extremely comprehensive
• Brilliant for tech nerds, not so much for luddites
• Encourages a lot of experimentation

The Dreame X60 Pro Ultra’s Dreamehome app is extraordinarily well featured, giving you an almost ridiculous level of control over what the robot does, when it does it and how it cleans each area of your home. The problem is that it can initially feel like you’ve been handed the controls to an airliner.

As per usual, once the X60 has mapped your home, the app presents a detailed map that lets you merge or divide individual rooms, define cleaning zones and set no-go areas and virtual barriers. You can specify whether particular areas should be vacuumed, mopped or given both treatments, and adjust cleaning intensity according to the room or floor type.

(Image credit: Future)

There are also extensive options for scheduling and short-cut tasks, along with a myriad of settings for suction power, water usage, mop behavior, carpet type, floor direction, obstacle crossing, obstacle avoidance, even if you want pet hair in the robot's bin to be compressed.

The sheer depth of the Dreamehome app is undeniably impressive, but it does come with a steep learning curve that encourages a lot of experimentation. It can also take a while to work out where everything lives, particularly if you’re the sort of person who prefers to press a couple of buttons and get on with more important things. Nevertheless, once you’ve mastered this app, it becomes one of the X60 Pro Ultra’s biggest strengths.

• App functionality score: 4/5

Should you buy the Dreame X60 Pro Ultra Complete?Dreame X60 Pro Ultra Complete: score card

Section

Notes

Score

Value for money

While in the premium bracket, I think the Dreame X60 Pro Ultra Complete is fair value for money, especially given the on-board tech and huge amount of included spares.

4.5/5

Design

Slick, sophisticated and impressively engineered, the X60 Pro Ultra and its glossy dock ooze premium quality.

5/5

Performance

The X60 Pro Ultra delivers outstanding vacuum performance and better-than-average mopping, but its obstacle avoidance system could be better.

4.5/5

App functionality

The Dreamehome app is impressively comprehensive, offering deep control over all aspects, but its sheer number of settings can feel overwhelming.

4/5

Average rating

4.5/5

Buy it if

You want great vacuum performance

Powerful 42,000Pa suction tackles stubborn household detritus with aplomb

You have hairy pets

This model has several functions designed with pets in mind

You have low furniture and high thresholds

The X60 cleans beneath very low furniture and climbs most thresholds

Don't buy it if

Your home is cluttered with tiny toys

The X60 most likely won’t evade small toys like Lego bricks

You're not very tech savvy

The X60 Pro Ultra's surfeit of features may overwhelm less technical users

Your home is on the small side

This model would be overkill, as would the large dock

How I tested the Dreame X60 Pro Ultra Complete

TechRadar takes its product testing very seriously so I lived with the X60 Pro Ultra for a few weeks having set up various weekly schedules and routines to gauge how efficiently it performed the tasks it was designed for. I also monitored the amount of times I had to rescue it from entanglement or some other anomaly but, aside from some minor obstacle avoidance issues, the X60 functioned reliably throughout its trial.

If only to prove its capability for cleaning up heavier doses of debris and mopping up spills, I also performed a few extreme tests that no robot should have to endure in its lifetime.

First reviewed September 2026.

Traditionally set in North America, the Silent Hill series has been known to stray beyond the continent on occasion, most notably with the recent Silent Hill f, which takes place in sixties Japan.

Review info

Platform reviewed: PS5
Available on: PS5, PC
Release date: September 24, 2026

With Townfall, we’re headed to nineties Scotland, but that’s not the only change: we also get a new perspective — in more ways than one.

Not only is the game in first-person, rather than third-person, but it's developed by a series newcomer, Screen Burn Interactive, formerly known as No Code. This small Scottish outfit doesn’t exactly have a long resume, but it's already proved itself with its two previous outings. 2017’s Stories Untold was an interesting, modern take on the text adventure, while 2019’s Observation, another first-person horror, won the Bafta for Best British Game in 2020.

Now the team brings its Caledonian spirit to bear on Konami’s flagship horror franchise, but does the new locale and camera angle represent another step in the right direction for a series keen on extending its recent purple patch?

Everybody’s aff to the rapture

(Image credit: Konami)

We open in typical horror story fashion: a man crawls out of the water onto a dock, not knowing where he is or where he came from.

Best bit

(Image credit: Konami)

Sneaking around the streets of St. Amelia. Its layout is ripe for creating your own paths to avoid enemies, and escaping their clutches by running down back alleys and hiding in cupboards is always a thrill. Plus, you get to explore a town that’s been lovingly and realistically rendered, with homes, shops, and features that offer a glimpse back to a simpler time.

This man is Simon Ordell, wearing an outfit almost identical to Travis Grady’s from 2007’s Silent Hill: Origins, and denoting his everyman-ness. As you make your way up from the harbour, a cutscene plays, and a tarp blows away from a sign, revealing the town’s name: St. Amelia

And so the game proper begins, as you make your way through the seemingly abandoned town to find out what’s going on. Along the way, you’ll take control of two additional characters besides Simon — Zoe and Richard — and the plot unfolds gradually to reveal the trio’s connection to one another.

(Image credit: Konami)

The first thing that struck me as I paced through the streets was the fantastic art direction. As expected from a Scottish team, Screen Burn has nailed the feel of a small Scottish town.

St. Amelia is lovingly crafted and perfectly captures its mid-nineties period, containing plenty of features recognizable to those this side of the pond, from period-correct storefronts and pebble-dashed houses to public telephone booths and post boxes.

Virtually every frame looks cinematic (there’s even a letterbox mode enabled by default). The perennial fog is, of course, present and correct here, and looks better and moodier than ever. Hues of white, red, and blue are prominent throughout as well, and look eerily beautiful as they shine through the mist and are backlit by streetlamps. This is a game made for screenshotting, although unfortunately it doesn’t have a native photo mode.

I played through the game in Performance mode, as is often my preference, and I couldn’t discern any visual downgrade from Quality mode. Naturally, though, Performance mode offered a much better frame rate, and I only encountered one section where things got a little choppy, but this was brief and far from ruinous.

(Image credit: Konami)

Complementing the visuals is the score, another feature that’s always stood out in Silent Hill titles. Here, though, regular series composer and sound designer Akira Yamaoka is, for once, not involved. Instead, the honors go to Pilotpriest, who’s done a fantastic job of replicating the off-kilter sounds and melancholic melodies fans have come to know and love, while adding a modern twist.

Pieces are more sparse and muted than before, often with heavy doses of reverb. They match the tone of the gameplay brilliantly, and compositions open up during cutscenes when some emotional weight needs carrying.

Unfortunately, the characters’ facial animations are less capable at this. They’re often stilted, which can work when they seem detached and dejected, but they fail to sell terror or emotional breakdowns when required.

This is doubly a shame since it lets down the largely great performances of the voice actors, of which Kezia Burrows’ Zoe is worthy of highlighting. She has a greater range of traits and emotions to convey than the other two, and she does so with aplomb.

Another disappointing aspect of the game’s design is the monsters — this for a series renowned for its strikingly original creations. Virtually every one is some kind of faceless alabaster humanoid with no details of note, and some are mere rehashes of past favorites, such as the fleshy straight-jacketed figure, although this time it has a plank driven through its head.

Less attention has been paid to their symbolism, too: there are suggestions that they reflect aspects of the inner turmoil of the main characters, à la most Silent Hill monsters, but there's little about their design that shows this in an interesting way.

Monster hunting

(Image credit: Konami)

Looks aside, though, the monsters are more dangerous than ever. There are a handful of basic pawns, such as those that wander around slowly, and the aforementioned straight-jacketed plank heads that move much faster and more erratically and are equally troublesome.

If either catches you, you’ll have a basic button-mashing QTE to avoid them if you have enough health, but if not, it’ll be instant death. In certain sections, a larger monster appears with blobs on its back, which is unkillable as far as I could tell. There's no chance to QTE your way out of its grip if it catches you, either.

Sneaking around enemies, therefore, can be a tense and terrifying experience, accompanied by a huge sense of relief when you succeed. Often, the environmental design expertly facilitates this stealth gameplay. The outside sections in particular are laden with cut-throughs and hiding places, and more impressive is the fact that this design is so firmly planted in the real world.

You’ll be sneaking around abandoned cars and peeking out over low walls, scooting down narrow back alleys and distracting enemies with bottles or radios. There’s a surprising degree of player freedom that makes these sections more satisfying to play through than your typical survival horror fare.

(Image credit: Konami)

What also helps is the excellent player movement. All three move in the same way, which is to say they’re responsive and agile, as well as being reasonably fast when sprinting, although this is no arena shooter — but then again, I wouldn’t want it any other way in a horror game. There’s also an option to use motion controls for aiming, with fast and precise movements, and it is my preferred method.

There’s a peek mechanic, too, which again is very useful. It may be lifted straight from classic first-person horror games, most notably Alien: Isolation, but it gives you a great deal of freedom to not just strafe laterally, but also to peer up over barriers or down over ledges, whether standing or crouched.

To detect threats more easily, Simon picks up a handheld TV (or CRTV) early on that can tune into enemy locations. They’re rendered in silhouette against a red background, serving as the game’s ‘detective mode’. This gives you a real advantage, but its screen is just obscure enough to avoid you using it the whole time.

You also lose focus on your surroundings while equipped, again preventing you from relying on it wholesale. Keeping track of multiple enemies can be hard as well, since you have to toggle between them one at a time, and those long passed or on a different level remain in the CRTV’s memory.

(Image credit: Konami)

Zoe, on the other hand, is saddled with a pager, a far less effective item since it doesn’t directly reveal enemy movements. Instead, bars around the sides fill up to show their approximate direction and proximity. Both gadgets, though, are effective at grounding Townfall in the real world. At certain points, the CRTV can be tuned to show recorded messages that help develop the story or provide hints. The pager, meanwhile, receives mission- and story-related communication.

This grounding is something the game as a whole does well. From the environments and puzzles, right down to the save system, which has always been a quirky thing in Silent Hill games. Usually, it’s something obscure and surreal, such as interacting with a well or a red square on a wall, but here you use landlines (no mobile phones in St. Amelia 1996) and dial 100 — the in-game number for the coastguard.

The map screen is as helpful and skeuomorphic as ever. Scrawls in red marker pen appear to show rooms you’ve entered, locked doors, and story clues. You can also zoom and pan around it with ease.

When the sneaking goes wrong, you aren’t completely defenceless. You don’t exactly have a military arsenal at your disposal, but some humble series staples are here, including wooden planks and metal pipes. They degrade this time, though, so they won’t last long. Firepower comes later, but ammo is scarce, making the game’s intentions for a stealthy play style clear.

(Image credit: Konami)

All this isn’t to say the weapons are useless. On the contrary, they’re very effective, and just the right amount of difficult to use. Melee combat is a basic affair, with one button for attack and one for block, but a lot of enemies can be disposed of quite quickly. The revolver and shotgun are even more effective, as you would imagine, but reload times are slow. All weapons feel impactful and satisfying to use, too, and the controls aren’t awkward.

Healing items are more generously distributed. There’s gauze for minor wounds and first aid kits for major ones. You can also combine alcohol with the former for a larger heal, while empty IV bags can be combined with IV fluids to automatically revive you if you die.

Aside from combat and sneaking, you’ll of course be solving puzzles. These are more ingenious than past efforts, and expertly combine that aforementioned verisimilitude with the series’ penchant for the weird.

One of my favorites had me reading out patient charts as Zoe to an aggressively condescending doctor, chastising her and me if I gave the wrong information. I then needed to find medicines dotted around the medical center, beginning yet another monster-avoidance section, which was as tense and as exciting as those in the streets of St. Amelia.

There are plenty of documents and computer emails to read and collect, providing clues to puzzles. There’s also a handy mechanic where you can hold a button to bring up the relevant clue while interacting with a puzzle, although, bizarrely, this is limited to only a few occasions.

Nothing personal

(Image credit: Konami)

While I certainly had a good time with Townfall, certain sections drag on a little too long. But on the whole, the game does well to vary the environments and puzzles (the enemies less so) throughout to keep you hooked.

I didn’t find normal difficulty too troubling: the puzzles are pitched just right, and those I spent a long time figuring out weren’t frustrating or unfair; in fact, they’re proved more satisfying to solve than the series’ typical offerings.

Sneaking and combat, on the other hand, are more challenging. Enemies can come upon you quickly, and since there can be long stretches before you reach a save point, you can find yourself trudging back through the same section if you perish. But mostly, their presence adds real stakes and rewards you with a sense of achievement once you make it past them.

(Image credit: Future)

It took me just over 20 hours to complete a single playthrough, and that was at a gentle pace. There’s little beyond the main objectives at hand to seek out, aside from some mysterious coded postcards that’ll have fans scrabbling to decipher, as well as hidden phone numbers to find. There are also unlockable outfits and weapons, and five different endings to get.

Townfall isn’t particularly scary. There’s the occasional jump scare, but it doesn’t rely on them for cheap thrills. The game is a slow burn, with plenty of quiet, contemplative moments as you explore your surroundings. As with many Silent Hill games, the horror is more psychological.

But herein lies a problem. While the story is commendable for sustaining the mystery for as long as possible, there’s a lack of character exploration, contrary to many other stories in the franchise.

The characters’ dialogue is mostly cryptic, so you never get much of a grasp on their thoughts and feelings until the end — and even then, there are still plenty of questions left unanswered, as they should be. But the three of them are by turns confused and forlorn throughout.

This mixture works to conjure that strange, uncanny tone the series is known for, but it also means the trio fails to elicit the same level of emotional investment garnered by past protagonists, such as Harry and Heather Mason, James Sunderland, and Mary.

(Image credit: Konami)

Social commentary seems to be Townfall’s primary concern: the main theme is one of corporate negligence and its devastating impact on a community, which is certainly relevant in today’s world. And it’s true that once the three collide and everything comes to a head, it starts to resonate emotionally. But this is a long time coming, and Zoe ultimately feels underserved by the story’s conclusion, a shame given the aforementioned excellence of her actor’s performance.

It’s entirely possible, though, that other endings will give her a larger role than the one I reached, although I have no idea how these are achieved yet.

Ultimately, I found Townfall less impactful than Silent Hill 2 or Shattered Memories. It won't haunt me for as long as those entries, but it’s an intriguing journey nonetheless, with mysteries that remain after completion to keep fans busy with their theories.

No doubt some will also see Townfall as departing too far from the Silent Hill formula, a debate that’s already begun. But to my mind, the new setting and perspective fit the mold perfectly, so despite a few reservations, it still marks another triumphant reimagining of the series.

Should you play Silent Hill: Townfall?Play it if…

You love sneaking and puzzling
Sneaking around enemies is often thrilling, and the memorable puzzles mix realism with the series’ unhinged tropes masterfully.

You want great art direction
Screen Burn has managed to haul a small Scottish fishing village into the world of Silent Hill beautifully. It looks and sounds fantastic.

Don't play it if…

You want the best monsters
The monsters are certainly less memorable this time round, a shame given the series’ pedigree in this area.

You want more emotion from your characters
Mostly, the three protagonists are quite stilted in their tone. The botoxed facial animations don’t help, either.

Accessibility features

Compared to other modern releases, Townfall’s accessibility features are a little lacking. There are no content filters for blood and gore, and visual assists for color and contrast, and the like, are also missing. You can toggle subtitles and closed captions separately, though, and you can adjust the opacity of their backgrounds.

There are three difficulty modes to choose from: Story, which reduces the threat posed by enemies; Normal, which is how the developers intend the game to be played; and Nightmare, which increases enemy power and makes your flashlight detectable by them.

Puzzle difficulty is handled separately, although there are only two modes for this, with the only difference being red highlights of salient text clues.

How I reviewed Silent Hill: Townfall

(Image credit: Konami)

I played Townfall for just over 20 hours, completing the game once through on normal difficulty. I also toggled the puzzle difficulty between Normal and Hinted modes at various points.

I played the game on a standard PS5, largely in Performance mode, although I also tested Quality mode for a time, too. I used a Lenovo Legion R27qe monitor and the Skullcandy Crusher PLYR 720 headset for audio.

I’m a long-time fan of the Silent Hill series and have played many games in the series, including the recent Silent Hill f and the 2024 remake of Silent Hill 2. The original Silent Hill 2 and Silent Hill: Shattered Memories rank among the best horror games of all time, in my mind.

First reviewed September 2026

Full spoilers immediately follow for Lanterns episode 6.

Lanterns episode 6 has made its debut on HBO Max — and while it didn't answer every big question I had heading into Lanterns' sixth chapter, it does include a cameo from someone who's an incredibly important part of the Green Lantern mythos.

Titled 'Bad Optics', this week's chapter sees John Stewart return to Rushville 10 years after the town's near-destruction to investigate the apparent death of Hal Jordan. Before his mentor's funeral, Stewart meets and interacts with a character who has strong ties to Jordan in DC Comics, and whose presence in this DC Universe (DCU) TV show further cements the wider worldbuilding — or should that be universe-building — of this nascent cinematic franchise.

So, who is this mystery individual? What's their relationship to Hal Jordan? And could we see their superhuman alter ego at some point in the DCU? Let's take a look at the latest comic book character to be referenced in the hit HBO Max show.

Who is Carol Ferris in Lanterns? Hal Jordan's love interest in DC Comics, explained

Carol Ferris has been a part of Green Lantern lore for over 60 years (Image credit: DC Comics)

Created by John Broome and Gil Kane, Carol Ferris is not only Hal Jordan's long-time love interest but was also one of his most noteworthy foes for many years.

Making her debut in Showcase #22 in late 1959, Carol was introduced to readers as the only child of aerospace mogul Carl Ferris and the vice president of her father's company, Ferris Aircraft. After hiring Hal Jordan as a test pilot for said corporation, Carol quickly falls in love with the soon-to-be Green Lantern, but their romantic relationship doesn't last.

The reason? Carol is abducted, brainwashed, and turned into one of Jordan's arch-nemeses in the form of Star Sapphire by the Zamarons. Formerly part of an alien race called the Maltusians, the female members of this species became the Zamarons after parting ways with the male Maltusians, who then rebranded themselves as the Oans — or, to call them by their more common name, the Guardians of the Universe.

With the immortal Guardians no longer requiring their female counterparts to procreate, the Zamarons left them to their devices, developed a warrior-like cult, and — to cut a long story short — soon became fierce rivals to the Guardians.

Carol Ferris was the first DC character to assume the Star Sapphire mantle (Image credit: DC Comics)

But back to Carol. Following her kidnapping, she's infused with an incredibly powerful space stone known as a Star Sapphire, which turns her into a superhuman capable of going toe-to-toe with a Green Lantern.

Abilities-wise, Star Sapphire's powerset includes — but isn't limited to — the ability to fly, create energy blasts, constructs, and force fields, survive in space, alter people's minds, and even heal individuals with the power of... well, love.

Read my Lanterns coverage

- Lanterns episode 1 ending explained
- Sinestro in Lanterns explained
- Lanterns' Manhunters explained
- Lanterns' Guardians of the Universe explained
- Lanterns episode 4 ending explained
- Lanterns episode 5 ending explained

Anyway, ever since Star Sapphire's first appearance in October 1962's Green Lantern vol. 2 #16, Carol and Hal have battled each other on a regular basis as their superpowered aliases. Whenever Carol isn't possessed by the space gem she's synonymous with, the pair get on pretty well (for the most part) and have been romantically entangled on an on-off basis through the decades.

More recently, Star Sapphire has taken on the role of a superhero rather than a supervillain. In fact, as of 2007's Green Lantern vol. 4 #20, Carol is the leader of the Star Sapphire Corps, otherwise known as the Violet Lantern Corps. This group has purple power rings, which are charged/powered by the emotion of love, and are now viewed as allies of the Green Lantern Corps.

Who plays Carol Ferris in Lanterns?

Amie MacKenzie, seen her in 2025 Netflix medical drama Pulse, plays Carol in the DCU (Image credit: Netflix)

That'll be Amie MacKenzie. Viewers might recognize her from Better Call Saul and The Madison on Netflix, Dexter: Original Sin and Yellowstone spin-off 1923 on Paramount+, Mountainhead on HBO Max, and MacGruber on Peacock.

Lanterns' latest chapter doesn't indicate whether we'll see Carol again. While her appearance at Hal's poorly attended funeral suggests they were together at some point prior in the DCU, it seems her inclusion is simply to make diehard DC comic book fans imitate that famous Leonardo DiCaprio meme from Once Upon a Time in Hollywood. Don't expect to see MacKenzie as Carol's superpowered alter-ego at any point in the DCU, then.

Anyway, this isn't the first time Carol has appeared in a live-action project. She was portrayed by Blake Lively in 2011's Green Lantern movie, which starred Ryan Reynolds as the titular hero and was slammed by fans and critics alike upon release. But, hey, Reynolds and Lively fell in love, got married, and had kids after meeting on that film, so something good came out of it for them!

How Carol Ferris' cameo might be hinting at the creation of more Lantern Corps

Has Sinestro founded the Yellow Lantern Corps in the DCU yet? (Image credit: DC Studios/HBO max)

In spite of what I've just written, though, it's possible Carol's appearance could point to the formation of other Lantern Corps in the DCU.

Okay, that already appears to be a given. Sinestro's brief appearances in the DCU Chapter One show and Lanterns' title card slowly changing from green to yellow imply that it won't be long before Sinestro creates the Yellow Lantern Corps, if he hasn't secretly done so already.

Regardless, with Carol showing up in the sci-fi crime drama, we've now met individuals with ties to three of the seven Lantern Corps that exist in DC Comics, all of which harness a specific feeling on the emotional spectrum: Green (willpower), Yellow (fear), and Violet (love). The others are Red (rage), Blue (hope), Indigo (compassion), and Orange (greed).

Will we see any and/or all of these interstellar organizations in the DCU? It's possible, although recent comments made by the head of HBO have me fearing the worst about Lanterns season 2. If another season of this successful TV series isn't greenlit, we'd have to hope that they're part of a Green Lantern movie — maybe an adaptation of the acclaimed 'Sinestro Corps War' comic storyline, perhaps? — on the big screen instead.

• Listen to One Man In audiobook for free with Audible trial (US)
• Abroad? Use NordVPN (Save 75% + Try 30 days)
• Audiobook runtime: 5 hours 26 minutes

One Man In tells the remarkable story of former British special forces soldier Christian Craighead and the events of January 15, 2019, when he responded to the terrorist attack on Nairobi's DusitD2 complex.

Published on September 15, 2026, Craighead's memoir is available in hardcover, ebook and audiobook editions in the United States – although not the UK currently.

If you're abroad and want to access your usual US ebooks or audiobooks, use NordVPN (get our 75% off deal below).

Here, we'll show you how to get One Man In online and when abroad, plus the different ways you can read/listen to Christian Craighead's jaw-dropping story – including for free.

Listen to One Man In audiobook for FREE

One Man In is out now in the US, with readers able to choose between physical, digital and audio editions of Christian Craighead's memoir.

Listen to One Man In for free with Audible's 30-day FREE trial (includes 1 book)

Or, if you're a Prime member, you can get a 2-month Audible trial for $0.99.

Running for 5 hours and 26 minutes, it is read by Craighead himself, meaning listeners hear the former SAS soldier recount the events in his own voice.

Prefer to read? Amazon lists the hardcover edition at $30, while the ebook has a $14.99 list price. The publisher's unabridged audio download is listed at $19.99.

TRAVELING ABROAD? ACCESS AUDIBLE FROM ANYWHERE WITH NORDVPN

Access from anywhere

If you're from the US but traveling abroad, you might find your usual streaming services, websites and subscriptions are geo-blocked.

Obi-One Nairobi fans traveling or working outside the States may need to use a VPN to access One Man In when roaming abroad.

There are lots of VPNs, but NordVPN is the one you can rely on to unblock almost any service – save 75% and get 30 days to try it out risk-free.

NordVPN – try it risk-free for 30 days
Looking to access content from abroad? If you're traveling abroad, use NordVPN to access sites as if you were back home. We test all the providers and we rate Nord as the best VPN. There's 24/7 support available, a money-back guarantee and, best of all, there's currently over 75% off with this deal.View Deal

Access Audible free trial when abroad

It's really easy to use a VPN to access your US Audible account.

1. Install the VPN of your choice. As we've said, NordVPN is the best choice.

2. Choose the location in the VPN app. For instance, if you're visiting Europe and want to use your US Audible account, you'd select US-based server location.

3. Sit back and enjoy the action. Head to Audible and get your local free trial or paid subscription for audiobooks.

Who is Christian Craighead?

Christian Craighead is a former British Army soldier who joined the military at 16 and served for 28 years.

His career included service with the Parachute Regiment and Pathfinder Platoon before he joined 22 SAS, where he spent almost 15 years.

Craighead served in locations including Iraq and Afghanistan before becoming internationally known following his involvement in the response to the January 2019 DusitD2 attack in Nairobi.

He retired from active service on September 7, 2020 and was awarded the Conspicuous Gallantry Cross.

In September 2026 he made his second appearance on the Shawn Ryan Show podcast and provided a blow-by-blow account of the DusitD2 attack.

Which devices are compatible with Audible?

• iPhone
• iPad
• Android phones
• Android tablets
• Apple Watch
• Mac
• Amazon Kindle
• Amazon Fire tablets
• Amazon Alexa devices

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

Looking for a different day?

A new Quordle puzzle appears at midnight each day for your time zone – which means that some people are always playing 'today's game' while others are playing 'yesterday's'. If you're looking for Sunday's puzzle instead then click here: Quordle hints and answers for Sunday, September 20 (game #1700).

Quordle was one of the original Wordle alternatives and is still going strong now more than 1,500 games later. It offers a genuine challenge, though, so read on if you need some Quordle hints today — or scroll down further for the answers.

Enjoy playing word games? You can also check out my NYT Connections today and NYT Strands today pages for hints and answers for those puzzles, while Marc's Wordle today column covers the original viral word game.

SPOILER WARNING: Information about Quordle today is below, so don't read on if you don't want to know the answers.

Quordle today (game #1701) — hint #1 — VowelsHow many different vowels are in Quordle today?

• The number of different vowels in Quordle today is 4*.

* Note that by vowel we mean the five standard vowels (A, E, I, O, U), not Y (which is sometimes counted as a vowel too).

Quordle today (game #1701) — hint #2 — repeated lettersDo any of today's Quordle answers contain repeated letters?

• The number of Quordle answers containing a repeated letter today is 1.

Quordle today (game #1701) — hint #3 — uncommon lettersDo the letters Q, Z, X or J appear in Quordle today?

• Yes. One of Q, Z, X or J appears among today's Quordle answers.

Quordle today (game #1701) — hint #4 — starting letters (1)Do any of today's Quordle puzzles start with the same letter?

• The number of today's Quordle answers starting with the same letter is 0.

If you just want to know the answers at this stage, simply scroll down. If you're not ready yet then here's one more clue to make things a lot easier:

Quordle today (game #1701) — hint #5 — starting letters (2)What letters do today's Quordle answers start with?

• W

• H

• P

• L

Right, the answers are below, so DO NOT SCROLL ANY FURTHER IF YOU DON'T WANT TO SEE THEM.

Quordle today (game #1701) — the answers

(Image credit: Merriam-Webster)

The answers to today's Quordle, game #1701, are…

• WAXEN
• HIPPO
• PETAL
• LEASH

For one of my words I had “rhino” stuck in my head as it featured an H, I and O, which meant I managed to ignore the other animal answer of HIPPO.

The far bigger challenge, though, was arriving at WAXEN — a word I had pondered earlier in the game but ruled out because of the rare letter X. Eventually I had no choice but to take the plunge.

Daily Sequence today (game #1701) — the answers

(Image credit: Merriam-Webster)

The answers to today's Quordle Daily Sequence, game #1701, are…

• ANGEL
• OLIVE
• GRIMY
• MARCH

Quordle answers: The past 20

• Quordle #1700, Sunday, 20 September: DERBY, LIMBO, GAWKY, REIGN
• Quordle #1699, Saturday, 19 September: SCOPE, HEIST, PULSE, CHAMP
• Quordle #1698, Friday, 18 September: MUCUS, MIRTH, APPLE, TILDE
• Quordle #1697, Thursday, 17 September: CRIER, CHIRP, NEIGH, FREER
• Quordle #1696, Wednesday, 16 September: SPEAR, PRICE, BLUNT, FISHY
• Quordle #1695, Tuesday, 15 September: SHUSH, LITHE, DELTA, BUGLE
• Quordle #1694, Monday, 14 September: ZEBRA, LEGAL, RATIO, TROLL
• Quordle #1693, Sunday, 13 September: LUCID, SPRAY, MANGE, EMAIL
• Quordle #1692, Saturday, 12 September: SUPER, BRUSH, RESET, SOWER
• Quordle #1691, Friday, 11 September: PLUSH, PAINT, RIVAL, AFOUL
• Quordle #1690, Thursday, 10 September: AGENT, EXILE, TITAN, PUREE
• Quordle #1689, Wednesday, 9 September: HAPPY, GONER, SMACK, MOTIF
• Quordle #1688, Tuesday, 8 September: FLUNG, DIARY, HALVE, BERTH
• Quordle #1687, Monday, 7 September: VISIT, CLANK, FEVER, OCEAN
• Quordle #1686, Sunday, 6 September: PILOT, HONOR, BOWEL, TAUNT
• Quordle #1685, Saturday, 5 September: GNASH, GEEKY, WHELP, PUTTY
• Quordle #1684, Friday, 4 September: QUICK, BEEFY, TRUCK, SNAIL
• Quordle #1683, Thursday, 3 September: SIGMA, ESSAY, MERRY, LLAMA
• Quordle #1682, Wednesday, 2 September: INCUR, FELLA, TITAN, FROZE
• Quordle #1681, Tuesday, 1 September: NASAL, ROUTE, ANKLE, MIGHT

Looking for a different day?

A new NYT Connections puzzle appears at midnight each day for your time zone – which means that some people are always playing 'today's game' while others are playing 'yesterday's'. If you're looking for Sunday's puzzle instead then click here: NYT Connections hints and answers for Sunday, September 20 (game #1197).

Good morning! Let's play Connections, the NYT's clever word game that challenges you to group answers in various categories. It can be tough, so read on if you need Connections hints.

What should you do once you've finished? Why, play some more word games of course. I've also got daily Strands hints and answers and Quordle hints and answers articles if you need help for those too, while Marc's Wordle today page covers the original viral word game.

SPOILER WARNING: Information about NYT Connections today is below, so don't read on if you don't want to know the answers.

NYT Connections today (game #1198) - today's words

(Image credit: New York Times)

Today's NYT Connections words are…

• COMPOSE
• CHECK
• FLY
• BEET
• BRAH
• WORM
• SPAM
• MATE
• CASTLE
• CHUM
• VIVA
• SENT
• CHOP
• CAPTURE
• DRAFTS
• SPINNER

NYT Connections today (game #1198) - hint #1 - group hints

What are some clues for today's NYT Connections groups?

• YELLOW: Messaging tabs
• GREEN: Used in angling
• BLUE: Words relevant to board game
• PURPLE: Classical open

Need more clues?

We're firmly in spoiler territory now, but read on if you want to know what the four theme answers are for today's NYT Connections puzzles…

NYT Connections today (game #1198) - hint #2 - group answers

What are the answers for today's NYT Connections groups?

• YELLOW: EMAIL BUTTONS
• GREEN: WAYS TO ATTRACT FISH
• BLUE: CHESS VERBS
• PURPLE: STARTS OF COMPOSERS' NAMES

Right, the answers are below, so DO NOT SCROLL ANY FURTHER IF YOU DON'T WANT TO SEE THEM.

NYT Connections today (game #1198) - the answers

(Image credit: New York Times)

The answers to today's Connections, game #1198, are…

• YELLOW: EMAIL BUTTONS: COMPOSE, DRAFTS, SENT, SPAM
• GREEN: WAYS TO ATTRACT FISH: CHUM, FLY, SPINNER, WORM
• BLUE: CHESS VERBS: CAPTURE, CASTLE, CHECK, MATE
• PURPLE: STARTS OF COMPOSERS' NAMES: BEET, BRAH, CHOP, VIVA

• My rating: Hard
• My score: Perfect

I came close to creating a group containing BRAH, MATE, and CHUM but stopped myself after failing to find a fourth tile of friendship slang.

CHUM, I then decided, must have something to do with fishing — a thought triggered by salty sea dog QUINT in the classic Spielberg movie Jaws and his bucket of CHUM.

EMAIL BUTTONS, meanwhile, contained four words that you see nearly every day of your working life but blank out for that very reason; somehow I was able to force myself to link COMPOSE, DRAFTS, SENT, and SPAM.

Yesterday's NYT Connections answers (Sunday, September 20, 2026, game #1197)

• YELLOW: ROUND FLAT THINGS: BLU-RAY, COASTER, SAND DOLLAR, TORTILLA
• GREEN: ASSOCIATED WITH KENTUCKY: BLUEGRASS, BOURBON, DERBY, FRIED CHICKEN
• BLUE: VAMPIRE MOVIES: BLADE, NEAR DARK, SINNERS, TWILIGHT
• PURPLE: _____BOARDS YOU RIDE ON: HOVER, SKATE, SNOW, SURF

What is NYT Connections?

NYT Connections is one of several increasingly popular word games made by the New York Times. It challenges you to find groups of four items that share something in common, and each group has a different difficulty level: green is easy, yellow a little harder, blue often quite tough and purple usually very difficult.

On the plus side, you don't technically need to solve the final one, as you'll be able to answer that one by a process of elimination. What's more, you can make up to four mistakes, which gives you a little bit of breathing room.

It's a little more involved than something like Wordle, however, and there are plenty of opportunities for the game to trip you up with tricks. For instance, watch out for homophones and other word games that could disguise the answers.

It's playable for free via the NYT Games site on desktop or mobile.

• Look out for the small blue 'hello' signs in Apple Stores
• They're to help test the iPhone 18 Pro and Pro Max cameras
• The new phones are now available to purchase in retail stores

The next time you head into an Apple Store, you should be able to spot a small blue 'hello' sign somewhere, as per reports on social media — and the appearance of these little signs is apparently connected to the launch of the iPhone 18 Pro and iPhone 18 Pro Max.

As per MacRumors, the signs are there to help customers test out the optical zoom capabilities of the new handsets. Both models offer up to 8x optical zoom, so having something specific to zoom in on is going to be helpful.

While the iPhone 18 Pro and iPhone 18 Pro Max have a variable aperture feature on one of their three rear cameras now, the optical zoom capabilities are actually the same as the iPhone 17 equivalents that launched in September of last year.

There are also new Pro controls on the latest handsets, so settings for lens aperture, shutter speed, and white balance can be tweaked shot by shot — something else you might want to test out in an Apple Store.

Now available

از امروز تمام اپل استورها این نشان hello رو یه جایی دارن که باهاش بتونن زوم دوربین آیفون‌ها رو به مشتری‌ها نشون بدن pic.twitter.com/gpLv7CPaweSeptember 18, 2026

The iPhone 18 Pro and iPhone 18 Pro Max arrived in Apple's retail stores on Friday, September 18, six days after preorders started. Prices for the handsets start at $1,199 / £1,199 / AU$2,099 and $1,299 / £1,299 / AU$2,299 respectively (for 256GB of storage).

New Apple CEO John Ternus made an appearance at the Apple Store in New York to help celebrate the wider availability of the flagship iPhones — as well as the Apple Watch Series 12, the Apple Watch Ultra 4, and the AirPods 5.

He didn't mention anything about the new 'hello' signs in stores, but they've been spotted by customers nevertheless. They look to be located on higher shelves and fixtures, perhaps to put off anyone wanting one as a souvenir.

And if you're wondering where the iPhone 18 is and when you'll be able to test that out, don't panic — it's rumored to be showing up around March time, alongside the iPhone 18e and the iPhone Air 2.

In the UK, we're exceptionally good at producing legal dramas. From Silk to Kavanagh QC, we've been offering the world a strictness and sense of traditionalism that the likes of Law & Order just can't touch.

Nowhere was this more apparent that The Split, one of the most popular dramas on BBC iPlayer in recent years. Exploring the world of divorce law, we follow the Defoes, a family of sisters working as divorce lawyers, who come face-to-face with their past after their estranged father returns to the fold.

The series came to an end in 2024, but two years later, spinoff The Split Up is ready to take up the mantle. This time, we're following Kishan Law, a powerhouse British South Asian family law firm.

Aria (Ritu Arya) is poised to take over leadership from her father Dhruv (Sanjeev Bhaskar), but the death of Aria's mother six months prior causes Dhruv to hesitate. Add in a workplace love triangle, complicated siblings and a messy divorce trial walking through their doors every week, and we've got something to get our teeth into.

It's a brilliant and delicious watch — so much so that I binged all six episodes in one sitting. Speaking to writer Ursula Rani Sarma, I learned that she already has ideas for more (though no official renewal has been made). On top of that, they're not ruling out a Nicola Walker cameo further down the line.

'If a storyline created that opportunity, we could 100% do it'

"My personal hope is for three seasons minimum," Sarma tells me. "I feel like I'm only just getting started with writing these characters and their journeys. As I was coming up with these storylines for season 1, I also had ideas for seasons 2 and 3 in my head."

This last statement alone immediately makes The Split Up worth investing in. We're living in an age of routine TV cancellations and one-season miniseries, so for Sarma to be taking the time to think further ahead means there's a natural richness that (surely) the BBC can't say no to.

"The three-season arc also emulates The Split," she continues. "But of course, that turned around and surprised everyone with a special. It's endless, my hopes and dreams for this family."

No spoilers, but we leave The Split Up at a place where enough is resolved and left unspoken simultaneously — which, frankly, is exactly where you want be heading into another season.

But what about a direct crossover? "I don't think anything's off the table," Sarma adds. "It was really important for the first series to establish this family to hopefully make audiences fall in love with them in the same way. To build that up without wondering what would happen if Hannah [Walker] were to walk in.

"But this is the gift of being part of The Split. If a storyline created that opportunity, we could 100% do it. Its cast are clearly so fond of that show."

That's the sense of optimism I like to see... now I just need to keep my fingers crossed.

When the original Steam Deck was first announced back in mid-2021, it was a revelation. Valve proved that handheld gaming PCs could be cost-effective and still powerful enough to run demanding software at playable framerates. The first-generation Deck would ship in February 2022, with the Steam Deck OLED model later innovating with a larger, prettier screen, improved refresh rate, and a superior battery in November 2023.

The Steam Deck impressed at the time with its respectable semi-custom AMD APU complete with its four-core (eight threads) CPU clocked between 2.4 and 3.5 GHz, and 8 RDNA 2 Compute Units clocked at 1.6 GHz.

In short, the hardware inside meant that pushing the best PC games to 30 FPS (and above in some cases) at its native 1280 x 720p resolution was a possibility, but the competition has proven that it's time for Valve to go back to the drawing board if it wants to stay competitive.

That's because the Steam Deck is now fairly underpowered when compared to the best PC handhelds on the market. The competition have Full HD (1080p) displays (or better), and feature vastly superior hardware in the form of AMD Ryzen Z1/Z2 or Intel Arc G3 APUs that run rings around the previous-generation frontrunner. So, what does Valve need to do to ensure the Steam Deck 2 can compete if, and when, it launches?

1. A 120Hz Full HD (1080p) OLED display as standard

All of the major Steam Deck rivals use a high refresh rate 1080p display, or better, so it's time for Valve to catch up (Image credit: Microsoft)

Both the Steam Deck LCD and Steam Deck OLED models feature a native resolution of 1280 x 720p with either a 60Hz or 90Hz refresh rate. Considering when the former was launched, this was an acceptable trade-off for providing playable framerates. However, every competing product in the oversaturated space has since eclipsed this, meaning the Steam Deck 2 absolutely needs a 120Hz 1080p display at a minimum to be viable.

You only need to look as far as the biggest rivals from back in 2023; the Steam Deck's panel was passé less than a year after the hardware debuted, as evidenced by the Asus ROG Ally and Lenovo Legion Go. Both devices could confidently be called true Steam Deck rivals. The first-generation ROG Ally had a 7-inch 120Hz 1080p panel, whereas the original Legion Go had a 144Hz 8.8-inch QHD+ screen.

I'll always argue that anything above 1080p on a display as small as seven or eight inches ultimately results in a waste of resources, and it seems hardware companies got the hint in 2024. More than two years ago, the MSI Claw A1M launched with a 7-inch Full HD screen, as did the Zotac Zone, Asus ROG Ally X, Asus ROG Xbox Ally X, Lenovo Legion Go S, and Lenovo Legion Go 2, as well as recent releases such as the MSI Claw 8 EX AI+. There's no need for Valve to look to 1440p (QHD), but the company needs to get with the times to avoid being left behind; lacking a decent screen would kill the Steam Deck 2.

2. A semi-custom AMD Zen 4 and RDNA 3.5 APU (or better)

It's likely that Valve would go the semi-custom route again, and there's plenty of life left in RDNA 3.5 yet. (Image credit: AMD)

It's never been Valve's style to throw standardized components into its hardware. While the competition was comfortable using the AMD Ryzen Z1 or Z2, or gave Intel Arc graphics a go, the Washington-based company had to go a different route. We saw this with the semi-custom APU inside the Steam Deck and Steam Deck OLED, but this design philosophy has since been repeated with the recently released Steam Machine.

Valve's living room PC features a semi-custom AMD APU that utilizes a six-core (12 threads) Zen 4 CPU and an RDNA 3-powered GPU with a total of 28 Compute Units and 8GB of GDDR6 VRAM. This home-console rival was said to offer "six times the horsepower" of the Steam Deck, and the tech could act as the perfect blueprint for what should be inside the Steam Deck 2.

Now, considering the portable nature of a Steam Deck successor and the lack of an active cooling solution, combined with the price tag (I'll get to that further down the page), it's unlikely that the Steam Deck 2 will be as powerful as the Steam Machine, but it could close the gap. The Steam Machine uses around 170W of its total power draw when plugged into mains, with 110W dedicated to the GPU, 30W for the CPU, and the rest of the components (RAM, fans, SSD, etc.) accounting for the other 30 to 60W of draw.

When plugged in, the Steam Deck OLED draws a maximum of 45W (or 15V at 3000mA). When on battery power, Valve's handheld draws anywhere from four to 27W depending on the task at hand. Obviously, battery life has to be a huge consideration here as well, so we aren't going to see upwards of 100W for a handheld. Today's competition, chiefly the MSI Claw 8 EX AI+, can provide 30-60 FPS in High settings in 1080p in either 17W or 35W on battery, but that comes at its own (literal) cost.

That's the power potential, but what about a close approximation of how a semi-custom RDNA 3 or RDNA 3.5-backed APU could perform inside the Steam Deck 2? Well, the closest we can get is with the performance jump from the AMD Ryzen Z2 compared to the Z1. At lower TDPs, the Z2 offers a generational leap anywhere from 4W to 40W, offering playable framerates of at least 30 FPS in 1080p, which is the only standard requirement for a Steam Deck Verified game in 2026. In particular, the Z2 Extreme features 16 RDNA 3.5 Compute Units with a total boost clock of up to 2.9 GHz and 24GB LPDDR5x-8000 memory; Valve could use something similar.

3. A sub-$1,000 / £1,000 price tag

The MSI Claw 8 EX AI+ is a spectacular machine, but it's eye-watering MSRP makes it completely unaccessable for most. Valve should not repeat this mistake (Image credit: Future)

When the Steam Deck was first revealed and distributed, it was sold at a loss. Prior to the original model's release, Gabe Newell said this decision was "painful" but "critical" for the hardware to break through to the market. Affordability and competitive performance took priority over outclassing the competition, and that was the core appeal of the Steam Deck, especially in the early years.

Well, it's nearing the end of 2026 now, and hardware companies (including Valve itself) have not learned from this example. The Steam Deck was successful because it was an affordable and powerful entry point into PC gaming, but that's a reality that's (unfortunately) becoming harder to remember as time marches on. The ongoing RAM crisis has meant the Steam Deck OLED shot up in price by 50%, and sales are also plummeting, and consumers have already begun to argue that handheld gaming PCs are back into the unaffordable niche product they were pre-2021.

To course-correct, Valve needs to offer the Steam Deck 2 for less than $1,000 / £1,000 / AU$2,000, at least for its base and mid-range configurations. We've seen growing backlash from gamers against rising prices recently, particularly with the eye-watering MSRPs of the new Asus ROG Xbox Ally X20 (starting at $1,299.99 / £1,299.99) and the aforementioned MSI Claw 8 EX AI's shocking $1,799.99 / £1,399.99 entry point. To remain competitive, Valve needs to undercut the competition here, even if it means sacrificing some superficial features and maybe even peak performance.

In other words, Valve needs to remember why the Steam Deck succeeded in the first place: it was powerful enough and affordable enough to make everyday people consider PC gaming. It's a tall order given we're currently in one of the worst times in the computing industry in recent memory, where the price of GPUs, RAM, SSDs, and anything else that uses memory are skyrocketing with no signs of slowing down. Even if the Deck 2 is originally sold at a small loss, it would go a long way to prove that Valve can be pro-consumer again, as it lost favor with the Steam Machine's so-so performance for the price tag.

4. TMR sticks and mechanical buttons

Valve has come a long way with its controllers over the years, but there's still work to be done (Image credit: Future)

Now that we've covered a prospective Steam Deck successor's price tag, specs, and screen, it's time to talk about ergonomics. While both versions of the Steam Deck are comfortable in the hand for long play sessions, the actual buttons and sticks leave a lot to be desired. Likely due to a price constraint, the handhelds utilize the standardized potentiometers, with physical parts that can cause wear and tear over time. If left unchecked for long enough, it can eventually lead to stick drift, which can be frustrating (and expensive) to repair.

Magnets are the way to go, then. It's tough when weighing up what would be the better technology for the Steam Deck 2's analog sticks, as both Hall Effect and TMR (Tunnelling Magnetoresistance) have their pros and cons. However, it's TMR that pulls ahead in this case for a couple of key reasons: it's more sensitive, accurate, and (thankfully) cheaper to implement now than a couple of years ago. There's a reason why so many of the best PC controllers on the market are already using TMR, so Valve would be missing an open goal if it cheaped out unnecessarily here.

TMR sticks and magnetic sensing technology as a whole are far superior and lead to fewer points of failure, ensuring longevity. It's essential for a handheld like the Steam Deck 2. Modern PC controllers, such as the Scuf Envision Pro 2 and the Scuf Omega, have led the charge with microswitch buttons and D-pads.

They not only provide instant actuation for more precise inputs, but they will also last longer than a pure membrane alternative. It's not just about immediacy and feel in the hand, but also about how this design decision could influence long-term use. Ideally, you would want a Steam Deck 2 to feel as good as the day you bought it, even five years afterwards.

5. A bigger and more efficent battery for longer (or more demanding) sessions

The Steam Deck OLED brought a 25% battery increase. It's time for Valve to consider another substantial bump to stay competitive (Image credit: Valve / Steam)

Any good portable device, whether it's the best phone or a Steam Deck successor, lives and dies on how good the battery life is. The original Steam Deck featured a 40 Wh (or 5200 mAh battery, whereas the Steam OLED bumped things up to 50 Wh (or 6,470 mAh). That netted you anywhere from two to 12 hours depending on the kinds of games you were running, with indies at low brightness yielding better playtime than a AAA release at full brightness would. On average, though, around two to four hours is considered the standard.

That's fine, sure, but Steam Deck rivals have since included larger batteries with longer average playtimes, and this is something Valve can directly address. For example, the Asus ROG Ally X has an 80Wh battery, meaning it can last anywhere from four to six hours of playback of intensive games played at 25 to 30W under the right conditions. That's a battery size increase of 60%; considering how many more lithium-ion batteries are being made in a compact or portable form factor, it means that this component cost could be cheaper than in 2021/22, and a larger battery could be among the top features of the successor device.

Ultimately, we want the ability to sit down and play our games on the sofa or on the move for longer, so offering a bigger battery with a superior lifespan would give the Steam Deck 2 the edge it would need. If Valve can bump the resolution up, tighten the ergonomics, increase the horsepower, and keep the costs down, then the second-generation of its handheld could be the best on the market.

• Samsung is shifting additional DDR5 and SSD production to outside manufacturers
• The company needs internal factory space for advanced HBM production
• DDR5 module assembly remains simpler than advanced HBM packaging processes

Samsung Electronics is shifting all additional production of conventional memory modules onto outsourced partners going forward.

The company is directing capacity growth for DDR5 modules and SSDs toward external firms, reserving its in-house capacity for the more advanced HBM packaging process.

Industry sources say this reflects Samsung's mounting need to use scarce back-end capacity more efficiently.

A deliberate pivot toward outsourced capacity

“Samsung Electronics lacks space to expand existing back-end production lines needed to manufacture semiconductors for artificial intelligence (AI), so it is reallocating space and equipment at packaging plants in Cheonan and Onyang to advanced packaging lines, including HBM,” a person familiar with the matter said.

“The company has decided to handle most increases in conventional memory module production through outsourcing rather than expanding its own capacity.”

Samsung has reportedly pressed its outsourced semiconductor assembly and test partners to expand capacity for DDR5 modules and SSDs.

The company has even gone a step further by asking partners to accelerate capacity expansions that were already part of its existing plans.

Several partners are already scaling up in response, including Dreamtech, Hanyang Digitech, and SFA Semicon.

Dreamtech began mass producing DDR5 modules in India last November and approved further equipment investment in June.

Its Noida facility, previously used for smartphone circuit board assembly, will eventually reach an annual capacity of over 50 million units.

Hanyang Digitech has invested more than 31.5 billion won in 2026 so far to automate its existing memory module plant located in Vietnam.

SFA Semicon is relocating DDR5 testing and assembly equipment from Samsung's Onyang campus to a new site in the Philippines.

That transition begins in November, with a second phase following during the second quarter of 2027.

Limited space drives the broader manufacturing shift

Samsung's own capacity constraints stem partly from a new HBM plant under construction at its Onyang campus, valued at 6 trillion won.

That facility will take multiple years to finish, leaving little room for conventional module expansion in the meantime.

A separate $1.5 billion back-end processing plant is also under construction in Thai Nguyen Province of northern Vietnam.

That facility will handle testing for older memory types, including DDR4, low-power LPDDR4, NAND flash, and universal flash storage.

DDR5 modules require mounting prepackaged DRAM and NAND chips onto circuit boards using surface-mount technology.

That process remains considerably less complex than advanced packaging methods used for HBM, making outsourcing comparatively straightforward for partners.

Micron previously pursued a similar route, shifting conventional memory production externally while concentrating internal resources on higher-margin products.

Whether SanDisk eventually follows a comparable path remains uncertain given how differently each company's manufacturing footprint and product mix are currently structured.

Via The Elec

• Satellite imagery shows China building a roughly ~200m ship in Shanghai that analysts say is a mothership for submarine-sized drones
• It pairs a bottom-opening well-deck and 12-arm cradle with an internal hangar that experts estimate holds four 115- to 148ft XXLUUVs
• While the ship itself is shorter than an Essex-class carrier, its dock clocks in at about 1,115ft

China appears to be building a type of ship no navy is publicly known to operate: a large surface vessel designed to carry, launch, and recover uncrewed submarines the size of conventional crewed boats

The hull was first posted on September 13 2026 by Japanese open-source analyst KAROTASU and assessed two days later by H I Sutton at Naval News.

It currently sits at a dock at the Hudong-Zhonghua shipyard in Shanghai, and Beijing has yet to formally acknowledge its existence, so it has no name, list of capabilities, or the military branch that will eventually operate it.

A large dock linked to a much smaller ship

KAROTASU's post places the vessel in what he calls a 340m-class dock, which works out to approximately 1,115ft. However, the dock is not the ship. The same post estimates the hull itself at roughly 200m long and 30m wide, and notes that a second, unidentified structure is being assembled in the same dock.

Defence Security Asia, using imagery dated September 15, puts the hull at about 205m by 27m, which clocks in at roughly 672 ft. This makes the mothership big, but not as big as the largest WWII-era fleet carriers, though larger than the US Navy's Independence-class light carriers from that era.

For context, an Essex-class carrier, the backbone of the US Navy's Pacific campaign from 1943, measured 872 to 888ft overall and displaced 36,380 long tons at full load. At 656 to 673ft, the Chinese hull comes up roughly 200 to 230ft short.

The primary use case of such a mothership would be to deploy extra-extra-large uncrewed underwater vehicles (XXLUUVs), which is a mouthful, and it builds around China's naval ambitions at a time when it seems certain that any conflict or show of power between the world's two superpowers will involve a maritime component.

At 35 and 45m (115 and 148ft), the XXLUUVs are the size of small crewed submarines and have an estimated range of 10,000 nautical miles based on a similarly sized concept that a Chinese shipbuilder has exhibited.

This, at the very least, challenges the current balance of power on multiple naval fronts that were not in play previously: until now, the drones have been tested from specially built floating docks, which conceal them in port and carry them out to sea for launch.

This works well on paper and for trials, which often use a mix of towing, favorable weather, and predictable shore infrastructure, but it limits range to their maximum operating capabilities and access to spare parts/repair potential, all of which the yet-to-be-named vessel addresses.

All in all, in a world that has conflicts flaring up left, right, and center, the world's largest surface ship is designed around full-size uncrewed submarines, with an enclosed hangar to house and maintain them, which is both a statement and a real threat to the US and its allies on the naval front from China, which is closing gaps on multiple fronts and taking the lead on others, even as territorial disputes in the South China Sea remain a key issue.

• Mouse Computer has fitted Ryzen AI processors into a laptop weighing just 981g
• The X2 uses a 12.2-inch display with 1920 by 1200 resolution
• A free SODIMM slot allows memory expansion from 16 GB to 32 GB

Japanese electronics maker Mouse Computer has introduced a compact laptop called the X2, a compact notebook weighing 981g.

The device pairs a 12.2-inch display with AMD processors, offering configurations that scale from modest to genuinely capable performance.

The laptop frame measures 277.1 mm wide, 210.2 mm deep, and 19 mm tall, retaining a single unoccupied SODIMM slot for later RAM upgrades.

Specifications built around three AMD chip options

Buyers can choose between three distinct AMD APUs, each aimed at a different performance and price bracket for portable computing.

The base option, a Ryzen 5 216, combines two full Zen 4 cores with four smaller Zen 4c cores in a six-core layout, which should perform similarly to AMD's existing Ryzen 5 8540U based on available benchmark comparisons circulating online.

A step up brings the Ryzen AI 5 340, part of AMD's Krackan Point lineup, also arranged in a six-core configuration.

Multicore results suggest this chip lands close to Intel's Core Ultra 7 165H and edges past the Core Ultra 5 325.

The top configuration, a Ryzen AI 7 350, uses eight cores split evenly between four Zen 5 and four Zen 5c units, which approaches Intel's Core Ultra 9 185H in combined multi-threaded and synthetic benchmark scores, according to early comparisons.

The portable display panel runs at 1920 by 1200 resolution and reportedly covers 100% of the sRGB color gamut.

A 180-degree hinge, 60 Wh battery, and 500 GB NVMe SSD round out the internal hardware package.

Connectivity includes gigabit LAN, Wi-Fi 7, Bluetooth 5, USB Type-A ports, HDMI output, and a microSD card slot for storage expansion.

A 2 MP webcam with a privacy shutter supports Windows Hello, while Windows 11 Home ships preinstalled on every unit.

Battery life reaches roughly 12.5 hours during video playback and about 25 hours in idle mode, measured under JEITA 3.0 standards.

Pricing stays confined to the Japanese market for now

The Ryzen AI 5 variant, equipped with 16 GB of RAM and 500 GB of storage, costs ¥239,800, or roughly $1,545.

The higher Ryzen AI 7 configuration carries the same memory and storage allotment but costs ¥259,000, or about $1,670.

Sales of this device will begin sequentially from October, though the rollout remains confined to the Japanese market via the company’s online store for now.

The laptop is currently available in three colour options — Cosmo Blue, Frost Gold, and Pearl White.

Both versions ship with single-channel memory by default, though the empty SODIMM slot allows an upgrade path from 16GB to 32 GB.

A compatible 16 GB DDR5-5600 module currently retails for around $229 through Amazon's storefront.

Interested customers elsewhere would need to rely on import services to acquire a unit, adding cost and complexity to the process.

Via Mouse | GDM (originally in Japanese)

Charles Cecil wants to make one thing clear: “I'm not anti-AI.”

The veteran British developer behind the legendary 1996 point-and-click game Broken Sword: The Shadow of the Templars speaks to TechRadar about AI's divisive role in modern game development - and, in particular, how it was used in the upcoming remake of the sequel, Broken Sword - The Smoking Mirror: Reforged, due out in 2027.

The original, a commercial success selling 759,000 copies, sees American tourist George Stobbart team up with freelance journalist Nicole Collard to track a mysterious Mayan stone across Paris, London, and Central America. This remake doesn't use AI, and, according to Cecil, there's a very good reason for that.

AI don’t think so

(Image credit: Revolution Software Ltd)

“I think AI has its place,” Cecil says. “It's just not relevant in our games because we want human animation, human characterization and human expression, and that's much better done by humans.”

A perfect example of that lies in the game's hand-drawn art, all of it redone for Broken Sword - The Smoking Mirror: Reforged.

This hugely labor-intensive process encompassed roughly 55,000 frames of animation, more than the 30,000 of Broken Sword: The Shadow of the Templars’ 2024 remake.

“George opens the newspaper and kind of just glances at the screen,” says Cecil, referring to the actions of the game's protagonist in an early scene. “He's looking at the viewer, and that's what human beings do. That is not something that AI can do.”

In low-res, eyes were just dots on a face. In high-res, eyes are windows to the soul and therefore have to be animated with precision and intent. Where applicable, artists were free to add flourishes, using their own interpretation to fill in the gaps.

For instance, the painting of Professor Oubier hanging on the wall in the rapidly burning study was always too pixelated to see clearly. Now, however, you can make out exactly what he looks like in 4K, at 36 times the original resolution, all while a visibly uncomfortable George jostles his rope bindings.

Developer Revolution Software didn't stop there. The team redrew scores of items that never really made much sense. Take the poison dart, which now looks less like a playing dart and more like something shot out of a blowpipe.

Human revolution

(Image credit: Revolution Software Ltd)

Despite the decision to abstain from AI use for the most part, it wasn't entirely absent. AI was used for the initial round of upscaling before artists added detail by hand, with backgrounds scanned as black-and-white layouts for them to build on using the originals as references. AI was more of a tool than a creative device.

“We experimented with AI for the first one, and we used AI to upscale and then redraw the characters. And as I say, it can upscale, but it can't put any character in.”

That's largely thanks to the luxury of time. The first game needed to be made in a year or risk cancellation. Revolution Software sprinted to the finish line, but made a few shortcuts along the way.

One oversight was 40 voice lines that were recorded and packaged in the game's files but never used. Now, in Broken Sword - The Smoking Mirror: Reforged, fans can finally hear them (one line involves George asking the Parisian waiter about the tequila worm in his pocket).

Another infamous example occurs in the Game Boy Advance version. “You have a choice to go to Spain or Syria,” says Cecil. “If you do one, you can finish the game, and it's fine. If you do the other, it takes you there, but you can't then finish. There's a bug that doesn't allow you to finish…It's an awful, awful bug. I mean, it's absolutely shocking.”

(Image credit: Revolution Software Ltd)

An AI might have spotted such mistakes, but then again, you could argue mistakes are what give art humanity. If nothing else, they make for a great story.

But the remake is far from a bug-fixing exercise. Instead, it's an opportunity to make the game better. Where most remasters seek to preserve the subject, Broken Sword - The Smoking Mirror: Reforged is almost an act of excavation.

“All of these things we now have the time to do,” says Cecil. “Because what Revolution has done for the last 15 years is we've self-published all our games. And I can decide when we’re going to spend a little bit more time getting this, right? Is it worth it?”

The devs found time to add features, too, like a new three-tier hint system that actively monitors your progress and feeds you steadily more obvious tips so you're not stuck for too long. The final tip tells you outright what to do. Thankfully, purists can turn it off.

Meanwhile, Nintendo Switch 2 players get four different control schemes: touchscreen, gamepad, one Joy-Con 2 detached and used as a mouse, or a versatile combination of everything.

Plus, there’s something Cecil has personally wanted to add for 30 years: a beautiful, detailed map of Paris. This is used to aid player navigation instead of having them rely on clunky icons. With everything in the game, these feel like choices made by humans, not AI.

“We've really gone out of our way to produce something that, in my view, is completely future-proof, because people were playing the first one 30 years ago, and millions and millions of people have played that game. And my ambition is that in 20 years' time, a fourth generation will be playing again.”

“Going forward,” he continues, “people are going to seek out human-crafted art. You know, there's going to be that sense, because obviously we all have a human connection with each other.”

Huawei, as a brand, is in a very interesting place right now. The company is genuinely innovating on a lot of fronts. It just launched the Mate XT2 tri-folding phone, which brings Samsung-like Privacy Display tech to the tri-folding form factor, as well as the Watch D3, which puts medical-grade blood pressure measurements into an even slimmer smartwatch.

But the problem with Huawei has always been that you can't buy a lot of the company's devices in markets outside Asia. And while that doesn't look like it's changing anytime soon, every once in a while, Huawei does launch some products in the West, and it reminds us (again) of what we're missing from other major hardware makers.

That's exactly what happened with the new Huawei MatePad Air. I usually use my still-going-strong M1 iPad Air (2022) as my main tablet, but I've spent the past month using the MatePad Air as my primary device. And it has shown me the one upgrade Apple could make that would probably convince me to finally buy the next iPad Air.

(Image credit: Sanuj Bhatia / TechRadar)

The MatePad Air itself doesn't look like some groundbreaking new tablet. In fact, Huawei has so many tablets in its lineup that you'd probably have a hard time distinguishing the new MatePad Air from some of the company's other models just by looking at the back.

But flip the MatePad Air around, and that's where, for me, it starts beating a lot of the competition. The MatePad Air comes with a 12-inch OLED display with up to a 144Hz refresh rate and a 1,200-nit peak brightness. But the model I've been testing for the past month is the higher-end PaperMatte version, which can reach up to 2,000 nits.

If you're unfamiliar with Huawei's PaperMatte display technology, it's essentially a special anti-glare treatment on top of the display that, Huawei says, eliminates up to 99% of ambient light reflections. It also gives the screen a more paper-like texture, which feels natural when you're writing with a stylus, while also helping reduce things like eye strain and smudges.

(Image credit: Sanuj Bhatia / TechRadar)

And this display alone is what has made me prefer using the MatePad Air over my iPad Air for the past month.

Think about what you actually do with a tablet. Whether you use an iPad Pro or the Samsung Galaxy Tab S11 Ultra, almost everything revolves around the display. You're watching content, taking notes, making video calls, or reading. For 99% of the time, you're looking at and interacting with a screen.

It's not like the iPad Air's display is bad. It's still a good-looking panel, and I've used my M1 iPad Air for years without any complaints. But Apple hasn't meaningfully upgraded the display technology in a while, and it's starting to show its age.

The iPad Air still uses an LCD panel. It's still stuck at 60Hz, and the bezels are quite thick as well. None of these things really bothered me before, but switching back and forth between it and the MatePad Air made the difference quite obvious.

The 144Hz OLED panel on the MatePad Air is obviously smoother, the colors look great, and the thinner bezels make watching content much nicer. And that PaperMatte display just makes everything twice as nice.

(Image credit: Sanuj Bhatia / TechRadar)

I've used the MatePad Air pretty much everywhere, be it the subway, airplanes, or my home, and not having reflections constantly bouncing off the screen has been great.

And it's not like Apple doesn't know how to do this. The company already offers its nano-texture display on the iPad Pro. The problem is that you need to buy one of the most expensive iPad Pro configurations (specifically, a model with 1TB or more) to even get the option, pushing the price to $1,899 / £1,899 / AU$3,249.

Huawei, on the other hand, makes the upgrade much easier to swallow. In the UK, you get the PaperMatte display regardless of which version of the MatePad Air you choose; the cheapest is just £699.99 (it's not available in the US). You're also getting 256GB of storage on the base model, compared to the 128GB starting storage on the iPad Air.

(Image credit: Sanuj Bhatia / TechRadar)

Of course, the MatePad Air isn't better than the iPad Air at everything. Apple's M-series chips still give the iPad a major performance advantage, and if I'm editing videos or doing other compute-heavy work on a tablet, I'd still rather have the iPad Air.

But most of the time, that's not what I'm doing on my tablet. I mostly use my iPad to watch content and browse the web, and for both of these things, I'd rather have the MatePad Air's display.

Plus, the hardware around the OLED display is pretty impressive, too. At 12 inches, the MatePad Air sits somewhere between Apple's 11-inch and 13-inch iPad Air models. It weighs just 509g and uses an aluminum frame and back, making it more than 100g lighter than the 13-inch iPad Air.

It's thinner too, measuring just 5.3mm compared to 6.1mm for the larger iPad Air, despite Huawei fitting a 10,100mAh battery inside compared to roughly 9,705mAh on Apple's tablet.

(Image credit: Sanuj Bhatia / TechRadar)

That said, none of this means I'm ready to completely ditch the iPad. I still prefer its performance, app ecosystem, and just how well it works with the rest of my Apple devices. But after a month with the MatePad Air, the one thing I desperately want to take back to my iPad is this display.

Give the next iPad Air an OLED display with (hopefully) an anti-reflective PaperMatte-like option that doesn't burn a hole in my pocket, and my M1 iPad Air might finally have a reason to retire.

The iPhone 18 Pro Max main camera's new variable aperture is a huge deal and, in my limited experience, I can already see that it could transform iPhone photography. Still, to know why this matters, you need to understand how it works.

The best analogy is your eyes.

If you've ever squinted to read something, you understand how an aperture works. The more open your pupil, the more light that comes in, but it's widely scattered, making objects in the distance out of focus. If you squint, that cuts the amount of light coming into your eye, but it also removes more of the scattered light, leaving only the sharper light and imagery. On a camera, the open aperture (represented by lower f-stop numbers) has the same effect as your wide-open pupils. Lots more light, but only stuff nearest you is in focus. A smaller aperture, represented by higher f-stop numbers, is the eye-squint: less light but more focus throughout the frame.

Ever since Apple announced the iPhone 18 Pro (and Pro Max's) big camera update, which also includes voluminous Pro-level controls, I've been itching to try them. As soon as I got my iPhone 18 Pro Max in black, I ran outside to see if the update makes an appreciable difference in photography quality and possibilities.

The short answer is yes.

Understanding this new variable aperture

A few things you should know about this new hardware feature. First, it's only on the 48MP main lens, so the 48MP, 4x telephoto, and 48MP ultrawide cameras only have fixed apertures: ƒ/2.8 and ƒ/2.2, respectively. Those are decently wide apertures, but the Main camera has Auto, ƒ/1.48, ƒ/1.8, ƒ/2.8, ƒ/4.0. That's a wide range. It's not quite DSLR-level, where you can stop down to, say, ƒ/11, which can be useful for deep focus and for capturing ultra-bright objects like the moon at night.

It's also worth noting that these are physical f-stops. The iPhone 18 Pro Max Main camera aperture features an array of interlocked blades thinner than a human hair, and they slide over each other as the aperture opens and closes. It's an engineering marvel. In the right light, you can see them adjusting.

I think it's fair to wonder why we need a physical, variable aperture when many of the iPhones since the iPhone 7 Plus have included portrait mode technology. In the early days, it used information from its wide and telephoto lenses to craft a bokeh effect. Eventually, Apple started using software to analyse the image and create a portrait mode with just a single lens (see 2020's iPhone SE). No matter how the iPhones have done it over the years, it's impressive and, for most users, good enough.

(Image credit: Jacob Krol/Future)

My goal, though, was to see how the portrait mode on the iPhone 18 Pro Max might differ from variable aperture photography (my preferred setting was f1.48). In other words, software vs. hardware. I did also fiddle with the other f-stops, but in bright daylight, the bokeh effect on ƒ/1.8 and ƒ/2.8 is quite subtle.

I took pictures of myself, photos of statues, and the image above of a drying leaf. That photo, or rather that one compared to the Portrait Mode shot of the same leaf, is what convinced me: this is a game changer.

When you use Portrait Mode on an iPhone, you can adjust the bokeh before or after photography. It's a reminder that this is, at least in part, a software effect that can include or discard as much background sharpness as you desire.

iPhone 18 Pro Max variable apertureLance UlanoffPortrait modeLance UlanoffA new leaf

Using the variable aperture is more akin to traditional photography. You have the lens, the aperture settings, and, now, if you choose, shutter speed, exposure, and focus. In the initial images, I left everything alone except for the aperture. To access it, you select the grid menu in the camera app, choose Aperture, and then select the aperture opening of your choice.

iPhone 18 Pro Max variable apertureLance UlanoffPortrait modeLance Ulanoff

I chose f1.48 and snapped a photo of the same leaf.

The result, which you can see above, was two pretty good pictures, but closer examination reveals how Portrait Mode struggled with the curling leaf. Look at the stem. It's almost gone. By contrast, on the variable aperture shot, it was clear that the phone wasn't trying to apply some software magic to separate the leaf from the background and create the bokeh. The leaf is complete. The whole image looks natural and, honestly, the hardware does what it would do on a traditional camera: let the lens and aperture opening make the call.

The leaf is in stark, colorful focus, and the background, while not as blurred as you get from Portrait Mode's default settings, is pleasantly out of focus.

iPhone 18 Pro Max variable apertureLance UlanoffPortrait modeLance Ulanoff

I saw similar results with my selfies. Again, Portrait Mode goes hard on the background blur, but because it's largely computational, some artifacts remain. In particular, the space between the right stem of my glasses and my head is oddly sharp. And, if I'm being honest, the default level of bokeh sort of disconnects me from the background environment.

The results on the Gertrude Stein statue are less equivocal. The default portrait mode seems to favor the figure a bit more, though I still like the variable apertures' more realistic background blur.

Let's go Pro

Later, I started fiddling with a combination of aperture settings, shutter speeds, exposures, and focus. There are limits on all of these tools, but that's really the limitation of a smartphone camera. Without the really big sensor, wider space between the lens, aperture, shutter, and sensor, there's only so much you can do.

Still, I was impressed. My variable aperture shots look closer than ever to pro-level photography, and the additional pro controls are just so much fun to play with.

For this black and white style image, I adjusted the shutter speed to 1/223s with f1.8Lance UlanoffHere, the shutter speed is just 1/121s with f2.8Lance Ulanoffshutter speed 1/602s at f1.8Lance UlanoffF2.2 at a slow 1/60s.Lance UlanoffThis is a high-speed shot (1/2088s) at f1.48.Lance UlanoffEven though there was a lot of shadow, I used the Pro settings to push this to 1/693s with an f1.48 aperture.Lance UlanoffF1.48s with high shutter speed (1/14286s).Lance Ulanoff

If there's any disappointment here, it's that we have just one variable aperture. I could do amazing things with more f-stops on the telephoto lens. Ah well, there's always next year.

Check out my gallery to see some of what I shot on my first day with the iPhone 18 Pro Max.

• New and improved solar panel materials have been analyzed
• Two key physics effects combine to create the improvements
• It could help panels break through the 33% energy conversion rate

There's a long-standing physics theory that suggests solar panels will only ever be able to convert a maximum of 33% of the sunlight that comes their way, but a team from the University of Groningen in the Netherlands has detailed a way to harvest extra energy from 'hot' electrons that could push through this limit.

Solar panels work by using the photons from sunshine to jump-start electrons. With the most energetic photons, the result is super-charged, so-called hot electrons, that have the potential to generate significantly more electricity — but the problem is they cool down too quickly and lose the energy as heat before it can be captured.

This new research builds on previous studies reporting a promising new tin-based solar panel material. The material has demonstrated it can keep the heat from hot electrons trapped for around 1,000 times longer — but there have been competing ideas about the mechanism making that possible.

Trying to clarify what was going on, though a series of computer simulations and experimental measurements, was the purpose of the new research: understanding the material's key properties will help scientists build on the potential of the material and potentially scale it up into something that's viable commercially.

Ultra-long cooling

Two physics effects are essential for the upgrades reported by scientists (Image credit: Faber et al., ACS Energy Letters, 2026)

The analysis carried out by the researchers determined that there were two different actions at work. First, an effect called a hot phonon bottleneck creates a heat trap: as the environment around the buzzing electrons warms up so quickly, the electrons end up reabsorbing the thermal energy again, keeping their heat for longer.

Second, what's known as the Burstein-Moss effect creates an atomic traffic jam. As hot electrons cool, they quickly fill up the lowest available energy states in the material, which means other hot electrons can't lose their heat as quickly — it's like an airplane filling up from the front, with passengers arriving later having to walk all the way to the back.

These two effects were already known about, but now we know they're the reason that these special solar panel materials work — and that in combination they provide the extended hot electron cooling that might help to nudge future solar panels above that crucial 33% ceiling, and create clean energy more efficiently.

"It is the simultaneous satisfaction of these electronic, phononic, and chemical criteria, operating under high-injection conditions, that enables the ultra-long cooling times necessary for practical devices," write the researchers in their published paper, which appears in the journal ACS Energy Letters.

It’s Batman Day and as both a Batman and 4K Blu-ray fan, I had to celebrate the occasion. The caped crusader first debuted in May 1939 (a whopping 87 years ago!) and, ever since, has become one of the most iconic figures of page and screen.

While many remember the classic 1966 Batman series starring Adam West and Burt Ward, the character’s first appearance on-screen was in 1943, in a 15-part serial also titled Batman. Since then, we’ve seen numerous animated shows and big-budget movies starring DC’s comic book hero.

I thought I’d pick out some of my favorite 4K Blu-rays (and even some standard Blu-rays) of the caped crusader to help you celebrate the day, including one disc I regularly use as a benchmark for testing the best TVs and best soundbars.

The Batman (2022)Warner Bros. / Future Future / Warner Bros. Discovery

One of the latest big-screen adaptations, The Batman, starring Robert Pattinson in the title role, follows Bruce Wayne’s early years as the Dark Knight, as he deals with a serial killer named The Riddler (Paul Dano), while also navigating corruption in Gotham and a deepening mystery involving Bruce’s own family.

Anyone who reads my work on TechRadar will know that this is one of my go-to discs for AV testing. Its low brightness and numerous high-contrast scenes make it an excellent test for any display, while the superb shadow detail and crisp textures provide plenty to scrutinise. The interesting camera work also gives the film a murky, gloomy look that perfectly captures the seedy underbelly of Gotham.

If you have a Dolby Vision-capable TV, you’ll be rewarded with some truly breathtaking images that really show off what your screen can do. The best OLED TVs are particularly impressive with this disc, making it a no-brainer for my TV testing.

The Dolby Atmos soundtrack is immersive and packed with detail. My reference scene is the Batmobile/Penguin chase sequence, which shows off just about every element of a home theater surround sound system or soundbar. From the roaring bass of the Batmobile’s engine to the precisely mapped sound of swerving traffic and hail of bullets, The Batman is essential for anyone looking to hear their sound system shine.

Batman Forever (1995)Warner Bros / Future Warner Bros / Future

Directed by Joel Schumacher, Batman Forever took a more colorful approach than Tim Burton’s previous movies, Batman (1989) and Batman Returns. Starring Val Kilmer as Bruce Wayne/Batman, he must deal with the joint threat of Two-Face (Tommy Lee Jones) and The Riddler (Jim Carrey).

This is the first Batman movie I remember watching growing up. While the film itself remains divisive among Batman fans, that shouldn’t detract from what is a fantastic 4K Blu-ray release. Its bold, neon colors really pop on-screen, especially in Dolby Vision, with the Riddler's vibrant greens, the neon light tubes used throughout Gotham and the movie's vivid reds all bursting with colour.

There’s excellent use of shadow here, too, with the often melodramatic shots of characters standing forlornly over lamps creating a dynamic image. On an OLED, such as the LG G6 I used for testing, blacks are rich and inky throughout, while the 4K upscale delivers crisp textures, particularly with skin tones.

Batman Forever's Dolby Atmos soundtrack is just as impressive. The sound of the batarang cutting through the air is precisely mapped to rear channels, creating a wonderfully immersive effect on a surround-sound system. Explosions have plenty of heft, bullets pack a satisfying punch, and Two-Face’s helicopter at the start of the movie is a great showcase for Atmos.

The Dark Knight (2008)Warner Bros / Future Warner Bros / Future

The second entry in Christoper Nolan’s Dark Knight trilogy, The Dark Knight, sees Batman (Christian Bale) go head-to-head with the Joker (played brilliantly by the late Heath Ledger), a criminal mastermind who seeks to plunge Gothan into chaos and push Batman to his very limits.

The 4K Blu-ray of The Dark Knight looks fantastic throughout, but it’s the IMAX scenes where it shines. The opening bank robbery is packed with intricate detail, from the Joker’s creased suit, the robbers' mask, to the densely packed streets of Gotham city. Later, as Batman stands amid smouldering wreckage in another IMAX-shot sequence, the rich black levels and superb shadow detail look sublime on OLED.

There are some seriously impressive high-contrast scenes, too. During the prison escort, as cars move into the tunnels, the contrast between the dark surroundings and the trucks and overhead lights create a fantastic picture.

The DTS-HD 5.1 MA soundtrack is phenomenal. In the opening scene, the score is perfectly balanced, with the scratching strings mapped to the left rear and the ticking clock mapped to the right, all while the front channels handle the main action. Subwoofers will get a nice workout, thanks to the monstrous explosions, thumping bullets, and punchy Tumbler engine. There’s some excellent detail, too, when Batman rides his bike for instance. As the camera’s perspective changes, the trajectory of the sound is mapped to the correct channel with real precision.

The Lego Batman Movie (2017) Warner Bros / FutureWarner Bros / Future

The Lego Batman Movie follows Batman (Will Arnett), Robin (Michael Cera) and Batgirl (Rosario Dawson) as they try to foil the Joker’s (Zach Galifianakis) plans to take over Gotham city. This was a spin-off of the highly successful The Lego Movie, where Arnett’s Batman was first introduced.

Again, this movie looks superb in 4K. Color reproduction is fantastic, with bright, bold and lively colors throughout. The greens of the Joker’s hair and Riddler’s costume really pop, as do all the other colors, too. On the LG G6, reds were rich and deep, bringing excellent texture to images. Speaking of textures, the crisp animation looks slick, with the 4K detail really capturing the Lego look.

There are plenty of high-contrast scenes here, too, with bold HDR highlights standing out against deep shadows — such as the contrast between Batman’s eyes and his cowl. Shadow detail is excellent, retaining plenty of texture and defintion in even the darkest areas.

The disc includes both Dolby Atmos and DTS-HD Master Audio 5.1 soundtracks. Atmos delivers weighty but controlled bass, with the Batwing’s engines and the impact of explosions providing plenty of low-end punch. Directionality is excellent, too, especially during scenes with a lot of movement. The sound follows movement precisely, and it’s all mapped around the surround sound. The DTS-HD 5.1 track isn’t quite as powerful, but it has a pleasing warmth and plenty of detail.

Bonus: Batman: The Animated Series (1992-1995)Warner Bros / FutureWarner Bros / Future

Debuting in 1992, Batman: The Animated Series is arguably one of the most famous adaptations of Batman’s adventures on-screen, with an all-star cast including Kevin Conroy as Bruce Wayne/Batman and Mark Hamill as the Joker.

This is a rare Blu-ray recommendation from me, but for Batman Day, it would be impossible to leave the series out. Textures have been upscaled nicely and the beautiful animation and art style really benefits from the HD makeover. Colors are bright, vibrant and accurate across every episode, while the lack of HDR is hardly a drawback given how good the series looks.

Audio comes courtesy of a DTS-HD MA 2.0 mix, so while it won't put your surround-sound system through its paces, it's crips and clear, with clean dialogue, punchy effects and surprisingly good directionality. The iconic score is beautifully rendered — which, if you’re like me, will scratch that nostalgia itch nicely.

Reacher season 5: key information

- Season 5 renewed before season 4 even aired
- Will be based on Lee Child's 2015 novel Make Me
- Alan Ritchson will likely be joined by new cast members
- No production schedule or release date has been confirmed

Reacher season 4 has only just ended, and I'm already thinking about what could be to come in Reacher season 5.

That's partially because I wasn't completely taken with the season 4 storyline. My favorite Reacher seasons are where we see him drifting through small US towns that we've never heard of, punching a load of people as penance for their crimes, and then leaving.

But season 4 moved us towards a big city political espionage narrative — a move that clearly divided viewers, not just me. But with season 5 already renewed by Amazon, the first plot details reveal that we're heading away from the city once again.

But what has Prime Video actually told us about what to expect from Reacher season 5? We're breaking down everything from release date speculation to potential cast below.

Reacher season 5 release window predictions

(Image credit: Prime Video)

With no production schedule currently public, it's going to be a little while before Reacher season 5 hits our screens.

Based on season 4's filming calendar, which ran from June-November 2025, we can guess that the very earliest we'd see season 5 episodes is late 2027.

While season 4 aired from August 2026, seasons 1-3 dropped anywhere from December to February, which again, could influence when we might see season 5.

Renewal news came in May 2026, with Peter Friedlander, Head of Global Television, Amazon MGM Studios, explaining in a statement: “The series’ ability to combine high-octane action with compelling character storytelling continues to resonate with tens of millions of viewers around the world.

"We’re excited to move forward with a fifth season ahead of Season 4's debut and to build on this incredible momentum.”

Reacher season 5 trailer: is there one?

Don't be silly. It took us until the end of July 2026 to get one for season 4, so expect the first full trailer to arrive around two weeks before the fifth season does.

Reacher season 5 cast

Michelle Pfeiffer leads our new cast. (Image credit: Prime Video)

As Reacher's supporting cast tends to change on a season-to-season basis, the only person we can guarantee will return for season 5 is Reacher himself: Alan Ritchson.

However, past cast members such as Neagley (Maria Sten) appeared in multiple seasons of the show, so there's currently nothing stopping the below season 4 cast from returning:

• Alan Ritchson as Jack Reacher
• Sydelle Noel as Tamara Green
• Christopher Rodriguez-Marquette as Jacob Merrick
• Kevin Corrigan as Detective Shaun Docherty
• Kevin Weisman as Russell Plum

The above hugely depends on how closely show creators will follow the original novel's source material. We can quite easily rule out Neagley's return now that she has her own spinoff.

New faces will likely be announced as filming kicks off, so watch this space.

Reacher season 5 plot predictions

(Image credit: Prime Video)

We know that Reacher season 5 will adapt Child's novel Make Me, which is the 20th installment in his Jack Reacher series.

In short, Reacher arrives at a remote, quiet prairie stop named Mother's Rest. There he meets Michelle Chang, a former FBI agent turned private investigator who is searching for her missing partner, Keever. Together, they trace clues across Chicago, Los Angeles, and the Deep Web, uncovering a horrific online operation.

With the local townspeople hostile and incredibly cryptic, it feels as though the season's tone will be more similar to seasons 1 and 3. However, the ties to bigger cities — particularly Chicago — provides a potential gateway for old characters to return.

At this stage, we're running before we can walk. I'm just glad that once again, the small town feeling wins out.