News
- Chinese firm turns RTX 5090 cards into two-slot AI accelerators using custom PCBs
- Factory automation in China converts Nvidia gaming GPUs into data center hardware
- Export restrictions are bypassed as RTX 5090 boards are reworked into rack-ready AI modules
A video circulating on Bilibili shows how enterprising Chinese technicians are converting GeForce RTX 5090 graphics cards into high-efficiency AI accelerators, bypassing export restrictions in the process.
Instead of using the retail GPUs as is, a local manufacturer is removing the GB202 chips and memory modules from the gaming boards.
The components are then transferred onto custom two-slot PCBs and paired with blower-style cooling, optimized for server rack environments.
Large scale productionThe operation appears to be being run at factory-level scale. Robotic arms extract GPU modules, pick-and-place systems reassemble them onto redesigned boards, and automated soldering equipment finalizes the electrical work, before visual inspection systems confirm alignment and build quality.
The process is fast, repeatable, and efficient, far beyond what might be expected of a small workshop, but what makes this enterprise even more notable is the sheer volume of RTX 5090 cards seen in the video, despite restrictions on their export to China.
Nvidia has not officially released the 5090D variant, designed to comply with U.S. sanctions, and yet shelves of full-power RTX 5090 units can be seen undergoing retrofitting.
Gaming models with three-fan cooling are too large for server chassis, and their power connectors are often positioned awkwardly for rack use, so the cards are physically modified to fit data center constraints.
The redesigned blower cards shift the power input to the tail end of the PCB, simplifying cable routing and avoiding risk of overheating.
From a hardware standpoint, the GPUs function identically.
By shifting them onto slimline, server-compatible boards, Chinese firms are essentially creating their own high-end AI accelerators.
If hardware intended for gamers can be so easily adapted for AI research and model training, then questions about the effectiveness of the current export controls surely have to be asked.
It’s clear from the scale of the operation in the video that there is strong demand for AI accelerators built from repurposed RTX 5090 cards.
For AI developers in China, it means they can get the computing power they need without having to rely on official supply channels.
Via Guru3D.com
You might also like- Amazon's CEO teased advertisements might be coming to Alexa+
- Other Alexa+ subscription tiers might be on the way, too
- No release date yet, but it sounds like a matter of when, not if
Amazon’s Alexa+ may have only been around for a short time, and while that would usually mean it’s in the rapid user-growth stage, it sounds like Amazon CEO Andy Jassy is already planning its demise: stuffing its conversations with ads.
Speaking at Amazon’s most recent investor call, Jassy explained that “People do a lot of shopping [with Alexa+]”
Adding “I think over time, there will be opportunities, as people are engaging in more multi-turn conversations, to have advertising play a role to help people find discovery, and also as a lever to drive revenue.” Basically, if you ask Alexa+ for shopping recommendations, some of its responses might eventually be sponsored.
Jassy went on to discuss the possibility of different payment tiers beyond the existing $19.99 a month model – it is already free with Amazon Prime – suggesting we may see an ad-supported and ad-free tier, but the idea sounds yucky. Right?
Because if AI’s responses can be paid for, why would you trust anything it says ever again?
Whatever Amazon is paid to tell you they eat. (Image credit: Amazon)We knew this was comingAI has sold itself as an unbiased guru that can answer many of life’s questions, but it’s increasingly clear how untrue that is.
After the bot couldn’t stop correcting its master, X AI’s Grok received some major changes leading to very public (and hate-ridden) crashouts, and now it seemingly serves as a mouthpiece for Elon Musk rather than any semblance of a reliable third-party.
The Chinese-run DeepSeek bot has been found to tow the Chinese Communist Party line on topics like Taiwan independence and Tiananmen Square, and I’m sure there are plenty of more subtle biases built into every AI chatbot out there.
Advertisements would bring in another layer of mistrust, especially depending on how they’re implemented.
In Alexa’s case, if it recommends three products and a fourth clearly marked as sponsored, why would you be incentivized to buy it? As you know, Alexa is suggesting it only because it was paid to do so. You’re much better off picking one of the other genuine recommendations.
Alternatively, if it instead prewarns you that some answers are sponsored, but doesn’t then label which answers were bought, you wouldn’t be able to trust that any of its shopping suggestions aren’t simply being paid for.
But ads were inevitable in many ways.
We offer unbiased tech buying advice, unlike AI (Image credit: Future)The race to the bottom on subscriptions means that many people don’t want to pay much for AI tools. Simultaneously, investors in their quarterly calls to big tech companies are starting to ask how the billions being invested in AI might eventually be turned into profit.
The now-not-so-quiet part that Amazon’s CEO just said aloud is advertising. People can have a cheap AI service, and the company can get paid.
I don’t know about you, but that doesn’t seem any different from the systems we already have through sponsored entries in Google or on Amazon’s store.
I was promised a Jarvis-like AI agent, not the same as what we have, but in a different wrapper.
Perhaps I’m getting tired of the perpetual AI hype, but with sponsored conversations looking like they’re headed our way, we might start to see some AI systems get ruined and die off – I can only hope.
You might also like- OpenAI has removed the ChatGPT feature, allowing people to search through public conversations with a search engine
- Many users learned too late that enabling the “discoverable” setting could make chats accessible to anyone online
- The decision came after several people saw their sensitive and private information publicized
OpenAI has abruptly shut down a feature in ChatGPT that allowed chats shared publicly to be searchable through Google. Users were unpleasantly surprised to discover that private information shared with ChatGPT was now publicly available as a Google search result.
Dane Stuckey, OpenAI’s Chief Information Security Officer, announced the shutdown on Thursday, with the feature disabled on Friday morning. OpenAI has also begun scrubbing search engine indexes of the previously shared conversations.
OpenAI designed the “Make this link discoverable” checkbox, which they had clicked on, to give more people access to potentially helpful conversations between ChatGPT and different users. In practice, it opened the door to a new kind of digital exposure, where criminal confessions, personal secrets, and corporate practices were just a few keywords away from turning up in a Google Search.
This was not some obscure opt-out setting hidden in a developer console. It was right there in the sharing menu of any ChatGPT conversation users chose to turn into a public link. However, while OpenAI believed the wording was clear about what users were doing, many users apparently did not understand. They thought the link would remain private or unfindable unless explicitly shared.
For the average person using ChatGPT to plan a resignation letter, troubleshoot a relationship, or work on a legal strategy, that’s more than a technical hiccup. It’s a gut check. If you’ve ever shared a link to a conversation, there’s a nonzero chance someone found it by a Google search.
Some of the reported indexed conversations included internal job applicant evaluations, sensitive mental health disclosures, confessions of crime, and proprietary software code. People were talking to ChatGPT as if it were private, but for anyone who toggled that setting, it wasn’t.
We just removed a feature from @ChatGPTapp that allowed users to make their conversations discoverable by search engines, such as Google. This was a short-lived experiment to help people discover useful conversations. This feature required users to opt-in, first by picking a chat… pic.twitter.com/mGI3lF05UaJuly 31, 2025
Searchable secretsTo OpenAI’s credit, the company pulled the feature quickly. But the fallout raises important questions not just about privacy, but about how much users can reasonably be expected to understand about the tools they’re using. The “discoverable” checkbox didn’t hide behind a wall, but it also didn’t do a great job of communicating the scope of its consequences. It’s one thing to share a link, but indexing it in global search engines indefinitely is something else.
It is fascinating what this says about how users behave with AI. They confide in it, test ideas, vent frustrations, and draft things they’d never type into a search bar. If you’ve ever shared a ChatGPT conversation, it’s worth checking whether that link is still live.
While OpenAI is working with Google and other search engines to purge previously indexed content, search crawlers have long memories. Some content may linger for a while, even if deleted. OpenAI disabled the feature quickly, but the damage may already be done for some users.
Hopefully, the lesson will stick with the public, much like the idea of not deleting embarrassing browser histories has transitioned from a common joke to something only the most clueless person would forget to do. The magic of tools like ChatGPT lies in how they create the illusion of a conversation. But if you forget that it is still an illusion, you might not notice risks like buttons that send your digital heart-to-heart straight to Google.
You might also like- The Honda N-One e: is unveiled in Japan and previews a global compact EV
- The diminutive car will measure less than 3,400mm (133.9inches) in length
- A range of 150+ miles is expected and V2L means it can power homes
The production version of the teeny, boxy Honda that wowed crowds at the recent Goodwood Festival of Speed has been unveiled and the N-One e: will go on to influence a global, low-cost EV.
Revealed in Japan this week, one of Honda’s smallest-ever electric vehicles will offer over 150 miles (245km) of range on a single charge from its equally small battery packs, while the single electric motor is expected to produce around 63hp.
Designed predominantly for town and city driving, the Japanese version of Honda’s Super EV concept complies with the country’s strict Kei Car rules, which mean it has to measure less than 3,400mm (133.9inches) in length.
It is likely that the European version will sport chunkier bodywork so it isn’t completely dwarfed by the myriad overweight SUVs, but we can expect to see a lot of the features carried over from the Japanese Domestic Market model.
(Image credit: Honda Japan)The interior, for example, is packed with practical touches, such as rear seats that fold upright or flat to provide ample stowage space for numerous, odd-shaped items.
There are also hard-wearing, easy-to-maintain fabrics and charging points for smartphones.
Honda also says that the N-One e: is capable of one-pedal driving and Vehicle 2 Load (V2L), which means owners can power homes during power outages, or simply top up a power-hungry laptop when on the move.
Small EVs are proving a big hit(Image credit: Honda Japan)While it is highly unlikely the N One e: (or whatever the Super EV concept will eventually be called) will make its way to the US market, Honda has said that it plans to introduce a small city EV to Europe and beyond next year.
The company was one of the first to market with a funky, compact EV game thanks to its innovative Honda e. Unfortunately, its cutting-edge styling couldn’t make up for its 100-odd mile real-world range, slow charging speeds and high sticker price.
The Super EV concept comes at a time when battery technology has improved, so the range could be a more usable 170 miles and the asking price will likely be more in line with something like the BYD Dolphin Surf, which starts at £18,650 in the UK (around £25,000 / AU$38,250).
Honda is synonymous with practical, affordable, compact vehicles, with the likes of the Jazz remaining popular among savvy buyers that want reliable, simple transport that laps up short hops without breaking the bank.
While the N One e: is predominantly aimed at the Japanese and Chinese market for now, we expect to see what a more global version will look like at the IAA motor show in Munich later this year.
Honda is also on an EV push in general, with the recent launch of the CUV e: battery-swap electric scooter and the announcement of a full-size electric motorcycle coming soon.
The thread that appears to be running through all of the products is Honda's commitment to excellent engineering solutions and a competitive price point. We might start to see battery-powered models that can compete on price with their gas-powered counterparts.
You might also like- I've ridden Honda's latest electric scooter with clever battery-swap technology – and it might just convince commuters to give up on petrol
- Honda unveils striking new 0 Series that completely rethinks its EV future
- Electric cars can be fun – the 6 hottest new EVs from the Goodwood Festival of Speed 2025
- Cheats and mods are now frontlines for cybercrime targeting gamers' wallets and private data
- Verified crypto wallets like MetaMask and Exodus are being drained through browser injection
- Trojan.Scavenger abuses overlooked flaws to disable browser safety and manipulate trusted extensions
Gamers seeking performance enhancements or special abilities through third-party patches and mods may be unwittingly exposing themselves to sophisticated malware, experts have warned.
Recent findings from Dr.Web revealed a malware family known as “Trojan.Scavenger” which targets Windows users by disguising itself as cheats or enhancements for popular games.
This seemingly harmless mod can ultimately compromise crypto wallets, password managers, and web browsers, posing serious risks to user privacy and digital assets.
When cheats become covert threatsThe infection chain begins when users download ZIP archives claiming to improve performance in games including the likes of Grand Theft Auto 5 or Oblivion Remastered.
These archives contain modified dynamic libraries, sometimes renamed with extensions like .ASI to resemble legitimate plugin formats.
When the user follows the installation instructions, the malicious library is placed in the same folder as the target game. If the game does not properly validate its libraries, the trojan loads automatically at startup.
In some cases, flaws in library search priorities are essential to the malware’s success, allowing it to hijack execution within the host application.
Once loaded, the malware establishes contact with a command-and-control server using encrypted communication. This process includes verifying encryption keys and checking timestamp consistency, which is meant to evade analysis and block antivirus detection.
The malware doesn’t stop with the initial payload. In more complex infections, it deploys additional trojans that embed themselves in Chromium-based browsers like Chrome, Edge, Opera, and Yandex.
These trojans interfere with browser sandboxing, disable extension verification, and replace legitimate extensions with modified versions.
Crypto wallets such as MetaMask and Phantom, as well as password managers like Bitwarden and LastPass, are among the affected applications.
Modified extensions collect mnemonic phrases, private keys, and stored passwords, which are then transmitted to the attackers' servers.
Exodus, a popular crypto wallet, is also targeted using similar techniques.
By exploiting library loading behavior, the malware extracts sensitive JSON entries, including passphrases and seed data required for generating private keys.
How to stay safeTo stay safe, always apply caution when accessing unofficial content.
Avoid downloading mods or cheats from sketchy forums or unverified sources, especially those shared on torrent platforms or through poorly moderated social media channels.
Antivirus software, while helpful, must be regularly updated to stay effective against evolving threats.
Android antivirus tools may protect mobile platforms, but on desktop systems, more capable solutions are necessary.
Good social media management also helps reduce exposure to malicious content. Limiting interaction with communities known for spreading cracked software or shady patches can lower the risk.
Lastly, checking file paths, verifying digital signatures when available, and restricting admin privileges on daily-use accounts can make it harder for malware to execute successfully.
You might also like- These are the best internet security suites available
- Take a look at our pick of the best best VPNs with antivirus you can use right now
- Hackers says they attacked Mailchimp and stole user data - and the community laughed it off
- Labels like “Verified” give a false sense of safety but don’t reflect real extension behavior
- Browser DevTools were never meant to track how extensions behave across tabs and over time
- Malicious extensions often act normally until specific triggers make their hidden features come alive
The unchecked spread of malicious browser extensions continues to expose users to spyware and other threats, largely due to deep-seated flaws in how the software handles extension security.
New research from SquareX claims many people still rely on superficial trust markers like “Verified” or “Chrome Featured,” which have repeatedly failed to prevent widespread compromise.
These markers, while intended to reassure users, often offer little insight into the actual behavior of an extension.
Labels offer little protection against dynamic threatsA central issue lies in the limitations of Browser DevTools, which were designed in the late 2000s for web page debugging.
These tools were never meant to inspect the far more complex behavior of modern browser extensions, which can run scripts, take screenshots, and operate across tabs, actions that existing DevTools struggle to trace or attribute.
This creates an environment where malicious behaviors can remain hidden, even as they collect data or manipulate web content.
The failure of these DevTools lies in their inability to provide telemetry that isolates extension behavior from standard web activity.
For instance, when a script is injected into a web page by an extension, DevTools lack the means to distinguish it from the page’s native functions.
The Geco Colorpick incident offers an example of how trust indicators can fail catastrophically - according to findings from Koi Research, 18 malicious extensions were able to distribute spyware to 2.3 million users, despite carrying the highly visible “Verified” label.
To address this, SquareX has proposed a new framework involving a modified browser and what it calls Browser AI Agents.
This combination is designed to simulate varied user behaviors and conditions, drawing out hidden or delayed responses from extensions.
The approach is part of what SquareX terms the Extension Monitoring Sandbox, a setup that enables dynamic analysis based on real-time activity rather than just static code inspection.
At the moment, many organizations continue to rely on free antivirus tools or built-in browser protections that cannot keep up with the evolving threat landscape.
The gap between perceived and actual security leaves both individuals and companies vulnerable.
The long-term impact of this initiative remains to be seen, but it reflects a growing recognition that browser-based threats demand more than superficial safeguards.
You might also like- Google Workspace is copying a very familiar YouTube feature to help you get through videos
- Check out the best AI phones on the market
- Here is our list of the best AI website builders on the web
- Wi-Fi 8 plans to fix edge-of-coverage dropouts where extenders usually fail to keep things running smoothly
- Wi-Fi 7 maxes out throughput while Wi-Fi 8 makes sure it works everywhere, every time
- Wi-Fi 8 to introduce Single Mobility Domains so devices roam between access points without getting interrupted
Only hours after the official ratification of Wi-Fi 7, early details of its successor, Wi-Fi 8, are already making waves.
A release from Qualcomm noted Wi-Fi 8 isn’t promising faster peak speeds; instead, it focuses on improving stability, cutting latency, and ensuring smoother performance in environments with many connected devices.
Where Wi-Fi 7 emphasized raw throughput and bandwidth gains, Wi-Fi 8 is being designed to sustain those benefits under pressure, ensuring consistent delivery even in crowded or interference-prone settings.
Rethinking the role of Wi-FiTwo core trends are shaping the direction of Wi-Fi 8. First is the growing ecosystem of personal devices such as AR glasses and next-gen health monitors, which demand seamless and low-latency connections to nearby companion devices.
Second is the rise of AI-powered systems that depend on fast, reliable access to edge or cloud-based intelligence.
Together, these trends are pushing local networks, including your Wi-Fi router or even a mobile 5G router, into territory they weren’t originally designed for.
Wi-Fi 8 is being framed as the foundational layer of connectivity that will underpin these increasingly dynamic and latency-sensitive systems.
And unlike a basic Wi-Fi extender that simply expands signal reach, the new standard rethinks how access points work together, how devices roam, and how signal quality is preserved at the edge of coverage zones.
What gives Wi-Fi 8 its edge is the ongoing development of the 802.11bn standard.
Several core features aim to make Wi-Fi more reliable in the real world, not just in ideal lab conditions.
One of the key innovations is “Single Mobility Domains,” which allow devices to move across multiple access points without experiencing disruptions, something particularly beneficial in environments like airports, hospitals, or multi-floor offices.
Another major improvement is performance at the edge, the outer limits of a signal’s reach.
Current setups, even with a Wi-Fi extender, often suffer drops and delays in these areas - Wi-Fi 8 introduces physical layer updates meant to improve consistency without relying on brute-force signal strength.
Multi-access point coordination is another shift, and this is ideal for large venues or dense urban housing - as rather than each unit operating in isolation, Wi-Fi 8 envisions coordinated networks that intelligently share airtime and avoid overlap.
On-device coexistence, especially where radios like Bluetooth or ultra-wideband compete for antenna space, is also receiving attention.
Wi-Fi 8, expected to be finalized by 2028, reflects a deeper rethink of what modern connectivity actually requires.
The push isn’t for headline-grabbing throughput figures but for a wireless experience that performs like wired infrastructure.
You might also like- Check out our pick of the best secure routers on offer
- Here are the best mobile workstations around today
- We've also listed the best monitors for every budget and resolution
This week DJI and Insta360 had a Freaky Friday situation as they announced a 360 camera and a drone respectively, tech the other is usually known for.
We also heard that OpenAI is scared of its next AI, and the UK faced the Online Saftey Act.
You can catch up on these stories and more in this week's catch up of the seven biggest tech news stories.
7. We Found My Skechers(Image credit: Skechers)This week, Skechers debuted the new Find My Skechers. They look like your run-of-the-mill sneakers on the outside, but hidden in each heel is a cutout that’s perfect for an Apple AirTag.
This Bluetooth tracker can then be used to help you keep track of your child if you’re out in an unfamiliar place, or help you locate any lost shoes, whether they were misplaced while out and about or while getting ready in the mad rush to get to school.
Though when we polled parents, the reactions were mixed. Predominantly because the target age range – toddlers and young children – outgrow their shoes every six months to a year. At their cheapest, Find My Sketchers cost $52 a pop, which is a lot to regularly shell out.
6. The UK faced the Online Safety Act(Image credit: Getty Images)This week Britain was hit by the latest provisions of the Online Safety Act, introducing age verification measures designed to help prevent young people accessing age inappropriate content. However the new requirements have frustrated many adults.
A petition to repeal the UK Online Safety Act garnered over 450,000 signatures in just a few days, and some have tried all sorts of workarounds including attempting to game the system with Death Stranding’s photo mode.
Most people are onboard with the idea behind the rules, but aren’t keen on needing to share their ID with websites – especially foreign websites – as they’re worried their data might be mishandled. Despite the concerns, the Government has said it won't be repealing the measures.
5. Google Home's crisis got worse(Image credit: Future)Last week, we reported on a growing number of complaints about Google Home devices, with the smart home gadgets failing to deal with commands properly or throwing up other bugs.
This week, it seems the problem got worse, with many reporting their voice controls for smart lights were now completely broken and not working at all.
The situation has gotten so bad that US law firm Kaplan Gore has announced it has "begun investigating a possible class action" against Google because so many users are now reporting broken functionality, despite most of these devices having worked fine in the past.
- Read more: Google Home devices are continuing to break
The 360 camera space has been dominated by Insta360 for years, but it finally has some proper competition following the arrival of the DJI Osmo 360. What's more, they could soon be joined by the GoPro Max 2, which was recently teased by GoPro.
Considering it's DJI's first foray into the market, the Osmo 360 impresses straight off the bat, delivering what DJI says are several industry firsts, the most notable of which are its twin 1-inch sensors with 8K video recording up to 50fps. Those sensors are effectively twice the size of the Insta360 X5's, our current favorite 360 camera, and should give DJI's model the upper hand for image quality, especially with richer 10-bit color depth.
Based on specs, the Osmo 360 has the edge over the X5, but it's real-world use that matters the most, and the X5 is our favorite model for a reason. It has stellar battery life, and its lenses can be replaced in the event of damage. We are currently conducting our hands-on versus of the two models, and it's making for a titanic battle.
- Read more: Can DJI's first 360 camera dethrone the X5?
DJI isn’t used to having competition in the skies, but this week, its first serious rival for years flew into town. The undisputed king of 360 cameras, Insta360, has announced its new Antigravity platform, which will soon give us the first drone with a 360-degree camera.
Okay, but what’s the benefit of a flying, all-seeing camera? The main one is that, in theory, you won’t need to worry about perfectly shooting aerial videos as you fly, because you’ll be able to reframe them later. We’re looking forward to experiencing that in person when the Antigravity drone fully launches later this month.
2. ChatGPT helped us study(Image credit: OpenAI)ChatGPT has added a new feature called Study mode to its paid-for and free accounts that turns the famous chatbot into an unpaid study tutor for your kids. Parents often panic about ChatGPT being used by their children to avoid actually learning anything, but Study mode is different. It’s pretty safe for kids to use because it teaches them how to arrive at the answer, rather than just giving them the answer. It’s more of an interactive back-and-forth chat about the question being posed, which encourages them to learn as they go.
Of course, you don’t need to let your kids have all the fun with it. You can use Study mode yourself to learn any new subject. Having the ability to turn ChatGPT into a proper tutor is pretty handy, so whatever it is you want to learn – whether that’s a new language or how to code – now you’ve got the perfect excuse to get started.
1. OpenAI was scared by its AI(Image credit: Shutterstock/ DIA TV)OpenAI CEO Sam Altman revealed details about his company's testing of GPT-5 (the next-gen brain for ChatGPT) by saying he got scared by the AI, and comparing it to The Manhattan Project.
The over-the-top analogy might evoke some level of keen scientific minds coming to a major breakthrough, but his other comments made OpenAI seem either reckless or incompetent.
Even if he's exaggerating, we don't know if either of those descriptions should be attached to the kind of company that might decide how next-gen AI power is deployed.
- Read more: OpenAI's CEO says he's scared of GPT-5
- Officially wrapped filming
- No trailer released yet
- Main cast predicted to return
- Plot to follow Mick Herron’s 'Joe Country' and 'Slough House' novels
- Season 7 expected to begin shooting in September/October
- Slough House spin-off, Down Cemetery Road, in the works
Slow Horses season 6 has already finished filming, which is huge news considering season 5 isn't even out on Apple TV+ until September 24.
The gritty spy drama is getting a bit of a name for itself, not only for its thrilling action-packed scenes, but its incredible release schedule never leaving more than a year between seasons.
And it appears there's plenty more where that came from as Jackson Lamb and his team of down-and-out MI5 agents are returning for season 5, 6 and 7 over the next couple of years.
Adapted from the books by Mick Herron, Slow Horses season 6 already has a plot to follow and filming in the can, but there's plenty more to talk about when it comes to trailer, expected release date, predicted case, news, rumors and more. Though, I'll be sure to update here once season 5 hits Apple TV+.
Potential spoilers follow for Slow Horses seasons 1 to 4.
Slow Horses season 6 release dateA post shared by Apple TV (@appletv)
A photo posted by on
Apple TV+ renewed Slow Horses for season 6 just days after the season 4 finale – and, according to Deadline, the cast and crew have already finished filming.
In an official statement from Apple TV+ on the show's renewal, Jay Hunt, creative director for Europe said: "Audiences around the world have fallen in love with the Slow Horses, and I'm delighted that Gary Oldman will be leading this star-studded cast on another acerbic and action-packed adventure".
So, when it comes to the Slow Horses season 6 release date, season 4 released in September 2024 and season 5 is scheduled for September 2025. With season 6's filming done and dusted, I think I can quite confidently say that season 6 is looking at a fall 2026 release date.
Slow Horses season 6 trailer: is there one?There's no Slow Horses season 6 trailer just yet (Image credit: Apple TV Plus)Considering Slow Horses season 5 is yet to receive its own trailer, it seems incredibly unlikely that we'll get a season 6 trailer any time soon.
But, we were treated to a season 5 teaser look at the end of the season 4 finale. If that's the case this time round, mark October 22 on your calendar for the season 5 finale for a potential first look at season 6. I'll be sure to update here when it drops, too.
Slow Horses season 6 predicted castGary Oldman will reprise his titular role as Jackson Lamb (Image credit: Apple TV Plus)Full spoilers follow for Slow Horses season 5.
Slow Horses season 5 will air six episodes on Apple TV+ between September 24 and October 22. Since we don't know what happens yet (and we don't want to spoil it by revealing the plot of Mick Herron's 'London Rules' novel), we can only predict the cast to appear in season 6, with one known addition:
- Gary Oldman as Jackson Lamb
- Jack Lowden as River Cartwright
- Kristin Scott Thomas as Diana Taverner
- Saskia Reeves as Catherine Standish
- Christopher Chung as Roddy Ho
- Aimee-Ffion Edwards as Shirley Dander
- Rosalind Eleazar as Louisa Guy
- Ruth Bradley as Emma Flyte
- James Callis as Claude Whelan
- Tom Brooke as J.K. Coe
- Hugo Weaving as Frank Harkness
- Joanna Scanlan as Moira Tergorian
- Jonathan Pryce as David Cartwright
- Lenny Rush as TBC
- Nick Mohammed as TBC
- Hiba Bennani as TBC
So, there's a new cast member that we know for certain is joining Slow Horses season 6 and that's Doctor Who's Lenny Rush who, according to the Radio Times, was spotted filming with Oldman in January. Other than that, I'm sure it'll be full of returning favorites.
Slow Horses season 6 potential plot synopsis and rumorsSlow Horses season 6 will be adapted from 'Joe Country' and 'Slough House' (Image credit: Apple TV Plus)Major spoilers follow for Slow Horses seasons 1 through 4.
Slow Horses season 6 is confirmed to be the first season that will follow not one, but two, of Mick Herron's novels. As confirmed by Deadline, these will be 'Joe Country' and 'Slough House', the sixth and seventh books in the series.
And the official synopsis from Apple TV+ for season 6 is as follows: "Season six sees the Slow Horses on the run as Diana Taverner embroils them all in a fatally high-stakes game of retaliation and revenge".
To delve deeper into the plot, it's also worth taking at the synopsis's for the books. For 'Joe Country' that's: "In Slough House, the backwater for failed spies, memories are stirring, all of them bad. Catherine Standish is buying booze again, Louisa Guy is raking over the ashes of lost love, and new recruit Lech Wicinski, whose sins make him outcast even among the slow horses, is determined to discover who destroyed his career, even if he tears his life apart in the process.
"With winter taking its grip Jackson Lamb would sooner be left brooding in peace, but even he can't ignore the dried blood on his carpets. So when the man responsible for killing a slow horse breaks cover at last, Lamb sends his crew out to even the score.This time, they're heading into joe country. And they're not all coming home."
The team at Slough House return for more espionage drama (Image credit: Apple TV Plus)And for 'Slough House': "Slough House - the crumbling office building to which failed spies, the 'slow horses', are banished - has been wiped from secret service records.
"Reeling from recent losses in their ranks, the slow horses are worried they've been pushed further into the cold, and fatal accidents keep happening.
"With a new populist movement taking a grip on London's streets, the aftermath of a blunder by the Russian secret service that left a British citizen dead, and the old order ensuring that everything's for sale to the highest bidder, the world's an uncomfortable place for those deemed surplus to requirements.
"The wise move would be to find a safe place and wait for the troubles to pass.But the slow horses aren't famed for making wise decisions."
Much like the rest of Slow Horses, season 6 will also be six episodes long and while I'd usually take a look at the previous season to give an insight into the plot, it doesn't air until September/October.
I'll be sure to update here as soon as the story at Slough House has unfolded once more.
Will Slow Horses get more seasons on Apple TV+?Slow Horses season 7 is officially on the way (Image credit: Apple TV Plus)Slow Horses season 7 was officially confirmed by Apple TV+ in July and Deadline reported in June, whilst in conversation with Oldman, that "season 7 will begin shooting in the UK in late September or early October".
With season 5 following books 6 and 7 in Mick Herron's Slough House series, season 7 will follow 'Bad Actors'. There's also the soon-to-be published 'Clown Town' coming September 9, which could hint at at least one more season to come, but any news on this is unconfirmed.
But, there's also a Slough House spin-off series in the works titled Down Cemetery Road, based on Herron's first novel of the same name.
With filming already happening, The Hollywood Reporter revealed it will star both Emma Thompson and Ruth Wilson as Sarah Tucker and Zoë Boehm respectively.
Apple TV+'s creative director, Hunt, said of the news: "Down Cemetery Road has all the hallmarks of Mick Herron's funny and acerbic writing, and I'm delighted we will be bringing it to life for Apple TV+ with such a stellar cast".
For more Apple TV+ coverage, read our guides on Ted Lasso season 4, Foundation season 3, Presumed Innocent season 2, and Silo season 3.
The UK's Online Safety Act officially introduced age verification measures on July 25, reshaping how people access certain websites and services.
Designed to protect children from harmful content and make platforms more accountable, the rollout has sparked heated debate. Critics worry about increased tracking, potential data breaches, and the loss of online privacy.
With new restrictions in place, VPN interest in the UK has surged as users look for greater control over their browsing and data security. While we don't encourage bypassing age checks, it's clear that VPNs have become a go-to tool for those concerned about privacy and the changing internet landscape.
We've tested countless VPNs over the years – you’ll find our full list of recommendations to the best VPNs and best free VPNs in our dedicated guides.
But if you just need a quick answer to which VPN to choose this weekend, we’ve got you covered here. These are the top options available today, plus a heads-up on what to look for in the small print.
1. NordVPN: best for most peopleNordVPN still sits at the top of our rankings because it’s the best all-in-one VPN we’ve tested, delivering world-class security and speed, plus versatility from its extra features. Yes, Surfshark is slightly faster, and ExpressVPN just beats NordVPN on privacy and security in our most secure VPN guide, but at $3.39 / £2.69/month (on the two-year plan) NordVPN is cheaper than Express ($3.99 per month), better at unblocking geo-restricted content (it’s our top recommendation for streaming), and offers more extra features than both rival services.
NordVPN’s Threat Protection feature blocks malware, ads, and trackers well, while its audited no-logs policy means that no browsing data is logged by its servers. (Those independent audits prove that NordVPN is operating exactly as it claims.)
The apps themselves are straightforward to use, making it a good choice for beginners.
In this new era of stricter internet rules, NordVPN is a great choice for most people, and TechRadar has an exclusive deal. Sign up and get:
✅ Up to 76% OFF
✅ Up to $50 Amazon Gift card
✅ 4 months free protection (TechRadar exclusive)
Choose the Pro, Complete, Ultimate or Prime plans to get the deal. There's a 30-day money back guarantee, so if it isn't right you can cancel your subscription and get a refund.
Read more: NordVPN reviewView Deal
2. Surfshark: top budget optionSurfshark is the cheapest VPN in our top three, at $1.99/ £1.89 per month on the two-year plan. It scored slightly lower than NordVPN and Express VPN in our privacy and security tests, but it’s still an extremely robust choice for keeping your data safe - and it beats the other two on speed. You also get unlimited simultaneous connections, so every device in your home can stay protected, which neither NordVPN or Express VPN offer.
Surfshark's CleanWeb tools block ads, trackers, and known malicious sites, while its audited no-logs policy ensures that your browsing remains private. In our experience, speeds remain consistently fast across both nearby and long-distance servers, making it perfectly capable for streaming, video calls, and general browsing without interruptions.
Pricing is one of Surfshark's strongest selling points: plans start at around $2 per month on a two-year subscription, while monthly and annual plans remain competitive for those who prefer flexibility. All plans include a 30-day money-back guarantee, and there's a seven-day free trial so you can see whether it's right for you.
Read more: Surfshark reviewView Deal
3. ExpressVPN: great for beginners, but not cheapIf you're new to VPNs and want something that works flawlessly from the start, ExpressVPN is an excellent choice. Its apps are simple, intuitive, and consistent across devices, so you can connect securely in one click – we’ve tested it with complete beginners and they were able to get it up and running with no issues.
It also scores highly where it counts, on privacy and security, beating NordVPN (only just though, as you’ll see in our most secure VPN guide). ExpressVPN's commitment to privacy is reinforced by multiple third-party audits and, like the two services above, it has a proven no-logs policy, giving newcomers peace of mind that their data isn't being tracked. Its Lightway protocol also optimises speed and reliability, even on slower connections.
One reason why it’s our third recommendation and not our first is that ExpressVPN costs more. Plans start at around $4 per month on a two-year plan. Every subscription includes a 30-day money-back guarantee, letting beginners test its performance risk-free.
Read more: Express VPN reviewView Deal
PrivadoVPN Free: bonus free optionWe don't recommend choosing a free VPN if data protection is your top priority, but if that’s what you’re looking for, then PrivadoVPN Free is one of the better services we’ve tested. It delivers AES-256 encryption, a strict no-logs policy governed by Swiss privacy laws, and industry-standard protocols like WireGuard and OpenVPN. It hasn't yet undergone an independent audit though.
You get 10 GB of full-speed data every 30 days. After that limit is reached, traffic is throttled to around 1 Mbps. That’s fine for light browsing or occasional streaming, not for heavy use.
Despite the cap, speeds during the full-speed window are surprisingly fast, with many tests showing performance comparable to our top-scoring VPNs. It can also unblock major platforms like Netflix and BBC iPlayer.
Still, if you’re serious about protecting your data, we’d recommend signing up for one of the services above.
Read more: PrivadoVPN Free reviewView Deal
Five red flags to avoid when choosing a VPNVPNs don’t always guarantee complete anonymity, but using a reputable one does significantly boost your privacy online.
However, not all VPNs are created equal. While the services we've highlighted above have proven in our testing over the years to be trustworthy and reliable, there are plenty of options that can put your privacy, and your device, at risk.
Exclusively free VPNs are particularly notorious for tracking users or injecting ads to make money, and some paid services cut corners on security or performance.
Here are five big red flags to watch out for when choosing a VPN:
- No independent security or privacy audit
- Vague privacy policies
- Slow connection speeds or frequent disconnections
- Severe data caps that make streaming or gaming impractical
- Limited server locations, restricting your connection options
- An image of the iPhone 17 Air battery has leaked
- We're expecting a capacity of around 2,900 mAh
- All of the iPhone 17 models should launch in September
We're expecting the iPhone 17 series to launch sometime during September, including the iPhone 17 Air – apparently a replacement for the iPhone 16 Plus. Now a new leak purports to show off the battery that will be fitted inside the Air model.
Images of this battery have been posted by well-known tipster Majin Bu (via 9to5Mac), and it comes enclosed in a steel frame, which matches the iPhone 16 Pro approach. Of particular note is how thin the battery is, from front to back.
According to Bu, the battery capacity is around 2,900 mAh, which is close to the 2,800 mAh figure we've heard from earlier rumors. For comparison, there's a 3,561 mAh battery inside the current iPhone 16, so we're looking at a significant drop.
However, given improvements in battery life technology, and tweaks set to be introduced with iOS 26, reports suggest iPhone 17 Air should still give you a day of battery life – even with the ultra-slim profile and reduced capacity battery.
Battery worriesiPhone 17 Air New Battery Design RevealedFull Article:https://t.co/dc9wI7LS3K pic.twitter.com/Y9MbDQDiRIAugust 1, 2025
If the rumors are to be believed, then the iPhone 17 Air is going to measure just 5.5 mm, front to back. That's a lot slimmer than the 7.8 mm iPhone 16, and it even beats the recently launched Samsung Galaxy S25 Edge, which is 5.8 mm thick.
Ever since we heard an iPhone 17 Air was on the way, there have been concerns about battery life. Putting together a super-slim phone is all well and good, but it does mean less room for the battery and other internal components.
However, we know that iPhones can traditionally get more battery life out of less battery capacity, because Apple develops its hardware and software in tandem – so let's hope the usual trick can be repeated with the iPhone 17 Air.
It should be unveiled alongside the iPhone 17, the iPhone 17 Pro, and the iPhone 17 Pro Max during September. Apple hasn't said anything officially, but well-placed sources have pointed to September 8-10 as the most likely dates.
You might also like- Norton Deepfake Protection is part of its Norton Genie AI Assistant on Norton 360 mobile apps
- The tool is available in Norton 360 mobile products in the US, UK, Australia, and New Zealand
- It currently supports English-speaking YouTube clips
Norton has introduced a new feature which aims to help users spot deepfake audio and video content on their smartphones
Norton Deepfake Protection will be part of the Norton Genie AI Assistant on Norton 360 mobile apps - as until now, it was only available on select Microsoft Copilot+ PCs, and was considered an “early access” phase.
Now, Norton is spreading the tool’s availability, allowing even those without AI PCs to defend themselves against sophisticated scams.
(Image credit: Norton)YouTube supportA deepfake video or audio is a media file created using AI to realistically mimic a person’s appearance, voice, or actions, often made by training machine learning models on real footage or recordings.
This type of content was already observed being abused in different scams, and Norton says the tool provides, “an added layer of contextual protection” by spotting inconsistencies or faint deformations in the physical features of people appearing in videos.
For now, the tool only supports English-language YouTube videos, but there are plans to expand platform and language support in the future.
Anyone interested in testing the tool can upload YouTube links to the assistant to receive “real-time guidance on the authenticity of the video”.
The tool is available in Norton 360 mobile products in the US, UK, Australia, and New Zealand, across Android and iOS devices. Desktop support is “coming soon,” Norton added.
One of the most widespread deepfake scams on the internet is the famous “Elon Musk investment scam”, in which crooks created a deepfake of Elon Musk promoting his new cryptocurrency coin. In the video, he talks about the coin being the future of money transfer, and promises “huge gains” for the investors.
While early iterations of the scam were rather easy to spot, AI has improved significantly over the years, and will continue to do so, further blurring the lines between truth and deception.
You might also like- Norton boosts AI scam protection tools for all users
- Take a look at our guide to the best authenticator app
- We've rounded up the best password managers