A few years ago, France introduced a mandatory repairability score for consumer goods like laptops and tablets. It involves five criteria that range from documentation and availability of spare parts to ease of disassembly, with the manufacturer using a government-provided checklist to determine their score.

Recently Asus determined that their Asus ROG Flow Z13 – model GZ302EA – scored a 10 out of 10 using this system. This led [iFixit] to run the same tablet/laptop hybrid through their own rating system.

You can find the filled-out spreadsheet for this device here, with this Asus-provided site showing a list of devices that all score a 10/10 or a measly 9.9/10 according to this system. As a self-reported score it is hard to take it as the objective truth, as there is every incentive for the manufacturer to tweak the truth to their own benefit and gloss over inconveniences. This is where it’s interesting to compare it with [iFixit]’s 7/10 score.

On documentation, Asus gives itself a perfect score but [iFixit] finds it to be incomplete. Removal of one fan requires the disassembly of the cooler with its liquid metal thermal interface on the CPU. The wireless card, and most ports, are soldered to the mainboard. On the bright side, after you get the screen off, the insides are quite modular, which is a plus.

[iFixit] dings three points: for documentation, soldered-down components, and a fan accessibility glitch. Parts accessibility outside of France is also significantly harder, but one can hardly blame the French system for that. Overall the French self-reported rating would seem to be a fair start, but depending on which criteria you define as required you may find yourself disagreeing with the score.

In the case of LPDDR5 RAM one could argue for example that with LPCAMM2 modules soldering RAM onto the mainboard ought to be a thing of the past, and Wi-Fi modules should always be removable as well. You can take that up with the French regulators.

The co-inventor of Apple’s FaceID and Vision Pro technology has spent the last six years building a frontier artificial intelligence model that could one day help decode electrical activity in the brain to diagnose cognitive disorders.

Now, Gidi Littwin’s startup, Hemispheric, has raised $52 million in funding after gathering data on 100,000 people’s brains to train deep learning models to examine the brain without the need for invasive procedures.

Littwin left Apple in 2020, looking for a change. He found it when his Hemispheric cofounder Hagai Lalazar cold-messaged him on LinkedIn. Lalazar had begun to develop artificial intelligence to study the brain without the need for surgery, and was looking for a commercially minded cofounder to drive the company forward. By the time he found Littwin, he had spoken to around 75 candidates.

Littwin had helped develop FaceID, and at that time was working on hand-tracking for an augmented reality product, the Vision Pro. As part of this, he had to collect what he told WIRED were “hundreds of thousands of subjects’ worth of data” to train the deep learning models powering the technology.

“There were massive data collection operations behind these projects and we knew we had to build something very similar at Hemispheric,” Littwin says, “and we have.”

Because each individual’s brain activity looks different, doctors have largely had to rely on subjective questionnaires and behavioral observations to diagnose depression, Alzheimer’s, and Parkinson’s. To get around that, Littwin and Hagai collected their “most prized possession:” a quarter of a million hours of brain data from 100,000 paid volunteers across Asia, as well as Tel Aviv and Boston. Subjects undertook a series of activities that look like games but activated different parts of their brains.

That data helped train a frontier model, which infers brain function from electrical activity within the skull in the same way that large language models deduce meaning by statistically analyzing text. They then tested the generalized model on subsets of people, including those diagnosed with PTSD, schizophrenia, and depression and said the model made accurate deductions about the individuals’ brain health. The team is currently working on a clinical study to test whether their model can diagnose and even predict Alzheimer’s.

The team will submit their first product, which will be used to study PTSD, to the FDA for approval early next year. They hope that will allow them to roll the product out to the public later in 2027.

To help diagnose a cognitive disorder, a patient wears a lightweight EEG headset that measures electrical activity in the brain for around 15 minutes while interacting with an app on a tablet. Hemispheric says its AI model will then help clinicians decode the signals to make diagnoses, select the most effective intervention by making predictions about treatment, and monitor progress.

“The future that we envision is one where this is akin to a blood test,” Lalazar says. “The device is going to be very, very cheap; it will be able to be sold and distributed throughout mental health clinics, hospitals, and even psychologists’ offices.”

AI-assisted diagnostic tools for conditions like lung cancer are already in clinical use and speeding up access to treatment across Europe. Meanwhile, AI giants including OpenAI and Anthropic are expanding into health care, intensifying competition for the raft of startups in the space.

Hemispheric has raised early-stage funding from investors including American and Israeli venture capital firms and individual investors, among them early Uber-backer Howard Morgan. They will use the money to advance partnerships with governments, healthcare organizations, and pharmaceutical firms, hire more in the US, and work towards regulatory approval. They also plan to measure more brain data from millions of people in an effort to improve their model

The pair are also developing their own brain scanners to obtain information that the company believes can provide more useful data for its models than traditional EEGs. “These devices were never built for machine learning and definitely not deep learning,” Littwin says.

Something to look forward to: League of Legends has joined the popular trend of long-running live service games that introduce classic modes to emulate their early years. However, instead of reverting to a specific moment in the game’s evolution, Riot Games will combine elements of how LoL felt between 2009 and 2013.

The League of Legends Classic game client is now available to download. The new mode, set to open at the end of this month, will focus on characters and gameplay elements from the game’s first three seasons.

Rather than simply relaunching League of Legends 1.0, Riot decided to start with Season 3 as a base and add features from the prior two seasons with a few modern touches. For example, while it revives the early Summoner’s Rift map, the company decided to improve readability by enhancing lighting, shadows, and textures.

Classic mode will include Summoner’s Rift, the original 40 playable characters from League of Legends’ 2009 debut and an additional 20 characters hand-selected from the game’s first four years. Items, runes, builds, masteries, skins, and other mechanics from this era will also return, including Atmogs, the Metagolem setup, and Zz’Rot Portal.

Playing rounds in Classic mode will unlock characters through a new progression system called Classic Level, which runs alongside progression in the main game – though players can immediately select characters they have already purchased. Another new progression system, called Summoner’s Journey, begins at Classic Level 10 and allows players to earn items without the pressure of ranked play.

League of Legends Classic will debut with a player-versus-player draft queue, co-op, and custom game modes. Runes and Masteries must be pre-selected before starting a Classic match. Both are unlocked only via gameplay and cannot be upgraded, as they are all set to the old Tier 3 baseline.

Although Classic mode aims to remind players how League of Legends felt before more than a decade of additions and balance changes, it will not remain static. Riot plans to release patches and balance updates according to player feedback, favoring users with higher Classic levels.

Meanwhile, fans are still awaiting League Next, which Riot teased late last year. The update, expected in 2027, will be the largest ever, bringing significant changes to the graphics, gameplay, and technology underpinning League of Legends. With a new onboarding experience, optional WASD controls, and changes to Summoner’s Rift, Riot aims to reverse declining player numbers by making the influential MOBA more accessible to newcomers.

Drake Anthony spends a lot of time in the workshop, continually pushing the boundaries of what you can produce with basic materials. His recent video on the styro pyro 2 channel attempts to replicate a phenomenon discovered over three centuries ago. The ultimate result is a glass flask that emits visible light when shaken, eliminating the need for batteries, wiring, or other external power sources.

The story begins in 1675, with French astronomer Jean Picard out on the streets of Paris on a dark, clear night. He’s carrying a mercury barometer, and as he moves, a small glow begins to form in the glass tube above it. This light appears only when he stirs the mercury and exposes some new glass for it to play on. The whole thing is a strange sight, which he describes to several of his scientific friends at the time and ends up giving the nickname ‘barometric light’.

YIERBLUE Rechargeable Spotlight Flashlight with 2000000 High Lumen LED, IP67 Waterproof Long Running Spot…

• EXTRA BRIGHT HEAVY DUTY FLASHLIGHT: Its brightness is several times that of the small model.This extra bright LED spotlight casts a super bright…
• WATERPROOF & FLOATING, IMPACT RESISTANT: This spot lights outdoor handheld is IP67 waterproof and submersible up to 3ft, and even float back up to the…
• LONG TIME LIGHTING & POWER PORT: Built-in rechargeable batteries, YIERBLUE’s proprietary battery technology give you up to 9 hours lighting on lower…

It takes the scientists some time to figure out what’s going on, but friction turns out to be the key to everything. When mercury flows smoothly over a clean glass surface, it leaves a trail of electric charge that can travel through the air inside the container. When that electric charge strikes the gas molecules within, they begin to emit light, which is similar to how neon signs work, except that the energy comes from friction and contact between the two materials.

Anthony sets out on a journey to recreate this entire effect from scratch. First, he takes a round glass flask and uses a propane-oxygen flame to create a slender glass stem for it. Glassblowing was a completely new ability for him, and things became even more difficult because he only has limited depth vision in one eye. Nonetheless, he completed it successfully. After attaching the stem, he connects the flask to a vacuum pump and gradually drops the pressure to approximately 5 millitors, leaving only a whisper of the original air inside. There was still some residual gas and mercury vapour within, which he later learned was necessary for the light to work.

The next step is to transfer the mercury into the flask. He puts a little more in and then screws the top back on. When he initially shakes the flask in full darkness, you may see a faint blue-white glimmer. The reason it’s so faint is that the mercury he’s using isn’t pure, which limits the amount of charge that accumulates when the mercury passes over the glass. Even so, the light illuminates whenever he shakes the device, which is a pretty good indicator given that the technology is nearly 350 years old.

Anthony wanted to brighten up his own light display, so he started experimenting with adding little amounts of noble gases to the flask after he had the vacuum running. It was a feeling that starting with neon at around 100 torr would be a good place to start, and boy was it correct, as shaking the flask generated a really dazzling light that appeared almost electric and could be seen in regular room lighting. The neon seemed to make the entire process more easier from beginning to end, getting the discharge started and maintaining it flowing.

The other experiments he carried out in identical flasks were quite interesting. Adding copper pellets appeared to work just as well as glass in producing a decent light through friction, but then he tossed in a few of chunks of Teflon, which resulted in a few little sparks. The Galinstan experiment was similarly a failure, as he attempted to use a liquid metal alloy instead of mercury, but it stuck to the flask’s glass walls, making it nearly difficult to generate a sufficient charge. As expected, employing a straight tube was far less effective than using a curved or bent one since the gas would simply break contact and re-form whenever it encountered an edge, giving him even more possibilities to build a charge.

Anthony decided to add a Tesla coil to the mix just for fun, to give the gas within the flask a little extra kick from the outside. And let me tell you, the results were just plain cool, as the coil supplied a little more juice to the previously charged region, resulting in some amazing plasma displays with nice pinching effects and distinct color zones.

If you’re looking to add more 4K Blu-rays to your collection, then Disney’s Summer Days, Movie Nights 4K Blu-ray sale has arrived at the perfect time, where it’s 2 for £30 on discs including titles from Disney, Marvel, Star Wars and 20th Century Studios. The same 2 for £30 sale is also live at HMV.

I regularly test 4K Blu-ray as part of our monthly Blu-ray Bounty feature, where we test the best new 4K releases. I’ve picked six discs from the sale that all featured as part of the Blu-ray Bounty, so I know just how good they look and sound.

• Microsoft’s July 2026 Patch Tuesday fixed a record 622 vulnerabilities, including 58 critical, two exploited in the wild, and one publicly disclosed, plus 428 Chromium bugs
• Actively abused flaws include CVE‑2026‑56155 (AD FS privilege escalation) and CVE‑2026‑56164 (SharePoint privilege escalation), alongside notable issues in BitLocker and Copilot
• Surge in fixes is linked to Microsoft’s use of Anthropic’s Mythos AI, with patch volumes rising sharply since its adoption

Microsoft has released its July 2026 Patch Tuesday download, marking another record-breaking update, addressing hundreds of flaws across the ecosystem.

The release, which is currently rolling out to Microsoft users, fixes a staggering 622 vulnerabilities, including 58 critical-severity ones, two that were observed as being abused in the wild, and one which has already been publicly disclosed.

Anthony Albanese has told the AI industry that Australian books, music, and journalism are not free training data, and that any large data centre built in the country will have to put more electricity into the grid than it draws out. Neither of those things is law yet.

The prime minister used a speech at the University of Sydney on Wednesday to announce an Office of AI inside his own department, effective immediately, plus Australian Standards covering energy, water, copyright, and siting.

It lands two days after Anthropic and others were reported to be weighing tens of billions in data centre investment against a copyright carve-out Canberra had already ruled out.

The energy obligation is the sharpest thing in the speech. Operators of the next generation of large data centres would be required to underwrite new power supply, pay their full share of grid connection so that no costs land on homes or businesses, and put at least as much energy into the grid as they take out of it.

“To be net-generators, not net-users,” Albanese said. That means funding new renewable generation and firming rather than joining a queue for someone else’s electrons, a heavier ask than anything hyperscalers face in Europe or the US, where grids are already buckling under connection requests.

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Water got similar treatment. Operators would have to minimise water use, maximise energy efficiency, and pay for any additional water infrastructure they need, on a continent Albanese called both the sunniest and the driest on earth.

Copyright got the rhetoric. “Let me make this crystal clear: not everything produced in Australia is up for grabs,” he said. “Not at all.” Australian writers, musicians, artists, and journalists “must retain ownership and control of their work”, and no company should train on it without the artist’s control of its price and value. “Anything less, is theft.”

What the speech did not contain was a mechanism. The policy has been read as obliging AI firms to reach agreements with local artists and media before using their content, but Albanese never said how that control would be enforced, and the attorney-general’s consultation on copyright is still open.

The distance between announced and legislated is the story here. Nothing unveiled on Wednesday binds anyone: the Office of AI is an executive creation, the standards go to National Cabinet next month, and legislation is only targeted for introduction early next year.

Albanese was candid that he does not want an exhaustive rulebook. “It is not our goal to try and legislate for every possible eventuality or risk,” he said. That is a lighter touch than the language around it implies, and closer to the ground Brussels has been retreating to than to the AI Act as drafted.

His claim that Australia “will be the first country in the world to bring these issues into a single, national framework” is doing work it cannot carry. The EU adopted the AI Act in 2024 and built an AI Office to run it, as legal scholars noted within hours.

Reaction divided on schedule. Greenpeace Australia’s Joe Rafalowicz called the facilities “water-guzzling energy vampires”, accusing the government of rolling out the red carpet while leaving them unregulated until at least 2027. Opposition Leader Angus Taylor said the office would just create more bureaucracy.

New York, hours before Albanese spoke, halted large data centre builds for a year, the pause Australia has now declined to take. Washington is still arguing over who pays when data centres raise power bills, the question Albanese thinks he has answered in advance.

Anthropic, which told Treasurer Jim Chalmers that its A$21.6bn Australian investment depended on copyright certainty, said it respected the process and would meet the terms the government sets. That is a company waiting for the fine print.

APRA AMCOS chief executive Dean Ormston welcomed the certainty but said the Office of AI “must seriously interrogate the numbers AI platforms are putting on the table”. The numbers are not on the table yet. Neither is the bill.

A new official filing marks the conclusion of another regulatory hurdle that has been preventing Apple Intelligence from launching in Apple’s crucial Chinese market.

In mid-June 2026, Apple’s Chinese partner Alibaba unveiled its latest AI model and, most significantly, revealed that it was now compatible with Apple Intelligence. Now according to Reuters, China’s cyberspace regulator has allowed Apple Intelligence to be registered for use on iPhones in the region.

China has complex requirements for regulatory approval, and sufficiently so that US firms typically need to involve a local partner. It’s not certain yet whether this latest listing is the final step toward Apple Intelligence launching in China, but it appears to be, and it marks more than two years of effort from Apple.

Apple has not commented on the regulator’s listing, and Alibaba has only confirmed certain details that were already presumed. Specifically, the company stated that its new Qwen AI model will be integrated into Apple Intelligence, when that becomes available for iPhone, iPad, Mac, and Apple Vision Pro users in China.

The report also cites an unspecified source saying that Apple Intelligence will also incorporate technologies from Baidu. This has also been presumed, as despite seemingly choosing against Baidu as a full partner in 2024, Apple reportedly signed with the company for specific functions such as Visual Intelligence.

Apple Intelligence has effectively required partnerships with local companies even though Apple has been able to release iOS in the country without the same level of scrutiny. This is because China has restrictive laws about generative AI software, large language models, and data privacy.

It’s sufficiently difficult to obtain Chinese government approval for any such services, that OpenAI is banned in the country. Apple needed to partner with firms such as Alibaba and Baidu which already had certain approvals. Use of these Chinese firms, though, also raises issues over how Apple Intelligence will be censored in the country.

Apple appears to have worked through all the technical issues, though, and has been waiting only for regulatory approval. That’s because in March 2026, Apple Intelligence was briefly enabled in China, apparently by mistake.