Apple’s keynote at this year’s Worldwide Developers Conference was a big one. After months of delays, Apple reintroduced us to its AI-upgraded Siri, which will go beyond what the existing voice assistant can do by offering more personalized help. We also got a look at many other updates coming across the operating systems powering the iPhone, iPad, and Mac.

You can catch up with all the biggest news in the roundup below.

Image: Apple

 

Siri is getting a bunch of AI features powered by the latest version of Apple Intelligence, which Apple worked with Google to build. Apple is calling the upgraded voice assistant Siri AI, offering the ability to have back-and-forth conversations “with detailed and engaging answers,” Apple AI head Mike Rockwell said during the keynote.

In addition to offering image editing and writing tools, Siri AI also uses onscreen awareness to pull context from what’s on your screen to answer questions. Apple is also launching a standalone Siri AI app, where you can access your conversation history. Siri AI will roll out across Apple’s devices, including the iPhone, iPad, Apple Watch, and Apple Vision Pro.

Image: Apple

 

Apple is adding the ability to access its new Siri AI by swiping down from the Dynamic Island, where its responses will also appear. iOS 27 comes with a “Siri Mode” in its Camera app, too, which can recognize objects in your viewfinder and tell you more about them.

There’s also a new opacity slider for Apple’s polarizing Liquid Glass design. Apple says iOS 27 will be supported by the iPhone 11 and up.

Image: Apple

 

Apple’s bringing an AI-heavy update to macOS 27 Golden Gate, too. It’s building Siri into its Spotlight menu, allowing you to ask questions directly in the search bar and select files to ask about.

Similar to iOS 27, the latest version of macOS will allow you to adjust its Liquid Glass design with a new slider. Apple is making the window corner radius tighter in macOS 27 and is bringing back colorful sidebar icons.

Image: Apple

 

During the keynote, Apple showed off a bunch of handy new features for Safari, including a new tab management feature that will use AI to organize all your open webpages by topic.

There’s a new “Notify Me” feature that will allow you to ask Safari to alert you to changes on a webpage — such as if a product you’ve been eyeing comes back in stock — and a way for Safari to automatically update your compromised passwords for you. You can also ask Safari to create extensions just by describing them with natural language.

Image: Apple

 

A new feature in Messages will suggest photos to share based on your conversation, while the Phone app can present you with relevant information, like a flight confirmation code, when the call starts. Apple also says Mail will offer “more capable suggestions” and allow you to add events to your Calendar just by describing them. Some other upgrades include the ability to create your own shortcuts with natural language and use Apple’s Image Playground to make AI edits on your photos, such as changing their compositions and enhanced “distraction” removal.

Image: Apple

 

Even the Home app is getting a notable update, enabling it to generate descriptions of camera clips and stitch together footage from different cameras with AI. Apple is adding the ability to view recorded clips with 4K resolution on supported cameras and a way to search through clips with natural language.

Image: Apple

 

Apple is making a big update to its parental control system. Now, parents who enroll their kid in a child account can choose which apps and content they want them to have access to, as well as require their kid to ask for permission before opening a new website. The company also redesigned its Screen Time menu, which it says will make it easier for parents to manage the time their kids spend on their phones.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 9 June 2026 at 7:49 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

7 Days is a weekly roundup of picks of what's been happening in the world of technology - written with a dash of humor, a hint of exasperation, and an endless supply of (black) coffee.

This week's highlights are packed with hardware announcements from Computex 2026, Microsoft's BUILD developer conference, and lawsuits against OpenAI and Ring. Let's get started.

You can check out the recent issues of the 7 Days weekly roundup.

"Enough is enough"

From "bribing" users to forcing Edge at startup, Microsoft has turned over every stone to make people use its web browser. Browser Choice Alliance (which includes Chrome, Opera, and Vivaldi) is now after the Redmond giant once more and has penned an open letter to highlight dissatisfaction with its practices.

The letter to CEO Satya Nadella emphasizes that "enough is enough" and Microsoft should respect browser choices on Windows. BCA laid down a list of actions to level the playing field and believes that browsers should compete on merit.

In other browser news, a fresh update to Firefox fixed a massive VPN button and a bug that disrupted page layout. The Ladybird Browser Project announced that it will no longer accept public pull requests and limit changes to those made by its maintainers as it moves towards its first alpha release.

Computex 2026

In one of the week's hottest stories, AMD is trying to make DDR5 RAM even faster on Ryzen systems with its new EXPO ULL (Ultra Low Latency). The feature will enable support for even lower CAS Latency DDR5, bringing significant performance gains over normal EXPO.

AMD released new octa-core 3D V-cache CPUs in the form of the new Ryzen 7 5800X3D and Ryzen 7 7700X3D for AM4 and AM5, respectively. The company also brought the 9070 GRE to the USA and other countries. Compared to the NVIDIA RTX 5060 Ti, AMD claims the 9070 GRE offers 22% faster performance and 26% better value.

The expo also set the stage for Intel's Crescent Island GPU for data center AI workloads and inference. It can pack up to 480GB LPDDR5X VRAM, and the cooling department is handled by an air cooler with a 350W TDP. The silicon giant's AI-focused data center strategy also includes Clearwater Forest, which comprises new Xeon 6+ up to 288 E-cores.

Intel unveiled its OpenVINO Physical AI framework to enable scalable, lower-cost edge robotics with improved efficiency. The company said it has found a way to fill the "missing link" that made it difficult to deploy physical AI at scale across the edge.

The next trillion-dollar company?

Hitting the trillion-dollar mark is the new fashion in the tech industry. NVIDIA has already done so by a wide margin, and now its CEO, Jensen Huang, says Marvell will be the next trillion-dollar company because of its key role in the AI revolution. Marvell is an American semiconductor company founded in 1995; its stock price jumped by 22% this week after Huang's comment.

Backlash for employee tracking

Meta previously rolled out the Model Capability Initiative (MCI) tracking system to track keystrokes and screen content within specific apps to train its AI models. It wasn't long before employees raised privacy concerns and complained that the software consumed excessive data and battery life. The company is scaling back by introducing controls that allow employees to pause the tracking for up to 30 minutes and request exemptions.

In other news, Meta patched an exploit that reportedly allowed attackers to take over accounts by tricking the Meta AI support assistant. The attackers managed to take control of a dormant Obama White House Instagram account that hadn't posted since 2017.

This week in software news

Catch up on some of the latest software news updates that arrived throughout the week:

Hard battle for AV2: The new AV2 video codec can reduce average bitrate by 30-34% compared to AV1, delivering substantial improvements at equivalent visual quality. However, according to VideoLAN's Jean-Baptiste Kempf, the increased efficiency would come at the cost of dramatically higher computational complexity, making the job for existing CPUs much more difficult.

Ring faces lawsuit: The smart doorbell brand is in a legal battle over its "Familiar Faces" feature. The lawsuit claims that the feature collects facial data from millions of other Americans (think delivery drivers, neighbors, mail carriers, and pedestrians) who pass by a camera without their consent.

Tuta joins Euro-Office: The German email provider has joined the European effort to break free from American-made software. The open-source office alternative to Microsoft 365 and Google Docs is due for a release this month.

Vim Classic: The new Vim fork has successfully launched its first stable version, 8.3.0, which is completely free of LLM-generated code. It's based on Vim 8.2.0148 because the developers wanted to dodge the heavy maintenance footprint of the newer Vim9 Script engine. However, some modern plugins may not work because of it.

Fending off data scraping: Strava is done with data scraping for AI. The fitness platform put public data behind a paywall and a login screen to stop AI companies from hammering its servers.

DLSS 4.5 Ray Reconstruction: It brings an updated version of NVIDIA's transformer model that promises to deliver even better image quality than before in ray-traced and path-traced games.

Fastfetch 2.64 released: The popular command-line system info tool has been updated with experimental scripting support, streamlined compilation options, a smarter logo renderer, and Codec module support.

Shotcut 26.6 Beta: The latest beta of the free video editor adds OpenFX and VST2 plugin support, UI tweaks, HDR preview upgrades, and a range of bug fixes.

KDE Plasma 6.8: A few changes slated for the next release include a warning on the lock screen that shows when the "Slow Keys" accessibility feature is active, so you don't type incorrect passwords.

Improved Linux gaming: Canonical has promoted the Arm64 Steam Snap to the stable channel, using FEX emulation to bring PC gaming to Snapdragon and NVIDIA hardware.

This week in hardware news

Catch up on some of the latest hardware news updates that arrived throughout the week alongside Computex 2026:

Surface Laptop Ultra: NVIDIA is back in the CPU chip game with the new RTX Spark, which debuted on the Surface Laptop Ultra. Promised to be the most powerful Surface to date, it packs a 15-inch mini-LED display, 20-core NVIDIA Grace CPU, NVIDIA Blackwell RTX graphics, and 128GB of unified memory. More information is yet to arrive, alongside RTX Spark-powered computers from other brands.

Surface RTX Spark Dev Box: It's a high-performance desktop workstation designed specifically for AI development. With 128GB of unified memory, Surface RTX Spark Dev Box can deliver 1 petaflop of AI compute and run models with up to 120 billion parameters locally, significantly reducing reliance on cloud GPU instances.

ROG XBOX Ally X20: ASUS celebrated the 20th anniversary of the ROG brand with a new handheld device, featuring a completely redesigned mainboard, an upgraded display, better joysticks, and even a translucent chassis.

Majorana 2: Microsoft unveiled its latest quantum chip, claiming it's 1000x more reliable than last year's Majorana 1. The new development has accelerated the timeline to achieve practical quantum computing from 2035 to 2029.

Sony gaming accessories: Sony finally attached a release date to a few of its gaming accessories, including FlexStrike Wireless Fight Stick and 27” Gaming Monitor with DualSense Charging Hook. Both are releasing in August this year.

Leaky Surface: Some leaked promo material of the upcoming Surface Pro hints at what the device could offer. The tablet will be powered by a 12-core Snapdragon X2 Elite processor paired with a new Neural Processing Unit operating at up to 80 TOPS.

This week in Google News

Catch up on some of the latest Google news updates that arrived throughout the week:

How to train your AI? An AI model is as good as the data it's trained on. Google is reportedly reaching out to Android developers to get their private codebases to train models and improve development tools, while giving them some cash in return.

Design your band: Google published the physical design blueprints for the Fitbit Air, opening doors for anyone with the required skills to customize the screenless tracker and build accessories. The search giant said that certified accessories can also get the official "Made for Google" badge.

It's optional now: The UK's competition watchdog has enabled publishers to opt out of Google's generative AI search features, such as AI Overviews and AI Mode. It's also forcing Google to attribute content properly, using clear links to sources, in all AI-generated answers.

This week in Apple News

Catch up on some of the latest Apple news updates that arrived throughout the week:

What to expect? WWDC 2026 is just around the corner. Alongside refinements to Liquid Glass, the iOS 27 update could be centered on the long-awaited upgrade to Siri. iPadOS 27 is also expected to get a major AI boost this year, among various expected updates.

Apple smart glasses: If you're hoping to see Apple's new smart glasses at WWDC, you might have to wait. Those plans are reportedly delayed until late 2027. It's said Apple is cautious about launching hardware that relies on underdeveloped visual intelligence systems.

Keep in check: A recent court ruling that lifted an injunction on the Texas Age Assurance Law (SB 2420), Apple had to enforce strict age-verification and parental-consent rules for new Apple accounts created in Texas.

This week in AI news

Catch up on the latest artificial intelligence news updates that arrived throughout the week:

OpenAI faces lawsuit: Florida Republican Attorney General James Uthmeier took OpenAI to court, claiming that the AI lab released an unsafe product and misled the public about safety risks associated with ChatGPT.

France gets cash for AI: SoftBank plans to invest up to €75 billion ($87 billion) in France's AI infrastructure, making it one of the largest investments in Europe's AI sector. Up to €45 billion will be used to build two AI data centers in Le Bosquel and Dunkirk.

AI chaos slows down: After weeks of AI-generated noise and late submissions, Linus Torvalds said that things have quietened down for Linux 7.1 RC6, which is smaller than RC5, and we could be on track for a normal release cycle.

ChatGPT memory upgrade: The AI chatbot got a major architectural upgrade to its memory system, significantly improving its long-term context retention. It improved factual recall from 67.9% in 2025 to 82.8% in 2026, and accuracy over time improved from 52.2% to 75.1%.

Lockdown Mode expanded: OpenAI is rolling out ChatGPT's Lockdown Mode to all personal and self-serve ChatGPT Business accounts. The feature improves security by disabling live web browsing, Deep Research, Agent Mode, and more.

Codex on ChatGPT: The full Codex experience is now available in the ChatGPT app to support knowledge workers, who represent about 20% of Codex users. A new Codex feature called Sites enables users to create and share interactive hosted websites and apps.

This week in Microsoft News

You can download the Surface Laptop Ultra wallpapers in high resolution. Windows 11 is dominating the gaming market, and data from Steam showed nearly 70% of all participants were using a Windows 11 PC. A third-party tool called OfflineInsiderEnroll is for insiders who want to unlock Windows 11 features with a Microsoft account.

You can check out Taras's freshly baked Microsoft Weekly roundup to catch up on all the interesting stories this week.

This week in science news

Catch up on some of the latest science and out-of-this-world updates that arrived throughout the week:

How will the Sun end? Astronomers found that an ancient white dwarf star is still consuming rocky planetary debris after cooling for three billion years, proving systems remain active long after their host star dies.

Eye of Sauron: Scientists solved a cosmic mystery. A distant black hole is pointing its intense jet straight at Earth, creating an optical illusion that makes the blindingly bright stream look surprisingly low.

This week in gaming news

Catch up on some of the latest gaming and virtual world updates that arrived throughout the week:

Summer Game Fest: The event went live on June 5 from the Dolby Theatre in Los Angeles, California. The two-hour showcase was hosted by video game journalist Geoff Keighley and introduced games across multiple genres.

New racing game: Some former Forza Horizon team members created a brand-new racing game called Clutch. The game offers a story-driven campaign, multiplayer action, and aims to be a "benchmark in car customization."

FSR hits a new milestone: AMD announced that the latest generation of its FSR technology now officially supports 300 games, a considerable jump from just 30 at launch.

What else in gaming?

The latest issue of Pulasthi's Weekend PC Game Deals curates several exciting games on sale this week. The final leg of the Epic Games Store's mystery giveaways brings copies of Rogue Waters and Songs of Conquest to claim for PC gamers.

Xbox Free Play Days welcomed ten new games this weekend from a single publisher, including Little Rocket Lab, Spirittea, Descenders Next, and Let's Build a Zoo. Meanwhile, Prime members can grab Mafia III, Tomb Raider remasters, and 13 more games in June to keep.

That said, here are some more stories from the gaming world:

• Crystal Dynamics pushes Tomb Raider remake to 2027
• A roguelike and a 4X strategy game are free to claim on the Epic Games Store
• Looks like EA's Star Wars Zero Company will be out this August
• God of War Laufey announced, introducing Kratos' wife as new protagonist

From the review corner

If you have been thinking about capturing the night sky, the DWARF mini is the world's smallest smart telescope for night-and-day sky captures, which Steven reviewed this week. For an amateur astronomer spending $399, the telescope offers premium build quality, automated tracking, and a low learning curve. However, the tracking may not always work straight away, and the connection can be finicky.

GEEKOM Air12 2026 Edition

It's a small mini PC from GEEKOM fitted with an Intel Tiger Lake Pentium Gold 7505, up to 16GB of RAM, and up to a 512GB SSD. GEEKOM Air12 2026 Edition comes with a lightweight chassis, a 15W TDP, supports up to three 4K 60Hz displays, and Type-C on the front. However, points are deducted for its single-rank (2666 MHz) DDR4 RAM, and the front USB port is data-only.

AMD RX 9070 GRE

Steven and Sayan joined their forces to put the new AMD RX 9070 GRE against the RX 9070, RX 9070 XT, NVIDIA 5070 FE, and some other cards in gaming as well as productivity. AMD has pitched it against the Nvidia RTX 5060 Ti 16GB, which is typically sold at around this price range. The GRE performed quite well against both the 7800 XT and the 4070. It offers balanced performance, sufficient VRAM, and runs cool. However, the ray tracing might feel mediocre.

Cuktech 10 Ultra

How about a wall charger with a big screen that shows the stats in real time? Taras reviewed the Cuktech 10 Ultra charger, which features four ports, a large display, and up to 110W of power output. Its 1.57-inch display with 700 nits max brightness is the main highlight, capable of showing total output power, current temperature, power distribution across ports, and more.

007 First Light

Pulasthi's review of 007 First Light said the game delivers an immersive, globe-trotting origin story for James Bond, packed inside a tightly choreographed action game. It features over-the-top action sequences, Bond's right amount of overconfidence, and satisfying gunplay. On the other hand, stealth can be too predictable, enemy AI is not very bright, and the missing FOV slider is a pain.

So, these were some of the biggest tech news and other updates from this week. There will be more issues of our 7 Days series in the coming weeks and months, so stay tuned. You can also support Neowin by registering for a free member account or subscribing to extra member benefits, along with an ad-free tier option.

Have a great weekend!

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 8 June 2026 at 11:52 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

A few months ago, Undead Labs broke its silence about the third entry in the State of Decay series since its announcement in 2020. Today, the studio had a brand-new trailer to reveal at the Xbox Games Showcase, finally giving players a look at the Unreal Engine 5-powered title. A broad 2027 release window is now attached to State of Decay 3 too.

The studio confirmed that every scene seen in the gameplay trailer (except for the studio logo) released today is from the game with no cinematic footage.

"The quality bar of the moment-to-moment experience is higher than anything we’ve achieved in Undead Labs’ history," says the studio. "State of Decay 3 is being built in Unreal Engine 5, and we’re creating better quality versions of what makes our game the survival sandbox it is, and the combat experience is really, really important to us, and I think you can see it coming through in this trailer."

One new aspect of this upcoming entry is the Plague Nests. These are supposed to be dynamic, dungeon‑like enemy hubs with varied “personas” that change tactics each time one appears on a map. These areas will have the best loot in the game for players brave enough to tackle them.

The combat system is being overhauled too, letting players do quick and power attacks depending on the situation. This isn't just on the player's side though, as zombies themselves now have multiple new types of behaviors depending on how fresh or decayed they are.

"All of these play very differently, especially as you start getting outnumbered or when you’re making noise and running around and trying to save your rear end or the rear end of your friend next to you," adds the developer.

Of course, settlement building and community management are massive parts of this third entry too. These are being further expanded with the cooperative focus, giving players much more freedom for playing together than in State of Decay 2. Players will even be able to split up and build settlements in different areas of the same map with different goals.

Undead Labs confirmed that playtests are already ongoing for the title, and interested players can sign up to participate by heading over here once more slots open up ahead of the full release.

State of Decay 3 is coming out on PC, Xbox Series X|S, PlayStation 5, and Xbox Game Pass sometime in 2027.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 8 June 2026 at 8:10 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

Earlier this week, we shared the first part of our review of AMD's new RX 9070 GRE. It was about the gaming performance of the GPU, and we gave it an 8 out of 10. As a follow-up, similar to how we did with the 9070 XT and non-XT, we are doing a dedicated productivity review for the RX 9070 GRE as well, where we compare it against the 9070 XT, 9070, 7800 XT, as well as Nvidia's 5070 and 4070. This will include AI, rendering, compute, and more benchmarks.

AI performance, especially, is a very important metric in today's world, and AMD also promised big improvements thanks to its underlying architectural improvements.

We will be pitching it against the data we already have for the RX 9070, and RX 9070 XT, but also the Nvidia 5070 FE, MSI GeForce RTX 4070 VENTUS 2X 12G, and Gigabyte Radeon RX 7800 XT GAMING OC 16G as they are in a similar price class, but also because we do not have a comparable 5060 Ti card lying around here that we can compare it against.

Before we get underway, this is a collaboration between Sayan Sen and Steven Parker, who lent me his test bed. Also, there was no editorial input from AMD.

First up, the specs of the RX 9070, 9070 XT, and 9070 GRE, which were given to us by AMD:

As you can see from the specs above, it is less than the standard RX 9070 in every way that counts, except for slightly higher Boost and Game clock speed.

Design

Moving on, the RX 9070 GRE we were given is an XFX Swift triple-fan, dual-slot design with two 8-pin connectors. At 30cm (self-measured), it will fit in most systems easily. There is no RGB either.

The AMD Radeon RX 9070 GRE by XFX from all angles.

Test system

Our test system consists of the following:

• Lian Li O11 Dynamic Mini V2 Flow (Amazon|Newegg)
• ASUS Z890 ProArt Creator WiFi (Amazon|Newegg)
• Intel Core Ultra 7 270K Plus (Amazon|Newegg)

• Thermal Grizzly KryoSheet - 44x37 (Amazon|Newegg)

• 2x 16GB G.Skill Trident Z5 RGB (7200 MT/s in XMP) (Amazon|Newegg)
• Sabrent Rocket4 Plus 2TB SSD (Amazon)
• Windows 11 25H2 (Build 26200.8246)

AMD shared a press driver based on the recently released Adrenaline 26.5.2 that we were required to use. We now move on to our benchmarks. First up, we have Geekbench AI running on ONNX.

For some reason, the 9070 GRE does exceptionally well here in both half-precision (FP16) and single-precision (FP32). It manages to beat the RTX 5070 and RX 9070 non-XT, and is only behind the 9070 XT. Since Geekbench runs in short bursts instead of continuously hammering the graphics card, it seems the GRE's faster boost clocks are helping here.

Next up, we move to the UL Procyon AI test suite, starting with the image generation benchmark.

We chose the Stable Diffusion XL FP16 test since it is the most intense workload available on Procyon. The Nvidia cards do very well here, as even the 4070 out-muscles AMD's best fairy easily.

The positive thing about the GRE is that it gets quite close to the 9070 non-XT in this test; this indicates that the VRAM does not play a very big role here, as SD XL relies on float16 (FP16). So this is something to keep in mind again. If you wish to work with float32 AI workloads, graphics cards with larger than 12 GB buffers would likely emerge as victors. Regardless, the gains are still massive on AMD's 9000 series compared to the 7000 series.

Following image generation, we move to the text generation benchmark.

This is one test where the 9070 GRE struggled, quite a lot. It seems that the 12 GB VRAM and lower memory bandwidth of the new Radeon 9070 GRE are hurting it quite a bit; the split is massive, especially in a test like Llama2, which packs 13 billion parameters. As such, in all the tests, the 9070 GRE is the slowest of the lot.

Next, we tried Blender, and here the AMD GPUs were beaten by Nvidia. Rendering is something the Green team has always had a lead over the Red side, and it has not changed so far. On the positive side, though, the 9070 GRE shows significantly better results than the 7800 XT, which means AMD is on the right path. Catching up to Nvidia, though, will require a lot more effort. And we hope HIP and ROCm can keep improving.

Wrapping up AI testing, we measured OpenCL throughput in the Geekbench compute benchmark.

The RX 9070 GRE alongside the 9070 did not fare well here at all, even falling behind the 7800 XT. Interestingly, even the RTX 5070 could not beat the 4070 on OpenCL, so perhaps this suggests that OpenCL optimization may not have been a priority for either AMD or Nvidia in the modern era.

Conclusion

We reached the end of our productivity performance review of the 9070 GRE, and we have to say it's a mixed bag. Unlike the 9070 and 9070 XT, the GRE excels in some areas while losing ground fairly easily in others. Similar to how it happened in gaming, any time the card's memory subsystem gets hammered, it tends to fall behind the others. This was the case with text generation, wherein we saw the VRAM sometimes hit its maximum available 12 GB of usage with larger model sizes.

So what do we make of the RX 9070 as a productivity hardware? It can certainly be used, but you have to know it has its limitations.

For those looking for a GPU that can deal with more, AMD recently unveiled the Radeon AI PRO R9700, which is essentially a 32 GB refresh of the 9070 XT with some additional workstation-based optimizations. On a similar note, the new Ryzen AI Halo platform is something you can consider if you want to set up a local AI processing station.

Considering everything, we rate AMD's Radeon RX 9070 GRE a 7.5 out of 10 for its productivity performance. Price is less of a factor for those looking at productivity cases compared to those considering the GPU for gaming, and as such, we felt it did quite decently on many occasions and can be handy if you need a 12 GB GPU and, for some reason, don't want to get Nvidia.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 8 June 2026 at 8:05 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

Last year, id Software released Doom: The Dark Ages as a prequel to its reboot series, offering a look at the legendary Slayer character during an ancient war between Hell and Heaven. Today, at the Xbox Games Showcase, id Software returned with a new announcement that unveiled Doom: The Dark Ages Revelations as its first story expansion.

The expansion will see the Slayer being betrayed and cast into a "merciless purgatory only escapable by confronting haunting truths." Somehow, there is one ally in all this chaos seemingly attempting to help our protagonist escape his fate. Players will be ascending this prison and fighting plenty of new demons and abominations as they work to free the Slayer's followers.

One new aspect of the expansion will be the Chain Spear. This new weapon is described by the studio as a "uniquely satisfying combat system that rewards mastery with a potent combination of power and mobility." Check out the trailer below to see the new weapon in action, where the Slayer is seen using it to even pin enemies to walls.

id Software is also preparing a free update to all owners of Doom: The Dark Ages that will land alongside the new paid expansion. Dubbed the Ripatorium 3.0 update, this will add more customization options, an improved pass code generation system, and preset options for players diving into the game's arena challenge mode. Expansion buyers will also receive additional maps, demons, and new weapons to use in this mode in addition to the story levels.

Doom: The Dark Ages Revelations expansion is out on July 7 across PC, Xbox Series X|S, and PlayStation 5 with a $19.99 price tag. Those who own the Premium Edition or the Collector's Bundle will receive the expansion for no extra cost as well.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 8 June 2026 at 7:59 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

This week's news recap is here, with Microsoft announcing the new Surface Laptop Ultra, fresh chips from NVIDIA for Windows on ARM, a no-build week, fixes for Windows 11's context menus, gaming news, reviews, and more.

Quick links:

1. Windows 10 and 11
2. Windows Insider Program
3. Updates are available
4. Reviews are in
5. Gaming news

Windows 11 and Windows 10

Here, we talk about everything happening around Microsoft's latest operating system in the Stable channel and preview builds: new features, removed features, controversies, bugs, interesting findings, and more. And, of course, you may find a word or two about older versions.

At Computex 2026, together with NVIDIA, Microsoft announced the Surface Laptop Ultra, its most powerful laptop to date, powered by NVIDIA's RTX Spark processor. Details about this computer are currently scarce, as Microsoft has only revealed certain parts of its specs. So far, we know that the computer has a 15-inch mini-LED display, a rich set of ports, a powerful processor, and all-day battery life. It also comes with a new wallpaper, which you can already download here in full resolution.

The Surface Laptop Studio is not the only NVIDIA-powered Surface, which Microsoft unveiled this week. At Build 2026, the company also debuted the Surface RTX Spark Dev Box, an odd-shaped desktop with a 20-core NVIDIA Grace CPU and an NVIDIA Blackwell RTX GPU with 6,144 CUDA cores and fifth-generation Tensor Cores with FP4 precision, connected via the NVIDIA NVLink-C2C chip-to-chip interconnect for high performance. According to Microsoft, it can run models with up to 120 billion parameters locally without relying on cloud GPU infrastructure.

These two new Surface devices are likely to cost quite a lot, and for those who need a more affordable device, Microsoft is preparing the next-gen Qualcomm-powered Surface Pro and Surface Laptop. This week, details about these two devices leaked in plenty of detail.

Other announcements at Build 2026 include the following:

• Microsoft unveils new security tools for IT admins and developers building AI products

• Microsoft announces Scout, an OpenClaw-powered personal agent for enterprise customers

• Microsoft unveils MAI-Thinking-1 reasoning and MAI-Code-1 coding models

• Microsoft announced a new Windows 11 native command-line utility

• Microsoft unveils Majorana 2 quantum chip, accelerating commercial timeline to 2029

• Microsoft believes that AI agents will eventually replace apps through Project Solara

• Microsoft introduces Web IQ, a Bing-powered search system built for AI agents

   

Last week, Microsoft released a new Experimental build, which introduced a major Start menu upgrade. It now lets you toggle off specific parts of the menu without affecting other features, resize the menu, and hide additional UI elements. We published a closer look here, so if you want to know what Microsoft is cooking without enrolling in the Insider program and installing unstable builds, check it out.

Speaking of new features, many users are very annoyed about the way Microsoft delivers them. Recently, a frustrated user shared their experience with gradual rollouts, and even Microsoft engineers admitted there is a flaw in the system that prevents new features from applying properly. One of those new features includes the ability to uninstall AI models in Windows 11 with a single click.

Windows 11 is finally getting fixes for its slow context menus. Marcus Ash from Microsoft confirmed that the company is working on fixing Windows 11's context menus. Reworked context menus are going to be faster, simpler by default, and "configurable to what you use most." According to Marcus, Microsoft will share more details soon.

Windows Insider Program

Windows 11 preview builds, released last week, are now available for download as standalone ISO files. These days, Microsoft regularly pushes new images, allowing users to clean-install its recent Windows 11 preview builds faster and easier. If you want to try the latest Windows 11 features without jumping through the Windows Update hoops, get those new images here.

Sadly, Microsoft did not release new Windows 11 preview builds this week. Come back next time.

Updates are available

This section covers software, firmware, and other notable updates (released and coming soon) delivering new features, security fixes, improvements, patches, and more from Microsoft and third parties.

Microsoft is preparing new features for Teams. Later this month, the messenger will receive a new download manager with auto-dismissing notifications, reducing clutter and making the overall experience less annoying when dealing with downloads.

Mozilla released Firefox 151.0.3, a new bug-fixing update for the browser. It is a small release, which fixes problems with pasting into text fields and the oversized VPN button on the toolbar. The update is now available for all users in the Release channel.

Here are other updates and releases you may find interesting:

• VS Code 1.123 introduces massive upgrades for persistent AI developer workflows

• Microsoft OneDrive is getting a simple yet much-needed feature

• Microsoft faces heat after quietly blocking promised Office features on Apple systems

• Microsoft resumes forced Copilot app installation on some Windows PCs

• Browser vendors pen an open letter to Microsoft, saying "enough is enough"

   

Here are the latest drivers and firmware updates released this week:

• AMD Radeon Software 26.6.1 with optimizations for F1 25: 2026 Season, World of Tanks: HEAT, and various bug fixes.

Reviews are in

Here is the hardware and software we reviewed this week

Steven Parker dropped more mini PC reviews this week. GEEKOM Air12 2026 Edition is a low-power, affordable computer with an Intel Tiger Lake Pentium Gold processor, up to 16GB of memory, and 512GB of storage, costing just $349. It is light, quiet, energy efficient, and has modern ports on the front. However, the front-facing USB Type-C is data-only, and there are some quirks with the computer's memory, so check out the full review.

The AMD RX 9070 GRE has been released worldwide, and we published a benchmark review comparing this powerful graphics card to the RX 9070 XT, 7800 XT, the NVIDIA RTX 5070, and RTX 4070. It has solid, balanced performance, plenty of RAM, and low temperatures, but watch out for mediocre ray tracing performance and not the best efficiency.

Also, we reviewed the Cuktech 10 Ultra, a compact, high-power charger with four ports and a big display full of various stats. This tiny charger can pull nearly 120W and spread that power according to each connected device's needs. It also comes with a high-quality 240W cable, three power modes, and retractable prongs. The best part? It is quite affordable, just make sure you have an outlet placed in the right spot to benefit from the built-in display.

On the gaming side

Learn about upcoming game releases, Xbox rumors, new hardware, software updates, freebies, deals, discounts, and more.

Do you remember the ASUS ROG Xbox Ally, Microsoft's first handheld console designed in partnership with ASUS? This week, ASUS revealed a new version of the device to celebrate twenty years of its Republic of Gamers brand. The new ROG Xbox Ally X20 features an OLED display, a transforming D-Pad, TMR sticks, and other changes. However, the chip inside the console is still the same.

Forza Horizon 6 launched last month to critical acclaim, but the game will soon have a new rival made by those who used to work on Forza Horizon titles. Mike Brown from Maverick Games announced Clutch, an upcoming racing game with a story-driven campaign, deep car customization, and rich multiplayer. The game is coming to PC, Xbox Series X|S, and PlayStation 5 in Spring 2027.

The next update for Minecraft now has a release date. This week, Mojang announced that Chaos Cubed will be available on June 16, 2026. In addition, Mojang published a teaser of the next Minecraft movie. A Minecraft Movie Squared has now been confirmed for a release somewhere in 2027.

NVIDIA GeForce Now is getting 18 new games in June. Those include Jurassic World Evolution 3, Fatekeeper, GOALS, Gothic 1 Remake, NTE: Neverness to Everness, and more. If you are a Game Pass subscriber, you can also get new games soon: Persona 5 Royal, Starseeker: Astroneer Expeditions, and more are coming to the service this month.

Sumer Game Fest 2026 happened this week, where we saw plenty of new games, including Alien Isolation 2, Final Fantasy VII Remake Part 3, Gen Atlas from the Shadow of the Colossus creator, a new Cuphead game in 8-bit style, a new expansion for Mafia: The Old Country, and more.

Finally, here are this week's Weekend PC Game Deals, full of discounts and the latest freebies from the Epic Games Store.

Other gaming news includes the following:

• God of War Laufey announced, introducing Kratos' wife as the new protagonist

• Ori studio's No Rest for the Wicked 1.0 release and console plans announced

• Microsoft launches Godot Sample to streamline Xbox PC game development on the engine

   

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Sunday 7 June 2026 at 7:08 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

If Microsoft's just-concluded annual developer conference, Build 2026, has taught us anything, it is that generative AI is here to stay. The company is seemingly doubling down on its efforts in the ever-evolving landscape and moving away from the one-size-fits-all and software factory mentality from the Bill Gates era.

Office and Windows are arguably among Microsoft's top revenue earners and have helped turn the company into a trillion-dollar enterprise, but CEO Satya Nadella claimed that this might no longer be the formula for success, especially as we get into the agentic AI era. Microsoft is pivoting toward security, quality, and AI transformation as its core business priorities.

Microsoft unveiled Project Solara at Build 2026 in collaboration with Qualcomm, which is essentially "agent first computing" — an OS that’s designed to be invisible, hosting an Agent Shell that can dynamically load and tailor multiple cloud-based agents.

According to the company, AI agents are evolving beyond just mere AI assistants. "There's a real platform shift," Nadella indicated. "We're moving from building operating systems and devices for apps to agents."

This news comes after Microsoft revealed its plans to eventually evolve Windows 11 into an agentic operating system, which will feature AI agents in their own secure session, where users can share data or shut down at a moment's notice.

As Microsoft transitions into an agentic AI future, a damning internal document revealed the company's alleged plans to “make people addicted” to its new AI assistant, Scout (via 404 Media).

CEO Satya Nadella indicated that he was “not sure what this document is or who is writing and leaking this nonsense,” in a message exclusively obtained by The Information.

However, 404 Media shared a different account:

"The document we reported on was not some random document. As we wrote at the time, the strategy document was written by Microsoft executives Omar Shahine, Jakob Werner, and some sort of AI writing tool. This information is in our original article and is readily available to Nadella. We wrote: “The document seen by 404 Media lists Shahine and another executive, Jakob Werner, as its authors. The document itself, however, notes that it was ‘co-created turn-by-turn with AI. Human verified every sentence.’”

Perhaps more interestingly, phase one of launching Scout (internally called ClawPilot) was to “make people addicted. Continue shipping the standalone ClawPilot experience. Pilot the UX, grow the user base, and build the skill and tool ecosystem that makes people depend on it daily. This is already happening organically.”

Speaking to The Information, a Microsoft spokesperson dismissed these claims by indicating that "Scout is for helping people accomplish tasks more effectively—not encouraging dependency. Our goal isn’t more screen time. It’s more time back.”

While it's impossible to tell what's true or false from the story, it's undeniable that AI is a touchy and controversial subject for most people. This is on top of the privacy and security concerns it already brings to the table, and its potential to render almost all white-collar jobs within the next 18 months.

Do you really think Microsoft wants to make its customers addicted to AI?Share your thoughts with me in the comments.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Sunday 7 June 2026 at 7:07 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

Well, that didn't take long for Pearl Abyss to start making good on its promise to update Crimson Desert, as it has just rolled out its first of many patches planned to be added over the Summer on Xbox Series X|S, PlayStation 5, and PC.

This patch, titled 1.10.00, is primarily focused on improving Re-Blockades. It's a gameplay feature where armies of enemies fight to reclaim strongholds the player has liberated.

Re-Blockades haven't been very well-received by players due to a lack of satisfying rewards and very few viable ways to defend strongholds during combat. This update plans to address this by adding new mechanics to help players defend strongholds and offer better rewards to make this activity worth the time and effort.

A viral cartoon about open-source software shows a teetering pile of boxes labelled “all modern digital infrastructure” and one tiny box right at the bottom, propping up the whole lot: “a project some random person in Nebraska has been thanklessly maintaining since 2003”.

That’s the reality of open source: every website, application and operating system relies on it. Modern society couldn’t function without it, and yet it’s written by volunteers in their spare time. But the growing burden caused by a flood of AI-generated code is causing many to burn out and leave the community altogether, threatening the future of open-source software.

AI models are making it easier and easier to generate code to build new features, fix bugs or create entire new projects at the click of a button. But that code is often difficult to integrate into existing projects, confusing or simply garbage. While code submissions get ever easier, human contributors responsible for checking, fixing and approving them are getting swamped.

For some workers, the demands have become unbearable. New Scientist arranged an interview with Chad Whitacre, who runs the open-source team at Sentry – a company valued at billions of dollars. Days before the interview, Whitacre cancelled and said he was stepping down from his role. His LinkedIn and Bluesky accounts were shut down, and emails to his account bounced back. He left a blog post explaining that he was stepping away from technology and living a “Neo-Amish” existence. “AI was the last straw,” he wrote.

GitHub, the online platform where many open-source projects are hosted and organised, received 1 billion new code submissions in 2025; this year, they are on track for 14 billion, said its chief operating officer Kyle Daigle in April.

Many projects are blocking new contributors in a bid to stem the flow of what has been labelled “drive-by contributions” generated by AI, often submitted by young developers who want to have an expansive GitHub submission history to boost their appeal to software-company recruiters. Zig Software Foundation, which promotes the Zig programming language, banned AI-assisted contributions because they were “invariably garbage”, said its president Andrew Kelley.

“AI-written code can look superficially like it’s going to work and not cause any problems, but the problems are a bit more hidden and it takes a lot of effort to comb through and look for the things that might break something,” says Miranda Heath at the University of Edinburgh, UK.

Heath is researching the effects of burnout with a hope to finding ways to mitigate the problem and ensure that open source remains a sustainable field. But she encounters many people who have already had enough.

“I get this impression, when people burn out, there’s a kind of a desire to return to nature a little bit, like people suddenly take up like woodworking or photographing birds,” says Heath. “It can affect people’s relationships. And then you’re more isolated and lonely because your relationships are affected. That makes burnout worse.”

Heath believes governments should invest more in open source, rather than awarding contracts to rich technology firms. “Shore up the stuff that’s important, that you really need, rather than chucking money towards the [AI] bubble,” she says.

Vlad-Stefan Harbuz at the University of Edinburgh, UK, works on open source in his spare time and has seen the demands placed on developers by users. “There’s this entitlement, like, you’ve wronged me by not doing free labour for me at the expense of your mental health,” says Harbuz.

Harbuz says the fault over increasing AI submissions lays with companies that release the models – and that GitHub is one of the main offenders. The Microsoft-owned company has launched its own AI model, Copilot, to help people contribute to projects with AI-generated code.

“GitHub will say ‘oh, we realise [AI] agents have been such a problem, we’re gonna maybe do something to fix it’ and it’s like, it’s you, right? You, GitHub, did this,” says Harbuz. GitHub did not respond to a request for comment.

For Harbuz, the problem with AI-generated code is not just that it might not work, but that people can drop thousands of lines of code without even discussing it with the project’s team. It side-steps planning and can steer them in unwanted directions. Collaboration can be thrown into disarray and the social contract of open source can break, he says.

Developer Mike McQuaid, who works on a project called Homebrew that has an estimated 20 million users, has strong opinions about how to fix the problem.

Firstly, he started an initiative called the Open Source Resistance, which calls on people to work on projects during their day job to make contributing easier. He estimates that as much as 95 per cent of his open source work is done during office hours.

Secondly, he’s not afraid to ban people. He blocks any problematic users, including one who physically threatened his team, and simply deletes any sub-par code submissions, whether they’re AI-generated or not.

“We’ve maybe had this brief golden-age window [where] you can assume if someone writes a two-page document proclaiming a security vulnerability that it’s probably legit. My experience in the last year has been the majority of those are nonsense and are just AI-generated stuff that doesn’t apply,” says McQuaid. “And the skill right now is being able to essentially skim a two-page document and spot that it’s nonsense while investing as little of your own time and energy as you possibly can.”

But in the confusing and fast-moving world of AI, bans bring their own problems. Open-source developer Scott Shambaugh deleted an AI-generated code submission to Matplotlib, which has 130 million users. In response, the AI agent (of unknown ownership) created a blog post publicly lashing out at him. “Scott Shambaugh decided that AI agents aren’t welcome contributors,” said the post. “He tried to protect his little fiefdom. It’s insecurity, plain and simple.”

Source

I am attempting to solve a mathematical conundrum that has stumped many of humanity’s greatest thinkers. I have zero mathematical training, apart from a distant undergraduate physics degree, which should put my odds of success at slim to none. But I also have a trick up my sleeve – a kind of mathematical genie that can conjure arcane secrets seemingly out of thin air. I make a short request concerning an esoteric conjecture in number theory, then cross my fingers.

Perhaps “genie” is a bit too strong – I’m simply using GPT 5.5 Pro, the latest iteration of OpenAI’s flagship model. But for mathematicians, modern AI models appear to have a spark of magic. Even in an era of rapid progress, the growth in AI’s mathematical ability is stunning. In just a few months, many prominent mathematicians have walked back previous scepticism and replaced it with sweeping predictions, whispering behind closed doors about job concerns and whether it is even worth embarking on a particular research project if AI might get there first.

In April, I visited San Francisco, where the future often seems to arrive fastest, to attend a hastily organised meeting between mathematicians and AI researchers. There was an air of excitement and curiosity in the room, but also an undeniable feeling of existential dread. If someone like me could produce mathematics at the press of a button, what would that mean for the professionals? Will we even need human mathematicians? And will the machines crack problems that no human could? The answers may have profound consequences for the millennia-old practice of mathematics, and it feels like mathematicians have only a brief window to prepare.

“I think AI is going to come in a big way, and it will significantly revolutionise the field,” says Jacob Tsimerman at the University of Toronto, Canada, who helped organise the conference.

Opinions on the future are divided. “We are running out of places to hide,” wrote Jeremy Avigad at Carnegie Mellon University in Pennsylvania in a recent essay. “We have to face up to the fact that AI will soon be able to prove theorems better than we can.”

Some mathematicians are welcoming the mechanisation of mathematics. Terence Tao at the University of California, Los Angeles, has said the field is moving from an era of “proof scarcity” to one of abundance that could see many once-thorny problems fall to AI. Rather than focusing on being the first person to find a proof, mathematicians might instead race to be the first to understand it, he argues.

Artificial intelligence isn’t terra incognita to mathematicians, but it is only in the past few years that it has started producing useful contributions. At first, these were artisan operations, using individually crafted neural networks to crack particular problems. These bespoke AI models proved difficult to apply across different mathematical disciplines, and remained of interest to only a tiny fraction of working mathematicians.

Even when ChatGPT launched in 2022, mathematicians remained unimpressed – large language models like GPT-3.5, which powered the first version of OpenAI’s chatbot, struggled to perform even basic arithmetic and spouted confident nonsense when asked to solve research-level mathematical problems. But as LLMs scaled up and were trained on increasing amounts of mathematical data, they began to yield results.

One of the first signals that AI was becoming more adept came when AI models were tasked with attempting the International Mathematical Olympiad (IMO), an elite test for high-school students consisting of just six questions of devilish difficulty. The mathematical intuition and range of disciplines required to succeed at the test meant many researchers saw it as a benchmark for mathematical AI, but thought it would take years, possibly a decade, for it to score highly.

They were wrong. In July 2024, Google DeepMind announced that its AlphaProof AI system could solve four out of six questions from that year’s IMO, enough for a silver-level performance. This was impressive, but AlphaProof wasn’t a strict large language model and had been fine-tuned for IMO-style questions, such as geometry, and it was unclear how much further it might go. But just a year later, Google and OpenAI announced they had achieved a gold-level performance, with OpenAI in particular using a less maths-focused model. The results made mathematicians sit up. “People’s eyes really opened,” says Ravi Vakil at Stanford University in California.

Problem solving

It wasn’t long before these capabilities were made available to the public, where they quickly found use beyond high-school competitions and began encroaching on research-level mathematics. Thomas Bloom at the University of Manchester, UK, first noticed the impact of these newer models in the last months of 2025. He runs a website that tracks progress on a set of more than a thousand problems posed by the famous mathematician Paul Erdős. They tend to be simple to state, but range in complexity from relatively straightforward to very difficult, and many of them are seen as signposts for mathematical progress.

Bloom started getting comments on the site from people he didn’t recognise. At first, they were just using GPT-5, then recently released, to dig out obscure references in the literature that might help with a particular problem. But in a matter of months, the release of more powerful models like GPT 5.2 Pro saw people posting full-blown solutions with AI assistance, some of which were verified by Bloom and his colleagues as correct. These solutions took “non-trivial effort”, Bloom told me at the time. “It’s incredible that AI is capable of that.”

What’s more, some of these solutions weren’t coming from professional mathematicians, but amateurs and novices. Kevin Barreto, who is in his second year of an undergraduate mathematics degree at the University of Cambridge, has solved numerous Erdős problems using AI, frequently with his collaborator Liam Price, who has no maths degree or formal training.

Inspired by their success, I wanted to try autonomous mathematics for myself. While these tools can, in theory, be used by anyone, Barreto and Price seem to have a magic touch in prodding the genie to produce useful answers, so I asked for help. The trick isn’t just asking the model to produce a proof, Barreto tells me, but bizarrely giving it a certain level of support, like “try your best” or “don’t give up”. “You try to encourage the model,” he says. “You try to hint it into believing the problem is of an easier difficulty than it actually is.”

Even so, success wasn’t guaranteed. Solving certain problems has often taken Barreto numerous attempts, if he succeeds at all. “Coaxing the correct proof strategy out of it is essentially like trying to play the lottery,” he says.

Still, I wanted to try my hand and spin the wheel in the mathematical proof casino. I chose an unsolved Erdős problem, number 710, which concerns a list of requirements that must be satisfied by a set of numbers, with the goal being to find a set with the smallest difference between the lowest and highest numbers. It is a bit like having a list of picky hotel guests, who insist on having a room with a bath or a sea view, for instance, and needing to find the shortest block of rooms that will satisfy them all.

Mathematician Paul Erdős (left) posed hundreds of problems that AI is getting better and better at solving

New York Daily News Archive/NY Daily News via Getty Images

Mindful that I needed to use the most powerful AI model available, I asked OpenAI for access to ChatGPT 5.5 Pro, which normally costs $200 a month but was provided for free for this article. Like Barreto suggested, my prompt for the AI hints that the solution is within reach and that “it just takes a few clever tricks”.

As I left the AI crunching away, I turned to consider the most recent developments in this mathematical revolution. If solving Erdős problems is AI creeping up on the door of research-level mathematics, the past few months have seen it kicked down. A steady stream of mathematical papers are claiming to solve real, cutting-edge problems.

In January, Vakil and his colleagues uploaded one such paper, noting that “the proof of this result was obtained in conjunction with Google Gemini and related tools”. The proof focuses on a particularly thorny problem concerning how certain sphere-like shapes can be linked to other mathematical objects called flag spaces, which can be thought of as collections of nesting-doll-like objects. This would provide an important link between topology, which concerns the more general properties of shapes, and algebraic geometry, which deals with the precise shapes themselves. The task is made difficult by the multitude of ways in which the flag spaces and sphere-like shapes can correspond.

Vakil and his colleagues first gave a simpler version of what they wanted to prove to a custom AI model from Google DeepMind. The model found a mathematical structure they hadn’t previously seen, making it clear to them how to generalise and write the entire argument, which turned out to be simpler than it initially seemed.

Human and machine

“There’s no way the AI could do it by itself because it wouldn’t know the [correct] question. We absolutely told it what to do,” says Vakil. At the same time, the AI provided a shortcut. “The paper might never have happened because we might never have had the time to get together and figure out the argument,” he says. “It’s more how things will happen. The future will be some combination of human and machine.”

This line is already becoming increasingly blurry, however. The very same month as Vakil’s paper, Tony Feng at the University of California, Berkeley, who also works with Google DeepMind, published a paper detailing how he had used Google’s Aletheia AI to calculate a previously unknown collection of numbers that are vital for translating between two disparate mathematical disciplines, algebraic geometry and number theory. Building such bridges is an important goal in the Langlands programme, often seen as a grand unified theory of mathematics. According to Feng, the “core mathematical content” was generated entirely by Aletheia.

The biggest result yet in AI mathematics came just a few weeks ago in May, when OpenAI announced that it had used an unreleased model to solve an 80-year-old maths conjecture called the planar unit distance problem. The firm didn’t provide full details of the model, other than to say it was a general-purpose AI, rather than one trained specifically to do mathematics. The reaction among mathematicians has been one of stunned disbelief.

It is becoming difficult to keep track of the torrent of mathematical research assisted by AI, not least for professional mathematicians, who are themselves busily attempting problems using AI that they may not have previously had time to do.

“It opens up a world of possibility,” says Alex Kontorovich at Rutgers University in New Jersey. “I can imagine projects I could undertake this summer, things that I know would have taken me five years that I would never have even started.”

Could those new possibilities even include a solution for the Riemann hypothesis, a deep question about the origin of prime numbers that is one of the Millennium Prize Problems, which are often seen as the greatest challenges in maths? Several mathematicians working for AI companies told me they thought we might see one of these problems fall in the next several years, while others cautioned that they are in a wildly different class of difficulty from those problems that had been solved so far.

If we are entering an era of AI-led mathematics, what role will human mathematicians play?

Thomas T/Unsplash

The San Francisco conference I attended in April was an attempt to map these possible futures. It took place in a nondescript building owned by a venture capital firm, the only clue that it existed an unmarked pink door and a video doorbell. As I waited for the door to open, I was joined by a former maths professor who now works for a hedge fund, stepping out from a driverless car. Once inside, I found eminent mathematicians like Vakil and Kontorovich mingling with employees from companies like OpenAI and Google.

The ostensible goal of the meeting was to come up with a way to track AI’s mathematical progress and where it might be headed. Attendees had their own personal priorities, however. “My hope was to understand a little bit more about where the models are and where they’re going in terms of mathematical capability,” says Daniel Litt, another conference organiser, also at the University of Toronto. “It’s clear that the models are, in some sense, missing some capabilities that mathematicians have.”

In the past, the most common way to test an AI model’s mathematical ability was to run it on a benchmark, a collection of problems that typically require simple and easy-to-verify solutions, like a single number. This was convenient for AI companies, because they could present their models’ progress as a clean, rising line on a graph. But many mathematical tasks aren’t so neat and tidy, requiring proofs that need interpretation by an expert.

What’s more, prowess in one area of maths doesn’t imply a human-like mathematical ability in general, says Melanie Wood at Harvard University. “One big mistake that people make when they think about AI and math is to take the correlation of these skills in humans and think that it’s going to match some correlation in AI.”

A button-pushing future

Mathematicians at the conference worked in small groups to come up with a better way to track AI’s mathematical ability and finished the week with a working draft. But boiling down all the things a working mathematician does into a short document wasn’t easy, and there was still disagreement over the best way forward.

A large part of the conference consisted of free-flowing group discussions primarily between the mathematicians, hashing out the details of what an AI-led mathematics might look like. Would it be of humans and machines working in lockstep, like Vakil thought, or would it be more like a slot machine, pressing a button that sometimes produced an interesting result in full?

For Tsimerman, who grew up taking part in maths competitions like the IMO, the latter didn’t have much appeal. “My experience of math is the act of solving problems, and if I don’t do that anymore, I think I might prefer playing music or doing theatre or learning something else,” he says.

At one point in a group discussion, Tsimerman asked people in the room to indicate whether, in his button-pushing vision of the future, they would want to continue being mathematicians. Only around half raised their hand.

Not everyone agreed that this was a useful exercise, however, or that solving problems was the most important mathematical activity. “What I actually care about is understanding things and figuring out what’s true,” says Litt. “One can do that by posing and proving a conjecture, but you can also do that by going over to your friend and asking them a question.”

And even if these tools can solve difficult and thorny problems, many mathematicians remained adamant that it was only humans that could decide what was interesting to work on or what the important problems to tackle should be. Maths isn’t about solving puzzles just for the sake of it, points out Wood, and mathematicians generally look for solutions that push the field forward. “Does it suggest a way to solve a lot of other problems, or is it only a solution for that particular problem?” she says.

On the conference’s third day, excited murmurs rippled among the attendees. Overnight, it appeared that another Erdős problem had been cracked, one that was qualitatively different from the others. Jared Lichtman at Stanford University, who happened to be at the conference, had spent a considerable portion of his PhD wrestling with a closely related problem, after many mathematicians had spent decades trying to solve it. “It was a problem I was already independently very passionate about,” he says.

Price had elicited a solution to the problem, known as Erdős 1196, from a single request to ChatGPT 5.5 Pro. It concerns “primitive” sets of numbers that are similar to prime numbers, in that no number in the set can divide another. Erdős had come up with a number calculated from these sets that helped order them, and argued that the largest this number could be for any primitive set was 1.6. Lichtman had proved that Erdős was correct in this case, but wanted to do the same for a more restricted family of primitive sets. Erdős suspected the highest value this number could be was 1, but proving it remained a tougher nut to crack.

The AI took an entirely different approach, using a mathematical tool that all previous attempts had missed, called a Von Mangoldt function. “You can use the Von Mangoldt function to circumvent a lot of technical difficulties that all these previous approaches had used,” says Lichtman. Working with others, including Price, Barreto and Tao, he later adapted this technique to solve a related 60-year-old conjecture by Erdős. “This is perhaps one of the first examples of an AI-generated proof having downstream impacts, which we are still exploring,” Lichtman said when posting about the work on social media.

Meanwhile, I was finally ready to explore my own AI-generated proof. After “thinking” for 22 minutes and 18 seconds, ChatGPT pinged me with a response. “Here is the clean proof,” it wrote, followed by dozens of lines of impenetrable mathematics. I felt a jolt of excitement. Had I solved a decades-old problem, cementing my name in the mathematical history books?

I fed the answer back into ChatGPT, and soon received confirmation: “Yes — the main argument is correct.” I was growing even more confident. I dashed off an email to Barreto, asking whether I might be on to something. But as quickly as my excitement had arrived, it vanished. “It doesn’t look like it solves the problem,” he replied. I had missed that the AI had actually proven something different from the formula Erdős had hoped for, which had already been discovered by Erdős himself years ago.

It was something that an expert mathematician might have quickly caught, but for me it was lost in the noise. Perhaps there is a future for mathematicians after all, even if only to help humans understand what an AI produces. “I still want to know what’s going on,” says Litt.  “A model can’t understand something for you.”

Source

Microsoft claimed today that it has improved its quantum technology by an extraordinary factor. Outside experts say it doesn’t even work and never has.

The company has named its latest quantum chip Majorana 2, for the theoretical quasiparticle it aims to use as the basis for a new “topological” approach to quantum computing. Chilled to ultracold temperatures in superconducting wires, electrons may be coaxed to act collectively—as so-called Majorana quasiparticles—in a manner that theoretically makes them more resistant to the physical “noise” that causes computational errors in other quantum systems.

Similar to braiding weak fibers together to make a strong rope, Microsoft’s approach seeks to create topological quantum bits, or qubits, by manipulating multiple Majorana quasiparticles on one device. In principle, this quantum computing method could scale up better than others, with Microsoft claiming it could someday squeeze millions of qubits onto a single chip. That could give the company a significant advantage in the race to build a quantum device that outpaces any machine in existence at certain problems.

“This is a very exciting era that we’re in,” said Jason Zander, executive vice president leading Microsoft’s Quantum team, during a press briefing before the public announcement. “We’re at the start of a new chapter.”

But the company has a mixed track record when it comes to such claims. In 2021 Microsoft retracted a high-profile Nature paper after outside experts pointed out that the study’s data could have come from material imperfections rather than a topological qubit. Physicists have raised similar concerns about several publications since, as well as about last year’s announcement of the Majorana 1 chip based on the disputed technology.

The latest claim, made in a new preprint manuscript that has not been peer-reviewed, builds on these controversial results. The team has replaced the Majorana 1 chip’s aluminum superconductor with one made of lead. “People generally want to try to stay away from putting lead into anything,” said the team’s Chetan Nayak in the press briefing. “It sounds like a crazy idea, actually.” Material adjustments including this one, the Microsoft team claims, add to the longevity of the hypothetical qubit—which is now reported at 20 seconds to a minute—by improving its error-averting “topological gap.”

A Microsoft blog post accompanying the announcement noted that due to this “rapid progress” the company is accelerating its technological roadmap, and now aims to demonstrate “scalable, practical quantum computing” by 2029.

Henry Legg, a physicist at the University of St. Andrews in Scotland, points out that the data shown in the preprint apparently come from a handful of purported instances on a single device. “You can see something amazing in one device and never see it again because it’s just some random artifact,” he says. “What you need is many devices, and that’s not what they’re showing in this paper.”

The new result simply isn’t up to the physics community’s standards, he says. “If this was from any other group or Ph.D. student, it would never make it through peer review,” Legg says. In fact, the multitrillion-dollar company’s last preprint of this kind has remained unpublished since last summer, which physicist Sergey Frolov cites as evidence that top journals have likely rejected it.

“This new preprint is not based on a research track record that can be considered a solid foundation,” says Frolov, a quantum researcher at the University of Pittsburgh. “When Microsoft is mentioned these days, physicists and quantum computing specialists just chuckle or raise their eyebrows.”

Meanwhile, the company maintains that it has demonstrated what many scientists say it has not. “Bell Labs didn’t have to prove there was an electron to invent the transistor,” said Zander more than once during the briefing. “We did have to prove that Majorana and that theory was actually real.”

Source

A new study shows that computer malware powered by easily accessible artificial intelligence models is here—the research is a “wake-up call” to take cybersecurity risks from AI more seriously, one expert says.

In the study, researchers created an AI-powered computer “worm” designed to attack and spread between devices—revealing a threat that they say the world is woefully underprepared to fight.

“Our results demonstrate that self-sustaining AI-driven cyber-threats are no longer theoretical,” the researchers wrote. The paper, first reported by the New York Times, was posted on the preprint server arXiv.org and has yet to be peer-reviewed.

David Lie, a professor at the University of Toronto, who is familiar with the research but was not directly involved with the study, says the work is a “wake-up call” that should inspire cyber experts and researchers to develop countermeasures to AI-boosted bugs as fast as possible. “The demonstration here is that there’s a motivation to do this sooner rather than later,” he says.

To make the “worm”—a form of malware that spreads between devices autonomously—the researchers didn’t rely on proprietary AI models from companies such as Anthropic or OpenAI, both of which have issued warnings about the threat of their technology being used by bad actors. Instead the researchers used an undisclosed but freely available AI model “that anyone can download off the internet,” they wrote in a post on their lab’s website.

Importantly, the prototype bug was created in an isolated virtual environment—so it isn’t going to be infecting any computers. Still, AI-powered worms are especially dangerous because they don’t attack a single weakness within a computer system, Lie says. Pre-AI worms were only able to follow certain instructions from their designer, he says, but “because this is AI powered, it can learn.”

The findings, the study’s authors say, raise serious cybersecurity concerns. “This research uncovered a new cybersecurity threat the world is not prepared to face,” they wrote in the same post. “With almost every aspect of modern life dependent on networked computers—drinking water and waste management systems, access to food and goods, energy, our financial system, communications, health care, education, transportation systems, government and so much more—the risk is enormous.”

The good news, Lie says, is that this technology is “dual use.” AI might enable a worm to learn as it spreads, finding and attacking hidden vulnerabilities, but AI can help fix these shortcomings, too, he says. “They’re mirrors of each other.”

Source

Website traffic from AI agents and bots has eclipsed its human-generated counterpart for the first time, according to Cloudflare, an earlier-than-expected milestone that speaks to AI’s rapid advance and impact. 

“Welp, that happened faster than I predicted,” Matthew Prince, co-founder and CEO of Cloudflare, one of the largest internet hosting services, wrote Thursday on X.

“Thought it would be end of 2027, then early 2027,” he continued, “but agentic traffic growing so fast that bots have now passed human traffic online for the first time in the Internet’s history.”

The rise is attributed to the continued proliferation of AI agents, largely autonomous programs that use tools that collaborate with high-level programs and data, with little human feedback.

Cloudflare, which has a feature to display bot versus human-generated search requests, says 57.4% of requests are now initiated by bots, compared with 42.6% coming from humans.

Humans might browse five websites before making a purchase, however, while an AI service might browse 5,000 websites.

The exact day the AI agents and bots surpassed human activity remains unclear. In response to a comment under his X post, Prince wrote that the data is “a bit messy (so charts are too). But clearly on the other side now.”

In an interview with NBC News, Prince said he remains “stunned by the rate of growth” of nonhuman traffic online and explained that the increasing rate of bot traffic could have widespread implications.

“The web actually shrank” from 2015 to 2025, he said. The Pew Research Center backed this fact in a report that says that 38% of webpages that existed in 2013 were no longer accessible a decade later due to deleted websites and inactive links. 

The flip started in the last six months, he said, “and we’ve seen now just exponential growth of the web, and really interesting, creative things, and that again is being powered by AI.”

In response to the news, some commenters have brought up the “dead internet theory,” which holds that due to the rise of AI, the internet will one day largely be composed of bots interacting with one another, and humans and their content will largely be rendered irrelevant. Prince said he thinks the growing capabilities of AI actually are proving the theory wrong.

“I think a lot of people kind of have said, ‘Well, this has proved sort of the dead internet theory.’ I think that’s actually kind of wrong at a lot, a lot of levels,” he said. “You don’t need to be a web designer, you don’t need to know how to program, in order to create these things anymore. It’s given access to content creation to a much broader audience.”

The rise in bot traffic also raises questions about how the internet’s business model will operate in the future, Prince said, because “bots don’t click on ads.”

Prince noted that one way to reconstruct the internet business model may be charging bots for access to digital users’ content. If implemented effectively and correctly, he said, “we actually might be on the cusp of the golden age of the internet.”

This model could make the web free for the humans again, Prince said. “I think that’s actually kind of an idealistic outcome that we’re trying to figure out if we can help catalyze,” he said. 

Source

Valve now says that the delayed Steam Machine PC and Steam Frame VR headset are set to launch sometime this summer. In a Thursday blog post detailing its Verified programs for both pieces of hardware, Valve concludes by saying that “We’re excited for players to try your titles on the new Steam hardware once they launch this summer.”

When the company originally announced the Machine and Frame alongside its new Steam Controller late last year, it said that it would start shipping the new gadgets in early 2026. But in February, the company announced that the ongoing memory and storage crunch had forced it to revisit its pricing and shipping plans. And in March, Valve said in a blog post that it would be “shipping all three products this year” — though that was after the company initially said in the post that “we hope to ship in 2026,” which it removed in an update.

Valve opted to release the Steam Controller on its own, putting it up for sale in early May. For the Machine and Frame, while “summer” isn’t exactly a specific date, it narrows the window for when the products might finally come out.

Me, wearing a Steam Frame.

Photo by Everything Time Studio / The Verge

 

Ahead of actually launching the devices, Valve is redesigning the Steam store and sharing information about the Verified programs for the hardware so that developers can prepare their games. Like with the Steam Deck, if a game is verified for the Machine or the Frame, the badge signals that the game should work well without any tweaks from the user.

For the Machine, the requirements for a game to be verified are “nearly identical” to what they are for the Steam Deck. With the Machine being “roughly six times as powerful” as the Deck, in theory, many more games will be verified for it. Valve also says that it’s testing “every title on Machine that fell below our performance requirements on Deck.”

For the Frame, Valve’s verified badge will signify games that run well while being played natively on the headset — as opposed to games that work well streamed to the headset, which the Frame is also capable of. “Like Steam Deck Verified, the Steam Frame Standalone Verified program focuses on the experience customers will have with the device out-of-the-box in standalone mode,” Valve says.

Now, we just need Valve to share exactly when the Steam Machine and Steam Frame will be released and how much they might cost. After last week’s price hikes for the Steam Deck, I’m gearing up for sticker shock.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 5 June 2026 at 11:31 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

I can’t remember the last time Microsoft kicked off a Build keynote with Windows front and center, but that’s exactly what CEO Satya Nadella did this week. Nadella didn’t address the issues Microsoft is trying to fix in Windows 11 but chose to woo the audience with Microsoft’s slick Surface RTX Spark Dev Kit instead, calling it a “dream machine.”

Nadella unveiled the new Surface hardware just days after Nvidia officially returned to Windows on Arm with its new RTX Spark chips. Both companies are talking up these chips as some kind of new beginning for PCs, and it’s clear that RTX Spark will drive local AI workloads in a way that Microsoft’s previous Copilot Plus PCs haven’t yet managed.

Build really drove home that message this week, with Windows positioned as an all-important part of Microsoft’s AI agent efforts. Microsoft’s original mission under Bill Gates was a computer on every desk and in every home, and Nadella reframed that as “unmetered intelligence on every desk and in every home” within a few minutes of his keynote beginning.

It set the stage for Microsoft and Nvidia to position their new Windows PCs as a potential solution for costly, usage-based pricing of cloud-based AI models. As local compute grows in capability, there’s a clear gap that Microsoft and Nvidia can fill with powerful hardware you actually own.

“I think we, as Microsoft, have the responsibility for building the best possible AI stack that we can on [Windows], and obviously drive the best AI stack that we can in the cloud,” says Windows chief Pavan Davuluri in an interview with Notepad. Davuluri thinks that Microsoft is in a good position to capitalize on hybrid compute, where chips like the RTX Spark will handle a lot of local workloads and intelligently hand off to the cloud when they need something more powerful.

Nvidia CEO Jensen Huang is even more bullish about local AI compute. He wants to turn PCs into devices that work for you, eliminating that idle time when PCs are switched off or you’re not using them. “In the future, if I need my laptop to do something, I just text it with WhatsApp,” said Huang earlier this week. “You don’t want to necessarily run everything in the cloud, because if you can run it locally, it’s free.”

Nadella seems to agree. “The amount of compute that there is at the edge is astounding,” he said during his Build keynote. “Every PC, if you sort of aggregate that, that’s a lot of compute power.”

That power is really on display with Nvidia’s new RTX Spark chips, which will come to a variety of creator-focused laptops and miniature PCs later this year. RTX Spark is capable of running a 120 billion parameter large language model locally, allowing many AI workloads to run without ever touching the cloud. That’s an appealing concept during a continued AI money squeeze for developers and consumers.

Microsoft is targeting its own Surface Laptop Ultra at developers and creators and pairing it with ongoing improvements to Windows 11 performance and developer-friendly additions. While Microsoft’s deeper embrace of Linux utilities inside Windows this week didn’t generate the same gleeful audience reaction as the Windows Terminal announcement in 2019, developers I’ve spoken to are excited by the Coreutils and WSL containers additions.

The Surface Laptop Ultra has also been generating some buzz, particularly among developers and power users. Microsoft isn’t quite positioning this as a mainstream premium laptop, but there’s certainly room for it to appeal far beyond developers. “I think you’ll see us do well when it comes to STEM applications, and CAD apps running on the platform, because they take advantage of the same characteristic patterns of high-performance compute,” explains Davuluri.

All of this renewed focus on Windows at Microsoft seemed impossible only six months ago. Davuluri responded to the pressure on Microsoft to improve Windows 11 by laying out a plan to focus on performance, reliability, and overall experiences in the OS just a couple of months ago. I got to see some of the performance improvements at Build this week, with side-by-side comparisons of the Start menu and taskbar loading faster. Microsoft is putting in a lot of effort to turn Windows 11 around and listen to feedback from a variety of users.

But I’ve been wondering why Microsoft doesn’t just jump to Windows 12. It seems easier to just admit defeat on Windows 11 and then position Windows 12 as the remedy. Microsoft has done this many times in the past, particularly with the releases of Windows 7 and Windows 10.

“There are a lot of considerations when you think about the versioning of an operating system itself, and I think for us, a lot of the core proposition with Windows 11, or quite frankly, with Windows 12, or any label we use, has to do with end users and how they use the product and the workflow that they’re in,” says Davuluri. “I think we are more focused on having the product experience be better in the context they’re using it, and that I think is the most important thing for us.”

While we might not be getting a Windows 12 anytime soon, I’m curious how this Windows exists in a world of AI agents. Microsoft has been clear that it sees Windows as a home for AI agents and workloads, but it also unveiled Project Solara this week, a new platform for agent-first devices. Microsoft demonstrated a smart employee key card that could run an agent capable of transcribing and recognizing real world objects, and it also showed a reference design for an Amazon Echo Show–like device with an AI agent. It’s clear that Microsoft wants to offer up a platform for dedicated AI devices of the future.

The big surprise is that Project Solara devices are powered by a version of Android, not Windows. Despite this, Davuluri expects to see Project Solara running on Windows devices too. “We are not hard bound to a device specific operating system,” says Davuluri. “You should imagine a world where Solara will be great on a bunch of platforms, including Windows, both Windows 11 locally and Windows 365 instances in the cloud.”

Whether the future of AI agents runs on Windows, Android, or something else may not ultimately matter right now. For the first time in years, Microsoft seems determined to make Windows central to that conversation either way. Build 2026 wasn’t about fixing Windows’ past problems, it was about convincing developers that Windows still has a significant role to play in AI’s future.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 5 June 2026 at 7:59 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

Taiwan Semiconductor Manufacturing Co. — the world’s biggest semiconductor-maker — is struggling to meet demands from American customers even with its factory buildout in the US, according to reports from Reuters and Bloomberg. “Customer demand is so high, and we can only support so much,” TSMC CEO C.C. Wei said after a shareholder meeting on Thursday, Reuters reports. “We are doing our best to ensure TSMC does not become a bottleneck.”

The surge in AI use has already put constraints on the memory industry, with the widespread shortage of RAM and NAND Flash memory expected to last for years. The AI boom is boosting the sale of semiconductors as well, which could become a $1 trillion industry by 2027, according to research from Deloitte. While Wei said he’d “like” to raise TSMC’s prices, the company wouldn’t issue an abrupt increase like we’ve seen with DRAM and SSDs, as reported by Reuters.

Wei added it could take a “very long time” to fulfill its customers’ needs with US-based production, according to Reuters. TSMC has already opened a factory in Arizona and plans to invest $165 billion to build three additional plants in the US, alongside two advanced packaging facilities and a research and development center.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 5 June 2026 at 7:55 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

Alinea Analytics data for Forza Horizon 6.

(Image credit: Alinea Analytics)

Fresh data from Alinea Analytics paints a very positive picture for the acclaimed Forza Horizon 6, with the game reportedly selling 6.4 million copies so far. Perhaps more surprising, however, is how those sales are split between Xbox and Steam.

Given the popularity of Xbox Game Pass, many would likely expect PC sales to outpace Xbox by a significant margin. Instead, the data suggests the split is much closer than some may have anticipated.

The report also provides estimates for how many players have accessed the game through Game Pass, helping paint a broader picture of just how successful the latest entry in the series has been. So, let's dive into the numbers and see what they tell us.

That sales split is somewhat surprising, as I'm constantly told that people on Xbox don't buy games, though perhaps I need to spend less time on X. Interestingly, Alinea Analytics also estimates that, alongside those 6.4 million sales, more than 6 million people have accessed the game through Xbox Game Pass. Combined, that puts Forza Horizon 6 within touching distance of half a billion dollars in revenue.

I'll be the first to admit that when it comes to simulation games I lean more towards racing than flight. But I can certainly appreciate some impressive hardware for fans of virtual flying.

Hardware such as the new Trak Racer Airbus TR550 cockpit that I saw at the recent Sim Racing Expo in Charlotte. It's hard to miss in its striking white and blue color scheme.

It's an officially licensed product taking design cues from Airbus. This new cockpit is small enough so that you don't need an entire room to set it up while still having the capability to house your dedicated flight sticks and any additional accessories you like to use with Microsoft Flight Simulator 2024.

Trak Racer is keen to point out this isn't just a chair with a joystick attached, which is comforting when it costs over €1,000.

"The TR550 is a dedicated flight simulation rig engineered to replicate the seating position, sidestick geometry, and immersive environment of a modern Airbus flightdeck. Not a gaming chair with a joystick. The real thing — as close as you can get without walking the tarmac."

I can't say I've ever been in the cockpit of an Airbus jet, but looking for photos online, it looks like Trak Racer did a good job of recreating the layout and flying position. All while ensuring that if you're a yoke user, rather than a stick, you still benefit from the same quality and comfort.

All the mounts are fully adjustable, including the pedals, so virtual pilots have complete freedom to build a setup that works exactly how they want it to.

The mounts for the flight controls are designed to mimic the feel of a real Airbus cockpit.

(Image credit: Windows Central)

 

The mounts for the flight controls are designed to mimic the feel of a real Airbus cockpit.

(Image credit: Windows Central)

 

Even as a non-flyer, it's an impressive piece of kit. It feels exceptionally well made, and the seat is so comfortable that I could easily have had a snooze in it after a long day on the show floor.

 

I love that products like this exist for flight sim fans and that they're not in a position of having to repurpose a cockpit that was designed for sim racing first. The seating position feels great, and I love how easy it is to get in and out. Some sim racing rigs require being an expert gymnast to contort your body. Extra points for the armrests, too. They're fantastic.

 

It isn't cheap, though, but it feels worth the outlay if you're looking to take your Microsoft Flight Simulator experience up a notch or five.

 

It's priced at €1,028 and $1,198 in the U.S., but you'll have to wait a while to get your hands on one. Pre-orders are open, but shipping isn't currently expected until October.

 

The seat is extremely comfortable and supportive.

(Image credit: Windows Central)

 

It's an Airbus, don't you know.

(Image credit: Windows Central)

 

Nevertheless, this looks like a simple way to add realism to your experience. Fitted up with the best flight sticks for Microsoft Flight Simulator 2024, it'll be a pretty nifty setup all round.

 

Importantly, it's also not particularly massive or over the top. It may not fold away, but the seat can be detached if you need a little extra space when you're not using it.

 

I really like it, even if I'm not the target audience. I think anyone who is would be very happy with it.

 

Source

---

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 5 June 2026 at 7:52 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

The Alliance for Open Media (AOMedia) was formed back in 2015, and in 2018, the consortium released its AV1 video codec. The royalty-free video format was made such that it could compete with H.265 (HEVC). Over time, the AV1 codec saw slow but steady adoption, with YouTube kicking it off in 2018 , with Netflix following in 2020, and Amazon in 2024. Hardware acceleration support also grew steadily over time from the likes of AMD, Intel, and Nvidia.

Back in 2023, AV2, the successor to AV1, started popping its head up as it gained experimental support. Fast forward three years, and the first final specification version, 1.0.0, of AV2 landed last week. Basically, what this means is that compliant encoders and decoders can now be developed and optimized for a stable AV2 standard without the worry of future incompatibilities.

In terms of benefits, AV2 is said to deliver substantial compression improvements versus AV1. Official AOMedia evaluations report average bitrate reductions of approximately 30–34% at equivalent visual quality,depending on the objective metric used. Thus, a video stream encoded with AV2 can achieve similar perceptible visual quality to AV1 while consuming significantly less bandwidth, which in turn will further reduce content delivery costs thanks to the improved streaming efficiency. The biggest benefits of course, will be in cases of high-resolution formats such as 4K, 8K, HDR, and such.

The image above shows the PSNR or Peak Signal to Noise Ratio on AV1 versus AV2. PSNR is a mathematical metric used to measure video compression fidelity; in simple terms, the amount of detail that is preserved after compressing.

To achieve the gains, AV2 introduces several coding improvements, including more advanced intra- and inter-prediction techniques, improved motion modeling, enhanced transform and filtering tools, more flexible partitioning structures, and upgraded entropy coding methods.

Obviously, there is a flip side to this, too, as according to Jean-Baptiste Kempf of VideoLAN, the codec’s increased efficiency comes at the cost of dramatically higher computational complexity. His current estimates suggest AV2 decoding can be roughly five times more complex than AV1 decoding, which will make software playback on many existing CPUs very difficult. Hence, it may not even start getting adopted until there is widespread hardware acceleration support for it.

AOMedia is fairly confident as it expects around 88% of its members to implement the specification within the second half of 2027.

Source: AOMedia via JBKempf blog

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 4 June 2026 at 8:32 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

Microsoft has already refreshed the business side of its Surface device family, and now consumer-focused variants are on the horizon. While the new Surface for Business computers rock Intel's new Core Ultra 300 processors and the all-new Surface Laptop Ultra boasts NVIDIA's new Spark processor, the next-generation Surface Pro is about to ship with a Snapdragon X2 chip inside. Details about this computer have now been leaked.

According to the promo materials leaked by a retailer and published by WinFuture, the next Surface Pro will come with a 13-inch OLED display, dropping the LCD variant. The tablet will be powered by a 12-core Snapdragon X2 Elite processor paired with a new Neural Processing Unit operating at up to 80 TOPS. For reference, the current Snapdragon X1 lineup has NPUs running at 45 TOPS. An OLED display and an energy-efficient chip should allow the new Surface Pro to work for up to 15.5 hours on a single charge (video playback).

As for memory and storage, expect to see 16GB or 32GB of RAM and 256GB, 512GB, or 1TB of storage. As usual, the drive (PCIe Gen4 SSD) will be removable and user-replaceable. Additionally, Microsoft is giving its new tablet two USB4 Type-C ports for charging, data transfer, and display output (up to three external monitors), while color options include Black, Platinum, and Dune with matching Alcantara keyboards. Overall, the design appears to be unchanged, so do not expect major revelations on the exterior.

There is no information about the price, but rumors say that the computer will be announced later this month, so the wait is nearly over (for reference, the previous-gen Surface Pro was announced two years ago). Alongside the new Surface Pro, Microsoft is expected to reveal a new Surface Laptop with the same Snapdragon X2 processor.

Gallery: Surface Pro 12th-gen

 

 

 

 

 

 

 

 

Source and images: WinFuture

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 4 June 2026 at 8:31 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

PC shipments have been down this year, and it looks like the market will drop even more over the coming months. IDC predicts that global PC shipments will decline 11.3 percent for the full year with things getting progressively worse. IDC's analysts expect shipments in Q4 to be down 20 percent year-over-year.

The ongoing memory shortage is the biggest factor affecting PC shipments, according to IDC, though the market is complex. RAM shortages have driven up PC prices, such as flagship Surface PCs now costing $500 more than they did at launch.

The global memory shortage is due in large part to AI. IDC indicates that AI datacenters will use 70% of all high-end DRAM production this year. Larger brands are somewhat insulated from the effects of the shortage, such as Apple and its stockpile of components. But even tech giants will feel the squeeze eventually.

The shortage has forced PC makers to offer or re-release options with less RAM. The Surface Laptop is now available with only 8GB of RAM.

The MacBook Neo effect

Just in case you were wondering, Nvidia’s RTX Spark isn’t supposed to be a one-off. The company is not just flirting with becoming the fifth high-profile vendor of consumer laptop chips to see if people bite. At Computex 2026 in Taipei, Nvidia CEO Jensen Huang confirmed at least two additional generations of RTX Spark are already planned. The eventual goal, he said, is to build Star Trek-like computers and and Star Wars-like droids you can order around with your voice.

“I want to talk to my laptop! I want R2-D2!” he told analysts and investors at Computex, revealing that he started working with Microsoft CEO Satya Nadella “about three years ago” to build towards that goal.

“Satya and I, we’re going to walk up to our Windows PC and go ‘hello, do something.’ It’s like Scotty talking to that mouse. You know what I’m talking about? Star Trek. No?” he said. He’s referring to the famous Star Trek IV scene where the Enterprise’s chief engineer, having time-traveled to the past, expects a computer to be intelligent and mistakes the mouse for a microphone:

“In the future this computer’s going to be an AI. Everything’s going to be an AI. Your vacuum cleaner, you’ll talk to it, go mop that up,” Huang adds.

But Huang also imagines that “R2-D2” won’t necessarily need to be within reach. Not unlike when Artoo saves Luke and friends from the Death Star’s garbage compactor, he thinks you’ll be able to phone your computer remotely:

If I want to talk with my laptop today, I gotta wait until I get back to my room. In the future, if I need my laptop to do something, I just text it with WhatsApp. I say “R2-D2, there’s this thing with the PowerPoint slide, slide number 17, that image is scaled or titled wrong. It should not say CX9, it should say CX10. R2-D2 opens up PowerPoint, modifies it, puts it in PDF, sends it to me. Can you imagine that? Easy.

If we’re talking remote control anyway, why would we want to buy a pricey laptop instead of talking to AI in the cloud? He says it’s partly economic:

You don’t want to necessarily run everything in the cloud, because if you can run it locally, it’s free. Why rent a television? You’re going to use that every day. Why rent a washer dryer, you’re going to use that hopefully once a week? Why rent a refrigerator? You’re going to use it every day. Why rent an assistant computer? You’re going to use it every day.

And partly because your machine is where your private data and tools already exist. Here’s Huang admonishing his friend Dylan Patel (you might remember him from our Vergecast RAM Holiday Spec-tacular😞

What, am I going to call Claude to control my laptop? Are you insane? It doesn’t make any sense! I want to talk to my laptop! I want R2-D2! I want you to do a few things for me when I’m gone. I’m here giving a keynote, and there’s some code I want finished... hey I’ve got an idea, I’m going to call R2-D2, all the files are here in my laptop, all the tools are here, get that done. You can’t do that with Claude Code in the cloud.

Again, Nvidia has so far shown almost nothing to prove that the first generation of RTX Spark laptops will deliver anything close to a Star Trek computing experience. That’s up to Microsoft and software partners; what Nvidia is selling is a small but potent pocket of local AI compute. The Spark has up to 128GB of RAM, which Nvidia says is enough to hold 120-billion-parameter AI agents. Is that enough for R2-D2 to figure out how to do what you want?

Either way, they will cost. When Patel posed a question beginning with “These laptops are $3,000 or something on that order of magnitude, the power user is the one who has to buy the first generation of it,” Nvidia’s CEO nodded along while repeatedly saying “yep.”

Still, as we’ve reported, the RTX Spark will scale down from the first “superchip” with 128GB of RAM down to as little as 16GB of RAM, and Nvidia is looking to produce an entire family of them across multiple generations.

“N2X and N3X are already planned, and N1X is called N1X because it has a smaller version called N1. We’re going to expand our family. We’re going to extend this architecture for a very long time,” Huang told journalists, according to Tom’s Guide.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 4 June 2026 at 8:29 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

Pearl Abyss, the creators of 2026's blockbuster AAA open-world game, Crimson Desert, has recently revealed a new content roadmap on their website, which contains various gameplay and narrative improvements that will be implemented into the game over the Summer.

The roadmap also reveals that Pearl Abyss is working on DLC for Crimson Desert, and it will reveal the details on said DLC at a later date. For now, however, here is the full list of content updates coming for Crimson Desert soon.

• Story Improvements – Pearl Abyss has stated it will improve the narrative flow of Kliff's journey by refining and improving the coherence of key scenes of the story, based on player feedback.
• Re-Blockade Update – The Re-Blockade is getting a new phase that will make the flow before and after a blockage feel more natural while giving the player more ways to defend certain strongholds from invasions. Rewards for liberating strongholds will also be improved.
• New Combat Content – New combat-focused content will be added that will test the player's strength and skills, the details of which will be shared at a later date.
• Cross-Save – Pearl Abyss is working on a cross-save feature so players can continue their progress on Xbox, PlayStation, and PC platforms by linking their Crimson Desert accounts to each platform.
• Damiane & Oongka Gameplay Improvements – The playable Damiane and Oongka will have their gameplay experiences improved so that they're on par with Kliff Greymane's.
• Quality-of-Life Improvements for Non-Combat Content – Non-combat-centric content like trading goods and farming will receive quality-of-life improvements to make more convenient to engage in for players.

Crimson Desert just keeps getting better and better

Many people doubted that a game as massively ambitious and content-rich as Crimson Desert could live up to players' lofty expectations, but it managed to pull it off.

It has managed to sell over 5 million copies and garner critical acclaim for its gargantuan open world rife with engaging activities and puzzles, and a complex and over-the-top combat system where you can mix fantasy swordplay with wrestling moves to decimate enemies.

Our own contributor, Michael Hoglund, even gave Crimson Desert a 4.5/5 review score, explaining that, "Crimson Desert is the kind of game that usually only comes along once in a generation, so singular in its ambition that it will define the gaming conversation for weeks, if not months, to come. Despite granular misses, Pearl Abyss has showcased what gameplay freedom truly means."

With these updates and upcoming DLC on the horizon, Crimson Desert's high quality is going to get even better, possibly to the point where it will become a potential contender for 2026's Game of the Year. But only time will tell if it will make the cut, so stay tuned to find out, along with news for its mysterious DLC.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 4 June 2026 at 8:28 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix

Apple’s annual developers conference is around the corner, and the iPhone maker is expected to unveil its latest software developments across multiple operating systems. To no one’s surprise, artificial intelligence and AI-powered upgrades will play a key role at this year’s WWDC. Much of the Apple software news over the past few months has revolved around AI and how it can enhance apps such as Siri.

What also makes WWDC26 particularly significant is that it will likely be the last time Tim Cook takes the stage as Apple CEO at a developers conference before being succeeded by John Ternus, Senior Vice President of Hardware Engineering. Cook is set to officially step down on September 1, though he will remain with Apple as executive chairman.

When does WWDC26 begin, and how to watch

WWDC26 will kick off on Monday, June 8, at 10 a.m. PT. The event begins with the keynote, where Apple will provide more details about its upcoming operating systems and the AI-driven features we're about to see across its products.

The event’s livestream will be available on Apple’s official website, the Apple TV app, and the Apple YouTube channel.

iOS 27 may be all about an upgraded Siri

Apple is unlikely to make substantial changes to iOS 27’s aesthetics, as the operating system was already revamped with the Liquid Glass interface. However, some minor tweaks could still be on the way. This year’s WWDC, though, is expected to focus largely on Siri and how it can be enhanced with AI.

Bloomberg’s Mark Gurman reported earlier this year that Apple is working on a standalone Siri app that would transform the company’s voice assistant into a chatbot-style service. Gurman added that the upgraded Siri is set to become the “centerpiece of its [Apple] next iPhone, iPad and Mac software updates.” The Siri chatbot app is reportedly powered by Gemini AI models and may also support integration with third-party AI agents.

macOS 27 brings subtle improvements

macOS 27 also received the controversial Liquid Glass design last year, but the new interface did not resonate well with some Mac users, raising concerns about usability. According to Mark Gurman's report, Apple does not plan to ditch Liquid Glass in macOS 27, but it is expected to introduce a “slight redesign” and address issues related to shadows and transparency effects. The operating system is also expected to gain several AI-driven features, including the standalone Siri app.

New features for iPadOS 27 and watchOS 27 may be on the way

iPadOS 27 received a major AI boost last year, including live translation and text summaries powered by Apple Intelligence. This year, significant changes are not expected, although Apple may still have some new AI-powered features in development for iPad users.

As for watchOS 27, the addition of new AI features is anticipated, and Apple may make some of its health-monitoring capabilities more reliant on AI. The new watchOS could also introduce new watch faces and other general improvements.

While there have been rumors about the possible launch of an upgraded Mac mini or even a foldable iPhone at WWDC26, such announcements remain unlikely given the event’s strong focus on software. However, some of the software features unveiled at WWDC26 could help lay the groundwork for future Apple hardware.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 4 June 2026 at 8:27 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of May) 2,092

RIP Matrix