Apple is allergic to nostalgia. In 2008, when the Macintosh was about to turn 25, I mentioned it to Steve Jobs and he instantly shut down the discussion. “If you look backward in this business, you'll be crushed,” he told me icily. “You have to look forward.” Now that Apple’s 50th anniversary looms, however, the company is begrudgingly engaging in a series of concerts and commemorations, and we’re being blitzed by books, articles, and oral histories of the company’s early years.

Rather than join the crowded trek down memory lane, I asked Apple to do what Jobs suggested—look forward. What does Apple want to happen in its next 50 years?

Earlier this month, I sat down with two senior executives to discuss just that. One was Apple’s SVP of worldwide marketing, Greg Joswiak, aka Joz, who joined Apple in 1986. The other was SVP of hardware engineering John Ternus, the putative front-runner to succeed Tim Cook as Apple’s CEO. He’s been with the company for 25 years. I also chatted briefly with Cook himself, just before Alicia Keys sang in front of the Apple Store at Grand Central Station—the beginning of Apple’s reluctantly splashy anniversary celebration.

After acknowledging Apple’s uncharacteristic party mode—“this is too special” to ignore, admits Joswiak—we tackle the future. After launching the personal computer revolution, Apple managed to navigate multiple inflection points. With the Macintosh, it mastered the graphical user interface that makes computers friendlier to use. The iMac positioned the company for the internet boom. And, of course, despite a late start, Apple absolutely owned the mobile era with the iPhone. These products have remained vital–just this month Apple released the buzzy new Macbook Neo, the latest version of a 42-year-old franchise. But now the future belongs to AI—a category where Apple seems to have whiffed so far.

These gentlemen disagree. Apple, they insist, is already at the forefront of the AI revolution. “We were doing AI before we called it AI!” says Joswiak. “Every single great chatbot works great on our products.” Ternus argues that even if Apple didn't take the lead in developing AI technology, it would still benefit. “Our products are the best place people will use the existing AI tools.”

I push them on this. After all, if we're looking decades into the future, shouldn’t we assume that we’ll move past our current computing paradigms and adopt something that specifically caters to the wonders of AI? That’s what Apple’s former design guru Jony Ive seems to be doing with OpenAI. They’re only one entrant in the race to come up with new kinds of hardware devices built specifically for AI. “I would assume you want one of them to be an Apple device, right?” I asked.

The answer seemed to be not necessarily. “Let’s not lose sight of the fact that nothing you just said is incompatible with the iPhone,” Joswiak says. “The iPhone is not going to go away. iPhone is going to serve a very central role in any of those things you’re talking about.”

Wait—Apple thinks that people will be using the iPhone 50 years from now?

“It's hard to imagine not,” says Joswiak. “That's where everybody else struggles. They don't have an iPhone, and so they’re scrambling for what to do. A lot of what they talk about ends up being accessories for an iPhone. We’re not going to get into future road maps, but I will tell you, iPhones are not going anywhere.” (Despite this bravado, I will be shocked if Apple does not come out with some AI-powered gadget in the coming years.)

Later in the day I have my greeting with Cook, and immediately ask him about Apple’s next 50 years. He launches into a rhapsodic description of Apple’s people, values, and culture, predicting that no matter what twists lie ahead, those factors will continue to make Apple unique and super successful. “Yes, the technologies of the future will change,” Cook says. “Yes, there will be more products and more categories. All of those things are true, but the things that made Apple Apple will be the same for the next 50 years, and the next 100 and the next 1,000.”

That of course presumes that superintelligence doesn’t totally rearrange reality in the next 50 years, let alone the next millennium. It also flies in the face of what the leaders of AI companies believe. OpenAI’s Sam Altman has even postulated that his own successor as CEO will not be a human but an AI model! Does Cook see that as a possibility for Apple anytime in the next 50 years?

Cook laughed merrily at the idea. “When you look at the leadership page,” he says of future Apple, “there will not be an agentic kind of model on there.” Left unspoken is what the people of 2076 will be using to look up that page.

This is an edition of Steven Levy’s Backchannel newsletter. Read previous newsletters here.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Saturday 28 March 2026 at 5:57 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

As many expected, GSC Game World announced an expansion for S.T.A.L.K.E.R. 2: Heart of Chornobyl at the Xbox Partner Preview event today. Dubbed Cost of Hope, the studio gave the first details of this story expansion today, which is releasing later this year.

The studio is saying the expansion will be pretty massive, taking "dozens of hours" to complete the storyline, which will take place across brand-new regions. It promises more engaging quests and characters and deeper answers to age-old lore questions.

"Cost of Hope will allow you to explore two long-awaited regions that have been cut off for years and that every stalker in the Zone has dreamt of reaching: Iron Forest and the Chornobyl Nuclear Power Plant," describes the studio. "A region isn’t just a single location. It’s a large territory, each with its own hub, along with dozens of smaller locations to discover — all filled with quests and activities."

Check out the brand-new gameplay trailer showcasing the story expansion above.

The studio also revealed that it is planning a trilogy surrounding S.T.A.L.K.E.R. 2, much like the original Legends of the Zone Trilogy. The Cost of Hope expansion is described as being the middle chapter of this “second trilogy," with another expansion also in the works to expand the story.

"The protagonist of the expansion is Skif, and the events unfold in parallel with the story of Heart of Chornobyl," adds GSC GameWorld. "After installing the DLC, during your playthrough of the main campaign you’ll receive a signal on your PDA that will start the expansion’s storyline."

S.T.A.L.K.E.R. 2: Cost of Hope expansion is releasing in the summer of 2026 across PC, Xbox Series X|S, and PlayStation 5. Both the base game and the DLC will be Xbox Play Anywhere supported.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 27 March 2026 at 12:22 pm AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

The most expensive Mac you could get is no longer available on Apple’s website, as 9to5Mac reports that the “cheese grater” Mac Pro workstation has been discontinued. It had been a part of the lineup since replacing the Power Mac G5 in 2006 when Apple shifted from Arm to Intel processors, but it’s had the feeling of a vestigial limb ever since Apple dropped Intel for its own M-series chips.

The M2 Ultra edition of the Mac Pro launched three years ago, with Apple hardware exec and possible future CEO John Ternus saying that “for those users who need the versatility of internal expansion, Mac Pro combines PCIe slots with our most powerful chip.” But without support for adding a new GPU, many of the professionals who should’ve been in its target market told us that they didn’t want one, with or without optional $699 wheels.

The Mac Studio offered similar options without those slots for $3,000 less at the time. Now it stands as the most powerful machine in Apple’s lineup, with M4 Max and M3 Ultra-powered options available now and, likely, M5 Ultra variants on the way.

Apple has acknowledged that 2013’s “trash can” Mac Pro design put it into a thermal corner it couldn’t find a way out of, as large single-GPU cards became preferred over multiple GPUs. Pro users wanted more options, with Red Digital Cinema president Jarred Land stopping during a talk in 2018 to tell an Apple employee that “We need Nvidia support… it’s not good for the community to just be on one platform so please just think about it, please.” Apple addressed those needs at the time by going back to the tower, but now, it’s leaving the Mac Studio and Thunderbolt 5 as its primary option for customers who need to add something extra.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 27 March 2026 at 12:21 pm AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

Text generated by artificial intelligence often has a particular vibe that gives it away as machine-generated, but it has become harder to pick out those idiosyncrasies as the tech has improved. We may be seeing a similar evolution of generative AI audio. Google has announced a new AI audio model called Gemini 3.1 Flash Live—as the name implies, it’s designed for real-time conversation. It’s rolling out in some Google products starting today, and developers will be able to start building their own chatty robots with the model, too.

Google says this AI is much faster and produces speech with a more natural cadence, aiming to solve a long-running issue with AI-generated speech. Like a chatbot, there’s always a delay between input and output in generative audio systems. Longer delays and unnatural inflection make conversations feel sluggish and harder to follow. Researchers generally believe 300 milliseconds of latency is about the limit for optimal speech perception, but Google has not specified any particular delay for Gemini 3.1 Flash Live. It just vaguely has the speed you need.

But benchmark numbers? Google has plenty of those, which it claims show that 3.1 Flash Live will be a more reliable way to have audio-to-audio AI conversations. For example, a big gain in the ComplexFuncBench Audio shows the new model is better at complex, multi-step tasks. Gemini 3.1 Flash Live also tops the charts in the Big Bench Audio test, which evaluates reasoning with a set of 1,000 audio questions.

Meanwhile, a strong showing in Scale AI’s Audio MultiChallenge means the new Gemini model is more able to cope with hesitation and interruptions in the audio input. Although it outpaces other real-time audio models, Gemini 3.1 Flash Live only manages 36.1 percent in this test. Audio models that are not designed to operate conversationally can reach scores over 50 percent in the MultiChallenge.

Credit: Google

The upshot is that Gemini 3.1 Flash Live should sound more like a person, to the point that Google felt it was time to integrate AI flags. The outputs from this model will have SynthID watermarks, which are not perceptible to human listeners. However, they can be detected if someone were to try to pass off Gemini AI speech as the real deal.

Google has partnered with companies like Home Depot, Verizon, and others to test the model. They all have glowing reports in the blog post on how well 3.1 Flash Live can mimic human speech. So the next AI assistant you encounter on a phone call might sound much more realistic. Maybe you’ll even think you’re talking to a person, and SynthID can’t help with that.

Developers can now access the model in AI Studio, the Gemini API, and Gemini Enterprise for Customer Experience. The latter is essentially a toolkit for agentic shopping. Gemini 3.1 Flash Live will be seen most prominently in Gemini Live and Search Live (a feature of AI Mode). The new conversational AI is rolling out in those products starting today.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 27 March 2026 at 12:19 pm AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

AMD has officially revealed its new Ryzen 9 9950X3D2 Dual Edition desktop processor, and you'll be able to get your hands on it on April 22.

This is AMD's (and indeed the world's) first desktop processor where both chiplets are equipped with the proprietary 3D V-Cache technology that is so coveted by PC gamers. With 104MB of cache underlying each chiplet, it offers a whopping 208MB total on-chip cache.

What does that mean for you? Well, with more cache, the CPU has more space to store the data it needs to perform as efficiently as possible. As AMD's Senior VP and GM of Computing and Graphics, Jack Huynh, explains in the announcement video, "208MB of cache means more game data, more assets, and more working data sitting right next to the CPU cores."

The Ryzen 9 9950X3D2 Dual Edition remains a 16-core processor based on the Zen 5 architecture and using the AM5 socket, just like its Ryzen 9 9950X3D sibling that we reviewed favorably here at Windows Central. However, it has jumped up to a 200W TDP compared to the 170W TDP of its sibling.

Last month, Google introduced Lyria 3, its latest generative music model that can create 30-second tracks with custom cover art generated by Nano Banana. It can create tracks even without user-provided lyrics, as the lyrics are generated based on user prompts. Lyria also allows users to control track elements such as style, vocals, and tempo.

Today, Google introduced Lyria 3 Pro, its most advanced music generation model. It can create tracks up to 3 minutes long, offering users more creative control and customization options. For example, users can now provide prompts with details on elements like intros, verses, choruses, and bridges. As a result, Google claims that Lyria 3 Pro can create more realistic and musically complex tracks.

The Lyria 3 Pro model is now available in the following Google products and services:

Vertex AI: Lyria 3 Pro is available as an API in public preview on Vertex AI, allowing large organizations to create soundtracks at scale for gaming and integration into different music and video platforms.

Google AI Studio and the Gemini API: Developers creating creative tools can now access Lyria 3 Pro APIs alongside Lyria RealTime in AI Studio.

Google Vids: Along with Lyria 3, Lyria 3 Pro is now available in Vids, Google's enterprise video creation service. These new models will allow Vids users to add custom music to match the videos they are creating. Lyria 3 Pro is rolling out to Google Workspace customers and Google AI Pro and Ultra subscribers starting this week.

Gemini app: Lyria 3 Pro is now available in the Gemini app for AI Pro and Ultra subscribers, allowing them to create longer tracks with improved customization and control.

ProducerAI: ProducerAI is a collaborative music creation tool targeting musicians. Lyria 3 Pro is now available inside ProducerAI for free and paid subscribers to help artists, producers, and songwriters iterate on comprehensive songs.

Similar to AI-generated images and videos, all Lyria 3 and Lyria 3 Pro musical outputs are embedded with a SynthID watermark for identifying Google AI-generated content.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 26 March 2026 at 5:40 pm AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

NVIDIA's latest upcoming version of its Super Resolution software DLSS 5 was announced last week, and the reveal ended up being one of the hottest stories in gaming for all the wrong reasons. The graphics and AI-focused company showed how the tech uses generative AI in an effort to enhance in-game scenes with photorealistic lighting.

What it was hoping would be a slam dunk that got people hyped for its release later this fall, however, was instead something of a disaster. Gamers and devs alike were shocked by how DLSS 5 made characters look like they were passed over by an uncanny AI slop filter, and NVIDIA CEO Jensen Huang only made things worse by telling consumers "they're completely wrong." He later backpedaled.

The overall backlash against DLSS 5 has been enormous, and despite Huang's clarifications that it doesn't touch model geometry or textures and can be fully controlled and finely tuned by developers, some remain skeptical at best and doubt the truthfulness of those statements at worst.

Others, though, believe that while DLSS 5 had a poor first showing, the tech is still very much going to catch on and become popular, especially as developers adopt it and tweak it to achieve the visual enhancements that they want. Among them is Kingdom Come: Deliverance 2 director Daniel Vávra, who's spoken up in defense of the software.

"I can imagine in the future devs will be able to train this tech for particular art style or specific people faces and it might replace expensive raytracing etc. This is just a little uncanny beginning. No way haters will stop this. Its way more than a soap opera effect every tv has when you turn motion smoothing on," he said in a post on social media.

OpenAI’s recently announced plans to shutter its Sora video-generating app have also scuttled the company’s planned $1 billion licensing partnership with Disney, according to multiple press reports.

“As the nascent AI field advances rapidly, we respect OpenAI’s decision to exit the video generation business and to shift its priorities elsewhere,” Disney said in a statement provided to media outlets. “We appreciate the constructive collaboration between our teams and what we learned from it, and we will continue to engage with AI platforms to find new ways to meet fans where they are while responsibly embracing new technologies that respect IP and the rights of creators.”

Disney and OpenAI announced the blockbuster three-year licensing deal in December, saying that over 200 Disney-owned characters would be available for use in Sora-generated videos. At the same time, Disney said it would be making a $1 billion equity investment in the AI company.

While that agreement was presented publicly as all but a done deal, an OpenAI statement at the time said it was “subject to the negotiation of definitive agreements, required corporate and board approvals, and customary closing conditions.”

Now, Axios cites “a source familiar with the situation” in reporting that no money ever changed hands in the planned deal. The Financial Times reports that the deal never got off the ground as OpenAI shifted its strategic direction, according to “two people familiar with the matter.” And Deadline cites a Disney insider who said point-blank that “the deal is not moving forward.”

Reuters cites an anonymous “person familiar with the matter” who described the Sora shuttering as “a big rug-pull” that blindsided the massive entertainment conglomerate. OpenAI published an updated version of Sora’s safety standards as recently as Monday, suggesting that many inside the company were not prepared for the late Tuesday announcement.

That said, Reuters’ source also suggested that Disney and OpenAI were still discussing whether there was another way the companies could partner with or invest in each other.

No more shockwaves

The December announcement of the planned Disney/OpenAI partnership sent shockwaves through Hollywood, with many publicly worrying about what it meant for the future of physical actors and human-created cinematic content. By January, Disney CEO Bob Iger was talking about Sora-generated content appearing on Disney+ as part of an effort to make the service a destination for daily short form video.

“The demand for Disney characters in particular from our users is sort of off the charts,” OpenAI CEO Sam Altman told CNBC in December.

In the months since, though, Hollywood’s worries and attention have largely shifted from Sora to upstart AI video apps like SeeDance 2.0, which went viral with detailed videos of familiar characters in Hollywood-style scenes, complete with realistic cuts and camera angles. Disney was among the companies that sent a cease-and-desist letter to SeeDance-maker ByteDance last month, calling the app a “virtual smash-and-grab of Disney’s IP [that] is willful, pervasive, and totally unacceptable.” Disney has also made legal threats to Google and other companies it says trained on its copyrighted works without permission.

When the Sora 2 model launched in October, OpenAI initially asked copyright holders to actively opt out of having their works used as the basis for generated videos. Following a public outcry, OpenAI quickly changed direction and instead asked IP owners to opt in to work with Sora, with vague promises of future profit-sharing.

Sora quickly became a hit on mobile platforms following its October launch as a standalone app. But the outsize attention from users was short-lived—estimates from Appfigures Intelligence provided to Ars Technica suggest the app peaked at about 3.3 million downloads across iOS and Google Play in November before falling to just 1.1 million downloads in February.

Across those months, Appfigures Intelligence estimates Sora grossed just $2.14 million in revenue from 11.7 million downloads. That’s a drop in the bucket for a company the size of OpenAI, especially when you consider the massive costs associated with generating AI video.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 26 March 2026 at 7:56 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

Following the launch of its Core Ultra 200S Plus desktop processors, which we reviewed this week, Intel, today at its Pro Day 2026, has now announced its "most advanced commercial client portfolio", headlined by the new Intel Core Ultra Series 3 with Intel vPro tech. This is essentially Panther Lake mobile for enterprises.

vPro, if you are not familiar, is a platform designed for businesses to deliver performance, security, and remote management features required for enterprise computing. vPro brings technologies like Intel Active Management Technology to Intel CPUs, which allows IT teams to remotely access and repair devices even if the OS is down or when the machine is powered off. Thus vPro essentially helps organizations and enterprises protect sensitive data, reduce downtime and streamline device maintenance.

While it is already available elsewhere, today's launch marks the first commercial PC platform built on Intel 18A lithography, its next-gen RibbonFET process node. Intel promises this architecture redefines managed PCs by enabling "seamless deployment, proactive security, and intelligent fleet operations."

In terms of what's on offer, Intel put some rough performance figures about what enterprise users can expect compared to four-year-old systems, as it believes that this is the typical refresh cycle for the majority of enterprises.

The biggest improvement will be seen in AI processing and that makes sense considering these new chips will pack Gen 5 NPU. Integrated graphics also sees substantial gains with the new Xe3 architecture which is inside the new Intel Arc Pro B390 graphics.

The company promises:

• Over 30% faster single and multi-thread performance
• Up to 30% higher productivity
• Up to 80% better graphics
• Up to 4x AI performance

Alongside the Core Ultra Series 3, Intel unveiled the Arc Pro B70 and B65 discrete graphics cards, targeting professionals who need high-end visualization and compute performance. The company also confirmed retail availability of its Intel Xeon 600 processors for workstations today.

Meanwhile, commercial PC offerings powered by the new Intel Core Ultra Series 3 and Intel vPro will begin shipping on March 31, 2026, and the company expects over 125 designs spanning major OEM partners.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 26 March 2026 at 5:05 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

On its Pro Day 2026, alongside new Core Ultra Series 3 vPro CPUs, Intel has introduced two new discrete graphics cards for its professional portfolio, the Arc Pro B70 and Arc Pro B65. These new GPUs are built for content creation, engineering, and, as expected, AI inferencing workloads. The company says the GPUs are optimized with a fully validated, open software stack such that the AI performance should scale from single to multi‑GPU deployments.

Under the hood, the Arc Pro B70 and B65 are built on Xe2 design and offer up to 32 Xe cores and 32GB of VRAM; both cards, Intel notes, are tuned for multi-user and multi-agent AI workloads. The company highlights the strong price-to-performance ratios for inferencing throughput across workstation and edge deployments.

According to Intel, the new Arc Pro B-series performance-wise delivers:

• Up to 2.2x larger context windows with the Intel Arc Pro B70 versus the competition.
• Up to 6.2x faster responses in multi-agent/multi-user workloads with the Intel Arc Pro B70 versus the competition.
• Up to 2x tokens per dollar performance with the Intel Arc Pro B70 versus the competition.

If you are wondering what kind of competition Intel is talking about, the company has compared its GPUs head-to-head with Nvidia's RTX Pro 4000.

In terms of availability, the Arc Pro B70 will be available starting March 25, 2026, both as an Intel-branded card and also through add-in board (AIB) partners like ARKN, ASRock, Gunnir, Maxsun, and Sparkle.

In terms of pricing, the Intel-branded B70 will start at $949, while partner pricing will vary by configuration. The Arc Pro B65 is scheduled for release in mid-April 2026 though there will be no Intel-branded variant for the smaller sibling.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 26 March 2026 at 5:03 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

Gamers Nexus (2.6m subscribers)

March 25, 2026

Video length: 34m 09s

Intel's Ultra 7 270K Plus (270KP) is one of the most promising CPU launches we've seen in a long time, and has some similarities to the Ryzen 1000 and 2000 era of CPUs. The key difference is that Intel's platform is a dead-end, unfortunately, but from a performance standpoint, Intel is showing a lot of positive similarities. In this testing, we benchmark the Intel 270K Plus vs. the 250K Plus, AMD's X3D CPUs (including the 7800X3D, 5800X3D, and 9800X3D), and Intel's predecessors, like the 14700K. Intel's 270K Plus does particularly well in workloads like video editing in DaVinci Resolve and in our code compile test.

00:00 - Intel 270K Plus CPU Review and Benchmarks

03:55 - Intel, Please Listen

06:12 - Intel Platform Performance Package

07:29 - Intel 270K Plus Specs

08:08 - CPU Pricing Update

09:05 - Frequency: All-Core 270K Plus vs. 250KP

09:53 - Frequency: Single-Thread Boost Intel 270K Plus

11:14 - Baldur's Gate 3 Best Gaming CPUs 2026

12:38 - Outer Worlds 2 Intel 270K Plus Benchmark

13:38 - Outer Worlds 2 1440p Benchmarks

14:04 - Stellaris CPU Simulation Time Comparison

15:15 - Kingdom Come Deliverance 2 Intel vs AMD CPUs

16:02 - Kingdom Come Deliverance 2 1440p

16:26 - Dragon's Dogma 2 Intel 270K and 250K Benchmarks

17:35 - Cyberpunk 2077 Phantom Liberty 1080p Medium

18:25 - Cyberpunk 2077 Phantom Liberty 1080p High

19:28 - F1 25 270K vs 7800X3D, 5800X3D

20:24 - Starfield CPU Benchmarks

21:11 - Chromium Code Compile Best CPUs for Programmers

22:42 - Blender CPU Rendering Benchmarks

24:14 - 7-Zip File Compression Best CPUs

25:23 - 7-Zip File Decompression Comparison

26:34 - Adobe Photoshop Best CPUs

27:10 - DaVinci Resolve Video Editing CPU Benchmarks

27:55 - Power Consumption - Blender All-Core

28:57 - Power Consumption - F1 25

29:18 - Conclusion

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Wednesday 25 March 2026 at 5:15 pm AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

On a Monday episode of the Lex Fridman podcast, Nvidia CEO Jensen Huang made a hot-button statement: “I think we’ve achieved AGI.”

AGI, or artificial general intelligence, is a vaguely defined term that has incited a lot of discussion by tech CEOs, tech workers, and the general public in recent years, as it typically denotes AI that’s equal to or surpasses human intelligence. In recent months, tech leaders have tried to distance themselves from the term and create their own terminology that they view as less over-hyped, more useful, and more clearly defined (although the new phrases they’ve come up with essentially mean the same thing as AGI). The term has also been the subject of key clauses in big-ticket contracts between companies like OpenAI and Microsoft, upon which a significant amount of money may hinge.

Fridman, the podcast’s host, defines AGI as an AI system that’s able to “essentially do your job,” as in start, grow, and run a successful tech company worth more than $1 billion. He then asks Huang when he believes AGI will be real — asking if it’s, say, five, 10, 15, or 20 years away — and Huang responds, “I think it’s now. I think we’ve achieved AGI.”

Fridman says, “You’re gonna get a lot of people excited with that statement.” Huang goes on to mention OpenClaw, the open-source AI agent platform, and its viral success. He said that people are using their individual AI agents to do all sorts of things, and that he “wouldn’t be surprised if some social thing happened or somebody created a digital influencer … or some social application that, you know, feeds your little Tamagotchi or something like that, and it become out of the blue an instant success.”

But Huang then seemed to slightly walk back his earlier claims, saying, “A lot of people use it for a couple of months and it kind of dies away. Now, the odds of 100,000 of those agents building Nvidia is zero percent.”

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Wednesday 25 March 2026 at 1:18 pm AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

In an interview with ANI, Chandak noted that the additional capacity will stem from several semiconductor projects that are scheduled to begin operations in phases. Once these facilities become operational, India’s chip assembly and testing capabilities are expected to expand substantially.

Chandak added that while part of the output will meet domestic demand, a considerable share is expected to be exported. As these facilities ramp up, India’s position within the global semiconductor value chain is also likely to evolve.

At present, India’s semiconductor push is primarily focused on chip assembly and testing rather than wafer fabrication. For instance, the recently launched Micron Technology facility operates as an ATMP (Assembly, Testing, Marking and Packaging) plant and has also been described as a smart packaging unit.

Other projects under development—including those led by Tata Electronics, Kaynes Technology, and CG Power and Industrial Solutions—are being built as OSAT (Outsourced Semiconductor Assembly and Test) facilities to perform similar assembly and testing functions.

According to Chandak, the Micron facility will handle memory products such as DRAM, NAND, and SSD, which are widely used across multiple industries. As AI workloads continue to grow, demand for memory devices is also on the increase, while sectors such as smartphones, laptops, and automotive electronics continue to face supply constraints. Chips assembled and tested in India are expected to support AI systems as well as applications in automobiles, laptops, and smartphones.

Initially, these chips will mainly fall within the 14nm to 28nm technology nodes, while the wafers themselves will continue to be sourced from overseas.

Among the upcoming projects, Kaynes Technology plans to assemble advanced power modules, including IGBTs and other power components, while also manufacturing printed circuit boards (PCBs) for electronic systems. These components will serve sectors such as automotive, industrial, consumer electronics, and defense.

Meanwhile, Tata Electronics is preparing to operate an OSAT facility in Jagiroad, Assam, focused primarily on power devices and multi-chip modules for industrial and automotive applications. Chandak indicated that the plant could achieve a capacity of more than 50 million units per day.

The semiconductor project led by CG Power and Industrial Solutions will focus on integrated circuits for industrial and automotive applications. The initiative will be implemented in two phases and could ultimately reach a daily capacity of approximately 15 million units.

(Photo credit: FREEPIK)

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Wednesday 25 March 2026 at 1:11 pm AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

Arm, one of the world’s leading chip design firms, announced Tuesday that it is producing its own semiconductors. The move is a departure from its long-standing model of licensing intellectual property to companies that manufacture and sell chips themselves. Speaking to a live audience in San Francisco, Arm CEO Rene Haas made his pitch for how the new Arm CPU could benefit the tech industry and why this is the right time for the company to step outside of its lane and go head-to-head with other chipmakers.

“Let me be clear: We are now in a new business for ARM, and we are supplying CPUs,” Haas said, holding up one of the company's new chips. Arm’s primary reason for moving in this direction, Haas said, is demand from customers.. But as artificial intelligence proliferates throughout the economy and demand for computing resources skyrockets, Arm is also trying to capture a sliver of the growing AI CPU market.

Arm’s in-house chip efforts had long been rumored, but now the company is finally offering a clearer picture of what it’s doing. The new chip is called the Arm AGI CPU, a nod to artificial general intelligence, an often-invoked but still hypothetical form of AI that could match human performance across domains. It’s designed to be coupled with other chips in high-performance servers inside data centers and to handle agentic AI tasks. The chip is being fabricated by Taiwan Semiconductor Manufacturing Corporation, the world’s leading semiconductor foundry, and is being built using TSMC’s 3nm process.

At the chip reveal event, Arm executives emphasized the company’s history of designing energy-efficient chips and claimed that its new AGI CPU will be the world’s “most efficient agentic CPU on the market.” Compared to competitors like the latest x86 chips made by Intel and AMD, Arm says this chip will deliver better performance per watt, or the amount of energy a computer uses to operate, and could save customers billions of dollars in electricity spending.

The first major customer of Arm’s new chip is Meta, which the company says has received samples of the CPU. OpenAI, SAP, Cerebras, and Cloudflare, as well as the Korean tech firms SK Telecom and Rebellions, have also agreed to buy the chip. Arm projects its AGI CPU will reach “full production availability” in the second half of this year.

Santosh Janardhan, Meta’s head of infrastructure, appeared on stage and said he thought the Arm chip was going to “expand the [chip] industry on multiple axes.” As Meta pushes toward “personal superintelligence"—AI that will make its apps deeply personalized—Janardhan said the company needs more silicon, and is especially interested in power efficiency.

OpenAI’s vice president of science and former chief product officer, Kevin Weil, also showed up on stage alongside Haas. “One of the most common things I hear inside of OpenAI: ‘I need more compute,’” Weil said. “It’s kind of the coin of the realm.”

Nvidia CEO Jensen Huang, Amazon senior vice president and distinguished engineer James Hamilton, and Google AI infrastructure chief Amin Vahdat appeared in pretaped video testimonials praising Arm’s new hardware. None committed to buying it, but all three tech giants already use Arm’s designs in their own processors.

Arm’s history traces back to the late 1970s, when it was known as Acorn and produced microprocessors. In the 1990s the entity changed its name to ARM (Advanced RISC Machines) and its then-CEO began licensing the firm’s chip designs to other companies. Arm, which has since dropped the all-caps “ARM” branding, saw its business boom during the mobile revolution. By the 2010s many of the world’s largest tech companies, including Apple, Nvidia, Microsoft, Amazon, Samsung, and Tesla, were all relying on its technology.

Arm appeared eager at the press event to demonstrate it has support from bold-faced names in the tech industry. While the company is mostly taking aim at chipmakers like AMD and Intel, which build CPUs based on a different architecture, it risks potentially alienating some of its longtime partners by releasing its own chip. Nvidia, which primarily makes GPUs, also bundles Arm-based CPUs into its rack systems. Earlier this year, Nvidia said it would sell stand-alone CPUs for the first time. Meta was one of its first buyers.

Ben Bajarin, CEO and principal analyst at the research firm Creative Strategies, says that Arm could be perceived more as a competitor than partner as its strategy evolves. Right now, Arm is launching a streamlined CPU with a relatively small number of cores—the chip’s built-in processing units—designed specifically for running AI agents, Bajarin points out. Over time, Arm may expand into more general-purpose CPUs, while AMD and Intel develop chips tailored for agentic AI. That would put the companies in more direct competition with one another.

But at the moment, Bajarin says, Arm is simply trying to get in on the fast-growing market for data center CPUs. Creative Strategies forecasts that demand for such chips will grow from $25 billion this year to $60 billion globally by 2030. That figure represents only CPUs for traditional cloud computing data centers. When CPUs for agentic AI are factored in, Bajarin’s demand estimate jumps closer to $100 billion by 2030. Even if Arm captures only a sliver of that, it could be a significant source of revenue for the company.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Wednesday 25 March 2026 at 5:27 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

Anthropic is joining the increasingly crowded field of companies with AI agents that can take direct control of your local computer desktop. The company has announced that Claude Code (and its more casual user-oriented Claude Cowork) can now “point, click, and navigate what’s on your screen” to “open files, use the browser, and run dev tools automatically” when necessary to complete tasks.

When possible, Anthropic says Claude Code and Cowork will still prioritize using Connectors to directly access and control outside apps or data sources. When that connection isn’t available, though, those tools are now able to ask permission to “scroll, click to open, and explore as needed” on the machine itself to do what’s asked. This kind of direct control of the computer can also be initiated and managed remotely via Claude’s Dispatch tool as long as the target computer remains powered on.

An Anthropic video shows some examples of tasks Claud Code can complete on your desktop via Dispatch.

The new feature is now available to Claude Pro and Max subscribers using MacOS in what Anthropic calls a “research preview.” That means the system “won’t always work perfectly” and will sometimes require a “second try” for complex tasks, Anthropic warns. Completing tasks via “computer use” also “takes much longer and is more error-prone” than performing the same task via Connectors, the company writes.

Pobody’s nerfect

Giving an admittedly imperfect and “error-prone” AI tool the ability to explore your computer desktop “as needed” could ring some justified security alarm bells for many users. That’s especially true given the widespread stories of security issues that have arisen when companies or individual users give AI agents access to sensitive resources.

Anthropic says it has safeguards in place to prevent common risks like prompt injection, and it will limit access to certain “off limits” apps (e.g., “investment and trading platforms, cryptocurrency”) by default.

Anthropic also notes on a support page that the model is trained to avoid “risky operations” such as moving or investing money, modifying files, scraping facial images, or inputting “sensitive data.” But the company also warns that such training safeguards “aren’t perfect” and “aren’t absolute,” meaning that “Claude may occasionally act outside these boundaries.”

Anthropic also warns that, when computer use is activated, Claude will be able to see anything visible on-screen, including “personal data, sensitive documents, or private information.” For all these reasons, the company recommends “starting with the apps you trust and not working with sensitive data” during this research preview stage.

Anthropic’s announcement comes just weeks after the rollout of Perplexity’s Personal Computer, Manus’s My Computer, and Nvidia’s NemoClaw, which similarly let their AI agents take direct control of the desktop. All of these corporate moves follow the viral spread of OpenClaw earlier in the year, which led OpenAI to hire OpenClaw creator Peter SteinBerger “to drive the next generation of personal agents.”

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Wednesday 25 March 2026 at 5:25 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

Last week, Nvidia’s public reveal of DLSS 5—and its “generative AI” enhanced glow-ups of gaming scenes—drew widespread condemnation from the gaming community. In a podcast published Monday, though, Nvidia CEO Jensen Huang tried to differentiate the technology’s optional, artist-guided graphical enhancements from the “AI slop” that Huang says he’s not a fan of.

As part of a nearly two-hour-long interview with the Lex Fridman Podcast, Huang was asked to explain the “drama” around DLSS 5 and “the gamers online [that] were concerned that it makes games look like AI slop.” Huang responded that he “could see where they’re coming from, because I don’t love AI slop myself… all of the AI-generated content increasingly looks similar and they’re all beautiful, so… I’m empathetic towards what they’re thinking.”

At the same time, Huang said DLSS 5 is decidedly separate that kind of “slop,” because it “is 3D conditioned, 3D guided.” The artists behind a game are still the ones creating the in-game structural geometry and textures that form the “ground truth structure” that DLSS 5 works from, Huang said. “And so every single frame, it enhances but it doesn’t change anything,” he said.

For the most part, though, gamers haven’t been worried about DLSS 5 creating trippy new content from the ground up like some generative AI world models. Instead, the worry is that DLSS 5’s visual “enhancements” could end up smoothing out many disparate games toward a single, flattened, homogenized photo-realism standard.

That’s a misunderstanding of how DLSS 5 works, Huang said. It’s not a technology where a game ships in one state and “then we’re gonna post-process it,” he said. Instead, DLSS 5 “is integrated with the artist, and so it’s about giving the artist the tool of AI, the tool of generative AI.”

Because DLSS 5 is “open,” Huang said artists can train the model for the specific kind of look they want. In the future, Huang said artists will also be able to prompt DLSS 5 with examples or a description of a desired look—“I want it to be a toon shader,” for instance. And if visual artists want to use DLSS 5’s models “to generate the opposite of photoreal, yeah, it’ll do that too,” he said.

The interview follows similar comments Huang gave in an interview with Tom’s Hardware last week, when he said “it’s not post-processing at the frame level, it’s generative control at the geometry level.”

But any “confusion” on this point from gamers is entirely understandable since previous versions of DLSS were explicitly sold as relatively turnkey post-processing to improve resolutions and/or frame rates by generating new frames that look like the ones rendered by the game itself. If Nvidia wanted to introduce a new artist-facing tool for using generative AI to create customizable shader-style effects, they could have done so without overturning the existing and well-known meaning and branding of DLSS.

Elsewhere in the new podcast, Huang threw in an aside that if artists don’t like the look of DLSS 5’s enhancements, “they could decide not to use it, you know?” Whether or not individual artists want to use DLSS 5, though, Nvidia’s announced partnerships with major publishers including Bethesda, Capcom, NetEase, NCSoft, Tencent, Ubisoft, and Warner Bros. Games. suggest that the companies behind many big-budget games will be pushing the technology’s use in their projects at least for the time being. And while gamers will also be able to disable DLSS 5’s “enhancements” if they don’t like the way they look, that’s not exactly the affirmative defense of the technology that generative AI boosters may have been hoping for out of Nvidia.

We still have months to go until Nvidia’s planned debut of DLSS 5 in any actual games. We expect Huang and others at Nvidia will be spending a good deal of that time trying to explain and justify the new technology to a skeptical gaming public.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 24 March 2026 at 10:44 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

LG Display is mass-producing laptop screens that automatically change their refresh rate from 1 Hz to up to 120 Hz, based on what’s on-screen, it announced this week. The display supplier said that it’s the first company to mass-produce these 1–120 Hz screens, which are supposed to boost battery life.

According to LG’s announcement, the LCD screens, which it’s calling Oxide 1Hz, will automatically use a 1 Hz refresh rate when detecting a static image on-screen and switch to up to 120 Hz when needed. Without providing more detail, LG said it created proprietary “circuit algorithms and panel design technology” and discovered “new materials and [applies] the oxide with the lowest power leakage during low-refresh-rate mode to the display’s thin-film transistor.”

In its announcement this week, LG said that “when performing tasks involving primarily still images—such as checking emails or reading e-books and research papers—the panel operates at the lowest refresh rate of 1 Hz. Conversely, it runs in high-refresh-rate mode at up to 120 Hz when streaming content such as movies or sports as well as playing games with frequent screen changes.

Lower-refresh-rate screens generally consume less battery power than equally specced faster ones, and LG claims its so-called “Oxide 1Hz” screens lead to “48 percent more use on a single charge compared to existing solutions.” Although, what users actually experience will depend on how they use their computer and screen.

In November, BOE and Intel announced a similar 1 Hz laptop display technology that works in conjunction with Windows and Intel GPUs to automatically change refresh rates. However, the companies haven’t said when the displays might be available in devices.

Some OLED smartphones and smartwatches have been using a display tech that’s similar to Oxide 1Hz, called LTPO (low-temperature polycrystalline oxide), since 2018. LG’s so-called Oxide 1Hz brings the capability to LCD displays and laptops.

The concept behind Oxide 1Hz is also similar to dual-mode displays that have existed in gaming laptops and monitors since 2024 and that LG Display produces. However, those displays differ from Oxide 1Hz in that they require the user to press a button to quickly switch from a high refresh rate and low resolution to a lower refresh rate and higher resolution and are targeted at gaming.

Dell’s 2026 XPS laptops currently offer Oxide 1Hz displays as a base option.

LG said that it will mass-produce an OLED version of Oxide 1Hz starting in 2027.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 24 March 2026 at 10:43 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

If you're still recovering from last week's DLSS 5 nightmare and Crimson Desert's gen-AI fiasco, we're sorry to re-open old wounds by telling you that Square Enix has announced it's teaming up with Google to put Gemini AI into Dragon Quest X.

The tenth mainline Dragon Quest is a Japan-exclusive MMORPG released in 2012 that never escaped the country despite its 2022 offline conversion, which probably explains why Square Enix will be conducting this experiment there and not inside a heavyweight champion like Final Fantasy 14.

VGC shared the news on 23rd March, reporting on a press event which took place this past weekend. Japanese publication Sankei was able to speak to head of development Takashi Anzai, who pitched the "Chatty Slimey" companion as a way to make new players feel "less lonely" as they begin their journey, which is a baffling thing to say about an MMO that is populated with real-life players that, usually, are very keen to welcome new players into the fold.

"When a player talks to Slimey using the chat function, it automatically generates voices and engages in conversation. The AI also analyses information from the game screen, and Slimey may even initiate conversations when a powerful enemy is defeated or a rare item is obtained," Sankei's report added. How Square Enix and the minds behind Gemini plan to limit and "guide" what the AI can and cannot say remains uncertain, and specific details regarding exactly how this will all work have yet to be shared.

As pointed out by VGC, Dragon Quest X won't be the first online game trying to make use of AI chat bots. Last year, Where Winds Meet released with a world full of AI chatbot NPCs, so you can imagine how things went after some players abused them. Similarly, Darth Vader went wild on Fortnite last year before Epic Games had to patch the AI-powered NPC into something less IP-threatening.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 24 March 2026 at 6:29 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

Crimson Desert has had a turbulent reception, with the game proving divisive among players. Here at Windows Central, however, Michael reviewed Crimson Desert and awarded it a 4.5 out of 5, describing it as a genre-defining experience.

That said, players eager to jump in on Intel Arc GPUs were left out. Arc users quickly found that the game simply didn’t work with their GPUs. Intel also stated it had reached out to Pearl Abyss to assist with development and help ensure the game runs properly on Arc hardware; however, were ignored by Pearl Abyss.

In a new post on X (formerly Twitter), the Crimson Desert account addressed the Intel Arc issues directly, stating:

“Regarding #CrimsonDesert support for Intel Arc:

We are currently working on compatibility and optimization support so that Crimson Desert can also be enjoyed on Intel Arc GPU systems. We are preparing to provide a smooth and stable gameplay experience, and we ask for your patience until the support update becomes available.

We apologize for any confusion our FAQ wording from several hours ago regarding playability on Intel Arc GPUs may have caused. Please regard this as our latest official news on the subject. Thank you.”

The studio has also updated the game’s FAQ to reflect this same messaging, after it previously stated that the game "does not support Intel Arc graphic cards."

Crimson Desert has also gone from Mixed reviews to Mostly Positive reviews on Steam as the studio tackles feedback.

Are you relieved to see Pearl Abyss address concerns so quickly about the game?

Apple has set a date for this year’s Worldwide Developers Conference (WWDC), scheduling the keynote and Platforms State of the Union for June 8th through June 12th at Apple Park. As usual, we will be tuned in to find out what’s new in the 2027-branded editions of iOS, iPadOS, macOS, watchOS, and tvOS. After last year’s round of glassy visual tweaks, many are waiting for Apple Intelligence to deliver on delayed features, such as an AI upgrade for Siri.

In its press release, Apple says, “WWDC26 will spotlight incredible updates for Apple platforms, including AI advancements and exciting new software and developer tools.”

According to the latest rumors, this fall we could see a staggered iPhone hardware update, with the iPhone 18 Pro arriving months before the standard model, while Bloomberg’s Mark Gurman recently reported that other devices, like a new smart home hub, HomePod, and Apple TV box, are also waiting for the new Siri release first.

Image: Apple

 

Apple recently launched a slew of relatively iterative hardware updates, including the cheaper MacBook Neo, chip bumps for the MacBook Pro lineup, a new Studio Display XDR monitor, new AirPods Max headphones, and even a second-gen AirTag, so we’ll have to wait and see if anything else is lined up for a big reveal this summer.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 24 March 2026 at 6:15 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix

A McDonald’s in the Chinese megacity of Shanghai is testing humanoid robots in roles usually the preserve of human workers, with other types of robots also let loose inside the restaurant to greet and entertain diners.

Truth be told, the robots don’t look particularly advanced, but a video (below) showing them in action does hint at a future where bipedal bots and other machines handle routine tasks at fast food restaurants, from welcoming customers and taking orders to delivering food and cleaning the floor.

The McDonald’s trial, using robots supplied by Chinese firm Keenon Robotics, comes at a time of economic contradiction in China, where businesses in some sectors are struggling to hire even as millions of young people face difficulty finding work.

It’s this tension that makes the McDonald’s trial stand out, with restaurant operators interested in deploying a reliable, potentially low-cost workforce in a strategy that raises fears of displacement among human workers in the service sector, which up to now has been a popular route into the workforce.

The reality, however, is more complicated. China’s workforce is shrinking as the population ages, while many younger job seekers are reluctant to take on low-paid, repetitive work. In that case, robot technology could be used to fill gaps rather than simply replace people. Still, the presence of robots in such a visible, everyday setting highlights how quickly that balance could shift.

While it could be a while before McDonald’s deploys humanoid robots in a more meaningful way, adding them to restaurants as greeters and entertainers could potentially draw curious diners, especially families with kids who might want to interact with the machines while waiting for their meal to arrive.

Even if the fast food giant eventually wants robots to run its restaurants, such a scenario is almost certainly many years away, simply because the technology isn’t yet up to it. What feels more likely, at least in the short term, is a hybrid setup where human workers handle the majority of tasks while the robots take on more basic, customer-facing roles out front.

Source

China has made history by becoming the first nation to approve a commercially available brain chip to treat a disability. NEO, the implant developed by Neuracle Medical Technology, translates the thoughts of a person with paralysis into movements of an assistive robotic hand.

After 18 months of testing that proved its safety, China's National Medical Products Administration authorized the implant for people aged 19 to 60 with paralysis caused by neck or spinal cord injuries that prevent them from moving their limbs.

According to Nature, the implant embedded in the skull is about the size of a coin. Eight electrodes protrude from the chip and are placed on the brain, in the area that processes body movement. When the user imagines moving their hand, the chip sends a signal to a computer, which then translates it into actions performed by a prosthesis.

In practice, Neo allows thought to activate a mechanical glove capable of performing basic tasks: picking up objects, manipulating utensils, or moving hygiene items. Sources consulted by the magazine indicate that 32 people have already tested the device without reporting any adverse side effects.

Although several companies are developing brain implants and brain-computer interfaces (BCIs), none have yet received authorization to market their innovations as treatments. Even advanced projects, such as Neuralink, have faced side effects that have stalled any regulatory approval.

The Plan to Dominate the Brain Implant Industry

This development comes just months after China issued a policy document that outlines the country's strategy to create a globally competitive BCI industry within five years. The plan outlines 17 steps to lead the research, development, and application of these technologies.

"The Chinese government has always supported disruptive technologies. I think from the government's point of view, this policy means that BCI technology has already moved from the concept level to the product level," Phoenix Peng, cofounder NeuroXess and Gestala, both BCI companies, told WIRED earlier this year.

The policy document promotes the mass production of non-invasive, non-implantable devices in various forms: forehead-mounted, head-mounted, and ear-mounted, such as hearing aids, visors, and glasses. It also proposes testing these technologies in high-risk sectors, such as hazardous materials handling, nuclear energy, mining, and electricity generation.

For now, brain implants have a clear objective: to restore autonomy to people with disabilities. Experimental devices exist that allow users to communicate with “thoughts ,” operate computers without moving a muscle, and even technologies that are beginning to restore vision. With the commercial approval of Neo, the industry enters a new phase.

Source

The largest single-player game to release this year, or at least so far (looking at your Grand Theft Auto 6), is Crimson Desert. The latest creation by Black Desert developers Pearl Abyss, Crimson Desert, is a massive-scale open-world game set in the same universe as their previous game.

On its first day, the game managed to reach nearly 240,000 concurrent players. Saddled with a mixture of up and down reviews, both on Metacritic and from gaming outlets at large, the barrage of players looking to take on the continent of Pywel.

Over the weekend, the game entered a new realm of discussion when it was discovered that some of the art assets were generated using AI. All while hitting an all-time new high of 248,530 players.

At least, that’s what I thought was about to happen.

Pearl Abyss issues an apology

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 include the new Firefox mascot Kit, Apple's powers on the App Store, Microsoft's threat to OpenAI, and NVIDIA's GPT moment for graphics. Let's get started.

How did we get here?

It's always advised not to hardcode sensitive keys into your apps while hardcoding. One of China's biggest cybersecurity companies made a rookie mistake that even beginners should avoid. Qihoo 360 accidentally shipped a highly sensitive wildcard SSL private certificate inside the public installer for its 360 Security Claw AI assistant, exposing its exact master key needed to authenticate traffic on its backend infrastructure to anyone who downloaded the software.

Stolen Made in America spyware

It was reported this week that Russia repurposed a stolen US-made spyware to target iPhone users in Ukraine. It's a highly sophisticated suite of mobile hacking tools based on the Coruna toolset, which itself was originally developed by the Trenchant division of the defense company L3Harris.

While everyday users are an unlikely target for such multi-million-dollar spyware, its actual targets probably had little to no idea that their devices had been compromised. That's because Coruna exploits fundamental flaws in iOS's architecture. Peter Williams, who stole the confidential tool, was sentenced to seven years in federal prison.

Firefox gets a cute mascot

You might instantly picture a red panda wrapped around a globe whenever you think of Firefox. Mozilla has appointed a cuter official mascot for its web browser, named Kit, which is neither a red panda nor a fox. It can be found on the product web page, onboarding screen, in pop-ups for new features, or in confirmations after changing a setting. The browser-maker said it wanted something that brings you "some warmth and familiarity when you're browsing with Firefox."

Mozilla also announced new features and updates coming to Firefox 149 over the coming weeks, including a free built-in VPN, split screen, and Smart Window to quickly summarize articles, compare products, and more.

Apple can remove apps at any time

Apple won a lawsuit against the streaming app Musi, which was removed from the App Store in 2024. Musi had sought a legal route against Apple, accusing the company of acting in "bad faith" and that it "knowingly relied on false evidence" to delist it. However, the judge ruled in Apple's favor and sanctioned Musi's lawyers for making up facts.

Meanwhile, Apple's low-cost MacBook continues to remain in the headlines. Parallels Desktop added official support for MacBook Neo to run legacy and essential Windows apps. Parallels said Neo outperforms Windows laptops in single-core CPU performance, so Windows apps will feel quick and responsive.

Vibe check 'failed'

Apple also went against vibe coding apps, such as Replit and Vibecode, by silently blocking their App Store updates. It's reported that Apple wants vibe coding apps to open AI-generated apps in a separate browser on the device and remove functionality that allows apps to be built directly for Apple devices.

Because Googling your symptoms wasn't scary enough

The newest love interest for AI companies is the health sector. Perplexity launched its suite of connectors, called Perplexity Health, which can connect to different sources, including Apple Health. It works with the Perplexity Computer AI agent to answer health-related queries and tracks health metrics and trends over time through a personalized dashboard.

Google is also updating its Fitbit Health Coach to integrate with your personal health records for personalized recommendations. In Europe, the search giant is moving beyond fitness tracking by deploying AI features for patients through a partnership with DocMorris while sticking to the European Union's strict regulatory environment.

Microsoft threatens OpenAI, Britannica sues

Microsoft has been the biggest financial backer for OpenAI from the start. However, the tables seem to have turned, and the Redmond giant is now considering a lawsuit against the AI company and Amazon. Microsoft is arguing that the recent $50 billion cloud deal between Amazon and OpenAI violates its exclusive Azure agreements.

OpenAI found itself in hot water after Encyclopedia Britannica and its Merriam-Webster subsidiary filed a joint lawsuit against the AI company over copyright infringement. According to their complaint, OpenAI illegally scraped more than 100,000 protected articles and dictionary definitions to train its ChatGPT models without permission.

The OpenAI super app

One side effect of having too many apps is that things can become overwhelming for users. OpenAI must have realized that this could happen in the future and has decided to combine its desktop apps into a super app. In other words, the ChatGPT app, the Atlas web browser, and the Codex coding app will merge to become a single app in the future.

The said development was confirmed by OpenAI's CEO of Applications, Fidji Simo. OpenAI aims to reduce fragmentation and accelerate development against rivals like Google. In other news, OpenAI launched the new GPT-5.4 mini and nano to bridge the gap in its lineup. These models are designed to compete with high-efficiency models like Gemini 3 Flash.

Gemini for Mac, Vibe design

In an effort to go native, Google is finally bringing a dedicated Gemini app for macOS. An early beta version of the app is currently being tested with a limited number of users. The initial version reportedly looks like the Gemini app on iPhone/iPad and only includes core features of the existing Gemini client.

The search giant also announced a new 'vibe design' update for its AI design tool Stitch, bringing an infinite canvas, voice support, interactive prototypes, and other new features. Google also pledged $12.5 million in a collective effort to "improve the stability and security of the open source community."

Apple's most expensive headphones

Apple announced the AirPods Max 2 this week, upgrading its priciest wireless headphones to the H2 chip we have already seen on AirPods Pro. While the sound drivers and overall design remain the same, Apple has added a new high dynamic range amplifier for cleaner audio, 1.5x better ANC, improved Spatial Audio, and a dozen new AI features.

Meanwhile, Apple-owned Beats by Dre partnered with Nike to launch a special edition of the PowerBeats Pro 2 wireless earbuds, featuring a dual-tone visual overhaul with different logos on each earbud.

The iPhone-maker is also buying the Polish company MotionVFX, which is among the most popular names in building plugins and effects for Final Cut Pro and Apple Motion. One possible motivation is that Apple can use MotionVFX's massive catalog to offer premium content through Apple Creator Studio.

Unofficial Task Scheduler gets updated

Microsoft may not have the time to revamp Windows Task Scheduler, but someone did it a while ago with a Fluent Design-based unofficial remake. Known as FluentTaskScheduler, it was updated this week to version 1.6.0, dubbed "The Universal Order," which brings native support for Windows on ARM to improve performance. With improved content organization, you can categorize tasks by tags/categories, import tasks into folders, and search across items by name, status, path, or tags.

"GPT moment for graphics"

NVIDIA is baking a new DLSS AI upgrade that boosts gaming performance. CEO Jensen Huang unveiled DLSS 5 during the GTC 2026 Keynote and called it the "GPT moment for graphics." Its newest version, supporting gameplay up to 4K in DLSS 5, will introduce a neural rendering model capable of adding "photoreal lighting and materials" to pixels in real time, using color and motion vectors from each frame as input.

Unfortunately, the initial reaction from gamers and industry professionals wasn't what NVIDIA had expected: criticism over the altered 'AI slop' looks of the characters. Huang responded to it by saying that the critics are "completely wrong."

The chipmaker is joining forces with Intel to enable faster AI processing and inferencing performance by combining Xeon and DGX. Intel announced that its new Xeon 6 processors will serve as host CPUs for Nvidia DGX Rubin NVL8 systems. The chipmaker also refreshed its high-end mobile lineup with the new Core Ultra 200HX Plus series, bringing performance upgrades and architectural refinements.

Show bookmarks bar

Chrome for Android is finally rolling out the feature desktop users might take for granted: the bookmarks bar, which sits below the address bar on the screen. If you use Chrome on tablets and foldables, you can go to Settings > Appearance > Show bookmarks bar to enable the feature on your device running Chrome 146.

That said, here is some more browser news from the week:

• Perplexity's Comet browser is now available on iOS
• Vivaldi 7.9 brings customizable full-screen mode, better split view, and more
• Chromium-based Opera 129 reaches stable channel with Gemini integration
• Gamer-focused Opera GX browser has finally arrived on Linux

WhatsApp gets a new home

The popular instant messaging app is now available on another platform: Garmin smartwatches. Garmin announced that WhatsApp now supports its Fenix, Forerunner, Venu, and Vivoactive lines of smartwatches, and if you have one of these, you can download the app from its Connect IQ store. While you can read and reply to messages, and get notifications for calls. You can not make calls, however, which is the case across most of the smartwatch industry.

AI Overviews are crushing small publishers

AI Overviews can make silly mistakes and ask you to glue cheese to your pizza. But it's also showing some real, unwanted consequences for publishers of all sizes. A recent report revealed that small websites and publishers (with up to 10,000 daily page views) saw a 60% drop in web traffic from Google Search over the past two years. Meanwhile, medium-sized publishers (up to 100,000) saw a 47% drop, and larger publishers (more than 100,000) saw a 22% decline in traditional search referrals.

Solving AI's biggest bottleneck

On the flip side, the AI industry is facing significant supply bottlenecks and bandwidth constraints amid the exponential rise in AI usage, prompting AI companies to seek even better hardware. Samsung and AMD signed an MoU to develop the next generation of hardware to support AI infrastructure.

As per the agreement, Samsung will supply next-generation HBM4 memory for AMD's upcoming Instinct MI455X AI accelerators. Samsung will also provide highly optimized DDR5 memory for AMD's AMD’s 6th Gen EPYC processors, with both components serving as the backbone of AMD's Helios rack-scale architecture.

Amazon buys Rivr, Alexa+ in UK

Amazon is trying to make your package delivery as futuristic as it can get. In the newest attempt, it has acquired the Swiss robotics firm Rivr, which specializes in machines capable of last-mile doorstep deliveries. Rivr has been working on products such as a four-legged robot on wheels. While in the early stages, Amazon said it will use the tech to figure out what devices could be integrated into its delivery options.

On the software side, its generative AI assistant Alexa+ made its way to the UK, bringing smarter home control and deep service integration for users across the Atlantic. For starters, Alexa+ is much more capable than the standard version of the AI assistant people have used for over a decade.

Advanced flow for sideloading

Android's sideloading feature has been in the news ever since Google announced controversial changes last year that require developers to get verified. In the latest change, Google said it's building an "advanced flow" to make it easier for power users to download apps from unverified developers.

According to the search giant, it's a one-time process designed to prevent coercion to download malicious software from scammers. Google is also working on "limited distribution accounts" that allow students and hobbyists to share their apps with up to 20 devices.

What happened at Microsoft this week

For those wondering what happened under the Redmond giant's roof this week: Microsoft blamed Samsung for making C drive inaccessible, revealed unique details about Windows 11 version support, retiring Access Database Compare tool in June 2026, and adding the ability to pause Windows updates indefinitely.

Someone built a Windows utility called NoCopilotKey to make the Copilot key less annoying by restoring the CTRL button's functionality. You can check out Taras's freshly baked Microsoft Weekly roundup to catch up on all the interesting stories this week.

Same same

After turning the Xbox app into a universal launcher, Microsoft is testing a feature that Steam users have had for years. In a preview version of the Xbox app, the My Library tab shows a plus symbol, which opens a new window to add any program to the Xbox library and launch it.

The Quick Resume feature on Xbox Series X|S is getting some much-needed manual gears. You will be able to choose which games can use the Quick Resume feature. The latest additions to Xbox Game Pass bring even more horror and RPG heavy hitters, including Resident Evil 7: Biohazard, Like a Dragon, Infinite Wealth, and Disco Elysium.

Minecraft...

The Tiny Takeover update is coming to all platforms on March 24, bringing dozens of redesigned baby mobs (including for tamable and wild animals, villagers, zombies, horses, and more) with custom sounds and animations.

Mojang is working with Merlin Entertainments to turn Minecraft into a physical experience. The block-based universe is officially becoming a theme park expansion to Chessington World Adventures in Greater London in 2027. Mojang is also readying a sequel to Minecraft Dungeons from 2020, which will bring back the casual dungeon crawling experience.

Lost in Simulation

The City Update 14 is now live for Microsoft Flight Simulator and its 2024 release, with upgraded visuals and accuracy of the Netherlands and Belgium. Its development team used advanced TIN (triangulated irregular network) surface texturing to update ten areas across the two regions.

Meanwhile, Electrician Simulation is free to claim on the Epic Games Store in its latest giveaway. As the name suggests, you can assume the role of an electrician and simulate fixing various equipment and appliances.

Death Stranding 2 for PC leaked on Steam

A story floated around piracy forums that the developer of Death Stranding 2's PC version accidentally uploaded an unencrypted version of the game for pre-loading on Steam. This came after a leaked build of Death Stranding 2 showed up on torrent and piracy-focused portals.

In other news, Valve is pushing the massive SteamOS 3.8 Preview update, adding Steam Machine support, KDE Plasma 6.4.3, kernel 6.16, better VRR, and broader handheld compatibility. Valve has also launched its store-wide first Steam Spring Sale for 2026, offering heavy discounts for thousands of games.

What else in gaming?

The latest issue of Pulasthi's Weekend PC Game Deals curates a number of multiplayer games on sale this week. Xbox Free Play Days for the weekend include Sea of Thieves: 2025 Edition, Hell Let Loose, Trailmakers, and other titles.

Meanwhile, the free tier of NVIDIA GeForce NOW is losing some big games though, including Cyberpunk 2077. However, support for VR devices has been updated from 60fps to 90fps for Ultimate tier members.

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

• Bethesda announces "countless improvements" to Starfield and a new story expansion
• Double Fine's pottery-based party game Kiln gets an April release date
• Ubisoft pulls the plug on Rainbow Six and Ghost Recon creator studio, begins mass layoffs
• Life is Strange: Reunion, Icarus, and over 40 new games are coming to Xbox next week

• Crimson Desert blocks Intel ARC GPUs from playing, studio asks owners to refund

   

From the review corner

This week, Steven got his hands on the HYTE Gundam Wing Y70 Touch Infinite case, a special edition case celebrating the 30th anniversary of Bandai Namco's Mobile Suit Gundam Wing anime. Each of the numbered units is extremely lightweight, with ample room for cable management, and comes with a pre-installed LCD and a PCIe 4.0 riser cable. However, it's a bit pricier than the "typical" Y70 Touch Infinite, and HYTE should have included a PCIe 5.0 riser cable to compensate for it.

Timberborn 1.0

If you think beavers can succeed humans in a post-apocalyptic world, you should check out Pulasthi's review of Timberborn 1.0. It's a Lumberpunk experience with surprisingly detailed water physics and resource management systems, and dense, puzzle-like maps. The game gets brownie points for replayability, as there are two beaver factions available to play with either.

On the flip side, Timberborn 1.0 comes with a glaringly barebones tutorial, no in-game wiki or resource page for mechanics, and districts seem underbaked. It's available for Windows and macOS on Steam for $34.99

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.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 23 March 2026 at 6:57 am AEST (my time).

News posts: 2023 5,800+ | 2024 5,700+ | 2025 5,700+ | 2026 (to end of February) 854

RIP Matrix