In late 2023, storage company Western Digital announced plans to split itself into two companies. One, which would still be called Western Digital, would focus on spinning hard drives, which are no longer used much in consumer systems but remain important to NAS devices and data centers. The other, called SanDisk, would handle solid-state storage, including the drives that Western Digital sold to consumers under its Blue, Black, Green, and Red brands.

That split effectively undid what Western Digital did a decade ago when it bought SanDisk for $19 billion. And we’re just now starting to see the way the split will affect the company’s existing consumer drives.

Today, SanDisk announced that mainstream WD Blue and WD Black SSDs would be discontinued and replaced by SanDisk Optimus-branded disks with the same model numbers.

WD Blue drives will now be “SanDisk Optimus” drives, starting with the Optimus 5100, a rebadged version of the WD Blue SN5100. Mid-tier WD Black drives will be branded as “SanDisk Optimus GX,” and the Optimus GX 7100 will replace the WD Black SN7100. And high-end WD Black drives will become “SanDisk Optimus GX Pro” SSDs, with the Optimus GX Pro 850X and 8100 replacing the WD Black SN850X and 8100 drives.

Given that these are all fast NVMe SSDs, I suspect the average user would have trouble detecting much of a difference between the low-end WD Blue/Optimus drives and the high-end WD Black/Optimus GX Pro SSDs. But the functional differences between the drives remain the same as before: the Blue/Optimus 5100 uses somewhat slower and less durable quad-level cell (QLC) flash memory, while the Black/Optimus GX 7100 uses triple-level cell (TLC) memory. The Black/Optimus GX Pro 8100 maximizes performance by stepping up to a PCIe 5.0 interface instead of PCIe 4.0 and including a dedicated DRAM cache (the 5100 and 7100 each claim a small chunk of your system RAM for this, called the Host Memory Buffer, or HMB). The 850X is a slightly older drive that keeps the dedicated DRAM but is also limited to PCIe 4.0 speeds.

Other drives are still in flux

Rebrandings are always fraught—you can end up just confusing people more in your attempts at clarity and simplification. But SanDisk’s new names at least make some logical sense: The more words the name has, the better the drive is meant to be. The model number continuity will help prevent confusion during this period of overlap between WD- and Optimus-branded SSDs, and the demarcation between the GX and GX Pro drives emphasizes the differences between those SSDs in a way that the catch-all “WD Black” branding didn’t.

Users of both WD-branded and SanDisk-branded SSDs should be able to use the SanDisk Dashboard software to check for firmware updates and perform other kinds of drive maintenance; the equivalent WD Dashboard software package was formally discontinued a year ago.

SanDisk’s press release doesn’t mention the WD Green or WD Red-branded SSDs, which are targeted toward entry-level and NAS usage, respectively. When contacted for comment, SanDisk representative Robyn Pagulayan told Ars that the company had nothing to announce about the fate of those drives.

SanDisk’s rebrand comes at a strange time for the memory business. Though not as badly affected as RAM prices, SSD prices have been spiking recently due, in part, to demand from AI data centers (and now, at least partly, by companies and consumers rushing out to buy things just in case prices get even worse). Shortages of both RAM and NAND flash chips will likely keep prices elevated in the short- to medium-term, at least. Due in part to this volatility, Micron announced last month that it would be discontinuing its Crucial-branded consumer SSDs and RAM sticks.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 6 January 2026 at 12:14 pm AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

Anyone can now try Alexa+, Amazon’s generative AI assistant, through a free early access program at Alexa.com. The website frees the AI, which Amazon released via early access in February, from hardware and makes it as easily accessible as more established chatbots, like OpenAI’s ChatGPT and Google’s Gemini.

Until today, you needed a supporting device to access Alexa+. Amazon hasn’t said when the early access period will end, but when it does, Alexa+ will be included with Amazon Prime memberships, which start at $15 per month, or cost $20 per month on its own.

The above pricing suggests that Amazon wants Alexa+ to drive people toward Prime subscriptions. By being interwoven with Amazon’s shopping ecosystem, including Amazon’s e-commerce platform, grocery delivery business, and Whole Foods, Alexa+ can make more money for Amazon.

Just like it has with Alexa+ on devices, Amazon is pushing Alexa.com as a tool for people to organize and manage their household. Amazon’s announcement of Alexa.com today emphasizes Alexa+’s features for planning trips and meals, to-do lists, calendars, and smart homes. Alexa.com “also provides persistent context and continuity, allowing you to access Alexa on whichever device or interface best serves the task at hand, with all previous chats, preferences, and personalization” carrying over, Amazon said.

The Alexa+ website’s homepage.

Scharon Harding/Amazon

 

Amazon provided this example of someone using the Alexa+ website for organizing lists.

Amazon

 

Amazon already knew a browser-based version of Alexa would be helpful. Alexa was available via Alexa.Amazon.com until around the time Amazon started publicly discussing a generative AI version of Alexa in 2023. Alexa+ is now accessible through Alexa.Amazon.com (in addition to Alexa.com).

 

“This is a new interaction model and adds a powerful way to use and collaborate with Alexa+,” Amazon said today. “Combined with the redesigned Alexa mobile app, which will feature an agent-forward design, Alexa+ will be accessible across every surface—whether you’re at your desk, on the go, or at home.”

Amazon provided this example of someone using the Alexa+ website to manage smart home devices.

Credit: Amazon

Alexa has largely been reported to cost Amazon billions of dollars, despite Amazon’s claim that 600 million Alexa-powered devices have been sold. By incorporating more powerful and generative AI-based features and a subscription fee, Amazon hopes people will use Alexa+ more frequently and for more advanced and essential tasks, resulting in the financial success that has eluded the original Alexa. Amazon is also considering injecting ads into Alexa+ conversations.

 

Notably, ahead of its final release and while still in early access, Alexa+ has been reported to be slower than expected and struggle with inaccuracies at times. It also lacks some features that Amazon executives have previously touted, like the ability to order takeout.

 

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 6 January 2026 at 12:12 pm AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

Qualcomm is starting 2026 the same way it ended 2025: fast, confident, and very aware that the Windows PC world is shifting under everyone’s feet. At CES 2026, the company officially unveiled Snapdragon X2 Plus, the newest member of the Snapdragon X2 Series and the chip that’s poised to matter far more to everyday buyers than the announced in September.

I had the chance to see early X2 Plus laptops in person during a press preview in New York City back in early December. Qualcomm let us run live benchmarks on pre‑production hardware, and—much like last year’s X Plus rollout—the numbers lined up almost exactly with what the company is now publishing.

Qualcomm has earned a reputation for giving reviewers honest, reproducible performance data, and the X2 Plus continues that trend (I even had one major OEM tell me this off the record, noting how Qualcomm always delivers on time and never overpromises; and yes, that was shade thrown at Intel).

And that’s important, because this chip is aimed squarely at the mid‑to‑upper‑range Windows laptop market. In this segment, most people actually buy PCs, where businesses refresh fleets, and where OEMs need predictable performance and efficiency.

The X2 Plus is the platform that will show whether Qualcomm’s momentum in 2024 and 2025 was a flash in the pan or the beginning of a long‑term shift toward Arm‑based Windows machines.

From what I’ve seen so far, it’s the latter.

A “Plus” chip that feels a lot like an Elite

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.

Do you remember Andromeda, Microsoft's cancelled project for dual-screen devices? One of its builds leaked, and enthusiasts put it to great use by porting it to the original Surface Duo. If you have this failed Android smartphone, you can try installing it right now, as all the files and guides have been uploaded to GitHub for everyone to discover.

Speaking of remembering, there was a time when we were able to activate Windows via phone and without an active internet connection. This feature is now unfortunately gone, as Microsoft quietly shut down phone activation.

This week's stats include the latest report from Valve. On Steam, Windows 11 continues crushing Windows 10, and in December 2025, Windows 11 reached a new all-time high of 70.83%. This time, we also have a detailed breakdown of Steam stats for Linux.

Windows 11 has a lot of ads across its user interface, which are often disguised as "Recommendations." It appears that you will soon be able to turn off some of them. Users discovered that Microsoft is working on updates to the Nearby sharing settings section and adding a toggle to remove app recommendations from the sharing UI.

Although we had little news because of the holiday season, there were still plenty of stories and useful guides. To start, here is a list of my ten favorite features that Windows 11 received in 2025. Not everything was AI slop last year, and some actually useful stuff made it to the operating system. Also, we have a list of things that Microsoft removed from Windows 11 in 2025.

Next, we have a few guides that explain how things in Windows work. For example, here is one about indexing files in Windows 11 and how this feature can make File Explorer and Search significantly better. Also, here is an explanation of what the "Refresh" button does for your desktop. Spoiler alert: it cannot boost performance.

Finally, we published a review of how the initial setup evolved over the last two decades in Windows. It showed how OOBE used to be fun and efficient, only to turn into a slow, ad-ridden nightmare years later.

Windows Insider Program

Nothing in the Insider Program this week due to the holiday season.

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.

Files, one of our favorite file managers for Windows, received a small but very useful update. The latest release now supports one of the best PowerToys modules. You can click a file and press Spacebar to preview it without opening the corresponding app. Besides that, there are some minor quality-of-life improvements.

Speaking of PowerToys, we continue our Closer Look series where we discover what PowerToys has to offer. This year, we reviewed the following modules: Command Palette, Color Picker, and Light Switch.

Finally, here is an overview of all the new features that Microsoft added to Teams in December 2025.

Reviews are in

Here is the hardware and software we reviewed this week

Christopher White published his review of the Synology DS925+, a solid four-bay NAS with good performance, two 2.5Gb ports, a solid operating system, and more.

On the gaming side

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

GOG, a popular store of DRM-free games, has a new owner. The store parted ways with CD Projekt Red, as it was taken over by Michał Kiciński. Despite the big change, GOG promises that its mission to provide gamers with titles that are "easy to find, buy, download, and play forever" will remain unchanged as an independent company. This means offering DRM-free games will continue to be the main pillar of GOG, along with offline installers and optional launchers.

NVIDIA kicked off 2026 with a new set of games in its GeForce NOW streaming service. It received 14 new titles, including My Winter Car, Banishers: Ghosts of New Eden, Eternights, and more. Keep in mind that there is now a 100-hour monthly play limit.

As usual, here is our weekly Weekend PC Game Deals article with a lot of various discounts, including the latest freebie from the Epic Games Store. You can grab Total War: Three Kingdoms and Wildgate for free until January 8.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 5 January 2026 at 4:45 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

Artificial intelligence is going to be everywhere. Not as a category, but as a label glued onto almost every product. Laptops, TVs, appliances, wearables, cars-if it has a chip, it will claim some form of AI advantage. Much of it will be backend automation rather than visible features, but marketing will make sure the word is front and center.

Smart glasses look set to dominate the show floor. After years of false starts, nearly every major brand now wants a piece of what comes after the smartphone. Expect a wide mix of designs: camera-focused glasses, audio-first "AI glasses," lightweight display glasses, and hybrids that sit somewhere between eyewear and XR headsets. No single design is likely to emerge as the obvious winner, but the sheer volume will be hard to miss.

TV technology is also back in the spotlight. Display makers will push new panel types, higher brightness, and higher refresh rates, whether or not those gains translate into real-world improvements. AI processing will be heavily promoted here as well, often tied to image enhancement, motion smoothing, and content upscaling.

Mobility continues to expand its footprint at CES. Electric vehicles, e-bikes, scooters, and concept transport devices will take up large sections of the show. One noticeable shift is the slow return of physical controls in car interiors, after years of touchscreen-heavy designs. Buttons and dials are quietly becoming a selling point again.

Robotics will blur further into the smart home category. Beyond robot vacuums, expect more humanoid and semi-humanoid demos focused on lifting, carrying, and basic household tasks. These systems are not close to mass-market pricing, but CES remains a favorite venue for showing what might arrive several product cycles from now.

Alongside all of that will be the usual CES staples: laptops, monitors, audio gear, accessories, and experimental form factors that may or may not ever ship. The mix is familiar, but the density feels higher this year.

CES 2026 looks less like a reset and more like a return to excess, with AI as the connective tissue tying together a very crowded hardware show floor.

Is someone heading to CES this year?

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 5 January 2026 at 4:43 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

As Windows 10 nears its end-of-support deadline, Windows 11 has emerged as the leading desktop operating system globally, though its adoption has lagged its predecessor, with enterprises delaying upgrades amid hardware constraints, cost scrutiny, and limited incentives beyond security.

In June 2025, Microsoft announced that support for Windows 10 would end in October 2025, closing a nearly decade-long run for the operating system (OS).

After the deadline, Windows 10 PCs will no longer receive security updates, feature upgrades, or technical support. Microsoft explained in a blog post that while the devices continue to function, the lack of regular security patches will increase exposure to cyber threats. For enterprises, running unsupported software could also raise compliance and regulatory risks. Application support is expected to taper off over time as developers shift focus to newer platforms.

To ease the transition, Microsoft said Microsoft 365 Apps on Windows 10 will continue to receive security updates until October 2028 and feature updates until August 2026, though without technical support.

Microsoft has positioned Windows 11 as part of a broader push to modernise its computing platform and embed AI across its products, citing improved security, manageability, and productivity as key drivers of the shift.

Desktop market

StatCounter data shows that by mid-2025, Windows 10 occupied 53 per cent of the global desktop market share compared with Windows 11’s 43.22 per cent until June. By December 2025, the balance had shifted, with Windows 11 taking the lead at 50.68 per cent and Windows 10 declining to 44.64 per cent in response to the approaching end of official support for Windows 10.

However, analysts noted that Windows 11 has demonstrated a slower adoption curve than Windows 10 at comparable stages in their life cycles. Kishor Fogla, Founder, Yellow Slice, highlighted that Windows 10 reached approximately 50 per cent global market share within the first three years of its launch, aided by its free upgrade and low hardware requirement strategies.

Windows 11, on the other hand, has been unable to similarly accumulate adoption four years post-launch, only entering the mid 30 per cent adoption rate.

“This gap is due to stricter hardware requirements, such as TPM 2.0, fewer upgrade options for older PCs, and a refined UI design. Many users were content with the previous Windows 10 design and therefore had no incentive to upgrade. The changes from a UX standpoint were minor and therefore demonstrated no urgency for an upgrade,” he explained.

Global supply chain

Nitin Lahoti, Founder & Director at Mobisoft Infotech, added that Windows 11 was launched during global supply chain pressure, due to which many organisations chose to wait. Slower uptake is driven more by enterprise delay than by consumer adoption since the latter usually upgrade when they buy new devices.

“Many new laptops already ship with Windows 11. Enterprises behave differently. They test longer and move more slowly. IT teams look at compatibility, security tooling, and training impact. Windows 10 still works well for most workloads. There is no urgent business pain, which reduces motivation. Consumers follow availability, and enterprises follow risk assessment. In this case, enterprise caution has had a bigger impact on overall adoption numbers,” he said.

Enterprises dilute the market for active Windows devices as they tend to upgrade less frequently. Despite Windows 11’s market share being over 50 per cent, a considerable portion of enterprise devices still run Windows 10. Enterprises care more about system stability, uninterrupted workflows, and retraining than visuals, which is a big reason they have not implemented organisational-wide upgrades.

Security updates

Lahoti shared that lifecycle extension plays a bigger role. Many enterprises made large Windows 10 investments during the pandemic, due to which devices are still usable, performance is stable and security updates continue. At the same time, PC refresh cycles have lengthened from around four years to five or six, as tighter budgets prompt CIOs to extract more value from existing hardware. As device upgrades slow, OS migrations tend to slow in tandem.

Windows 11 is being seen as a forced upgrade because there is a deadline approaching to support Windows 10. According to Fogla, this increased adoption rate is a direct result of the perceived urgency to upgrade. Using Windows 10 and Windows 11 are the only two options available. Windows 10 will be unsupported in a few years, creating a compliance-driven Windows 11 landscape.

The analysts added that although Windows 11 offers stronger baseline security and a more modern architecture — important for regulated industries — the day-to-day user experience remains largely unchanged. As a result, many organisations are deferring upgrades until clearer business value emerges through hardware refreshes, AI-driven features, or deeper ecosystem integration.

Source

The Streisand effect continues to be real, as Microsoft CEO Satya Nadella's AI comments go viral.

A couple of days ago, Nadella penned a short note on his hopes for artificial intelligence going into 2026. As you know, Microsoft is very much "all in" on AI, with Azure providing a significant chunk of the backbone for OpenAI's ChatGPT. Microsoft has been baking its ChatGPT-powered Copilot app into virtually every product it has, whether you like it or not. The brute force by which Microsoft is introducing these products has led to an unrelenting backlash on social media, and Nadella's latest comments reignited the commentary in a big way.

In the piece, Nadella said that he hoped society would "move on" from questions of "slop" for AI — emphasizing that for the technology to gain acceptance, it needs to move beyond spectacle.

Indeed, the vast majority of AI usage in the mainstream consciousness right now revolves around misinformation, dumb memes, and at worse, illegal abuse. xAI's Grok is currently being investigated by various authorities for allowing sexualized AI images of children, and OpenAI's ChatGPT is being sued for potentially causing a dreadful murder-suicide. But hey, at least we can generate cat memes more quickly than we could previously.

As such, it certainly doesn't seem like society at large is ready to accept AI as Microsoft's Nadella hopes. "Microslop" began trending on X yesterday, in the wake of our coverage on the topic.

The biggest tech show of the year kicks off next week, as some of the industry’s top players show up to Las Vegas for CES 2026. We’ll be there to see all the new product demos we can and to bring you the most exciting news from the show. Follow our coverage for a preview of all the new tech these companies are planning to launch in 2026.

Expect to see the usual suspects: laptops, smart home gadgets, and TVs, and a whole lot more wearables and health tech. We’re anticipating seeing more products with AI integration. Also, robots. Perhaps, humanoids even.

CES 2026 officially starts on Tuesday, January 6th, but stay tuned for news and announcements starting Sunday ahead of the show floor opening, when there are also lots of press conferences. Here are the major beats we’re expecting to see at the show.

Intel’s upcoming Panther Lake chip was preannounced in October.

Image: Intel

 

Laptops are always big at CES, with most major companies — save for Apple — announcing new models. They range from iterative spec bumps to whole new designs. Plus, there’s always the occasional concept that may or may never come out.

CES 2026 should bring new laptops featuring three new chips: Intel’s Panther Lake, Qualcomm’s Snapdragon X2, and AMD’s rumored “Gorgon Point” processors. Intel and Qualcomm are hyping up the efficiency of their new chips while simultaneously emphasizing graphics power — which always sounds too good to be true and often is. But battery life ranges from good to excellent among current offerings, and AMD proved with Strix Halo that integrated graphics could be downright impressive. I’m intrigued to see how things progress from here, especially in a year that doesn’t have new Nvidia GPUs. (But then again, Nvidia may soon compete in a different way.)

Regardless of how the latest chip showdown shakes out, I hope we see more unique ideas in laptop form factors. Give me more rollables, dual-screens, foldables, or any other clever uses of screen real estate that can be dreamed up. But like my favorite laptop from the last CES, let’s hope the radical stuff makes it past the concept stage.

- Antonio G. Di Benedetto

Robots and smart locks and AI, oh my! CES 2026 will be overrun by robots. From ever-wilder vacuums, pool cleaners, and lawn mowers, to humanoid bots with hands, limbs, and deeper intelligence, this will be the year smart home robotics moves from science fiction to science fact.

Will we all have humanoids doing our laundry by the end of the year? Nope. But the idea of robotic helpers, in some form, will soon be mainstream, thanks to huge advances in computer vision. And it all kicks off next week in Vegas.

Similarly, AI is powering advances in security cameras, moving them beyond pure surveillance and into a more integral role in home automation. Expect new capabilities that use deeper intelligence to provide much-needed context. Cameras becoming part of Matter is also a step forward here, and I hope to see companies announce support for the standard — even though the platforms are lagging behind.

The other big trend will be smart locks. Yes, I said this last year (and I was right), but delays around the Aliro standard slowed the onslaught. Still, smart locks remain the hottest thing in the smart home — being both an excellent entry point (ha) and a lynchpin for a fully integrated home. Expect to, once again, see a slew of palm, facial, and UWB-based unlocking locks, along with better form factors, as tech companies finally accept that not everyone wants Star Trek–style hardware on their front doors.

- Jennifer Tuohy

CES 2025 was a big year for PC gamers — new flagship Nvidia graphics cards, SteamOS handhelds, and 27-inch 240Hz 4K OLED displays — but don’t expect 2026 to offer the same! On the PC side of things, Nvidia’s RTX 50 Super has likely been postponed, we can’t trust Intel’s latest desktop GPU tease, and we’re not expecting gaming GPUs from AMD. The historically splashy Razer doesn’t have a booth on the floor for the second year in a row, and the “Gaming / XR” chunk of the Las Vegas Convention Center looks like it’s only tangentially about gaming this year: it looks like we’ll see more video glasses there than anything else!

It seems like Lenovo will have a rollable gaming laptop and a SteamOS version of the existing Legion Go 2 handheld, and I’m really curious how integrated graphics in Intel’s and in Qualcomm’s latest laptop chips might perform and whether they’ll fit in handhelds too. Lastly, Lego is going to be at CES for the first time in years…

- Sean Hollister

Every major TV manufacturer will compete with their own mini or micro RGB TVs in 2026.

Image: LG

 

What began with last year’s CES — when Hisense revealed its first RGB Mini LED TV, which it officially released this past summer — is primed to blow up at CES 2026 when every company will likely show their own RGB LED TVs. Samsung and LG have already announced plans to release smaller sizes, with Samsung offering its 2026 Micro RGB LED TVs with panels ranging from 55 to 100 inches, and LG debuting its first Micro RGB evo TVs in three sizes of 75, 86, and 100 inches.

Last September, TCL announced two new RGB LED TVs for the Chinese market and there are expectations of similar news for the US and Europe at CES. And I hope Hisense will extend its UX line into smaller screen sizes. (While Sony showed off its RGB LED in 2025 and filed a patent for something called True RGB, it hasn’t made major TV news at CES for a few years. I don’t expect that to change in 2026.)

The big question that has yet to be answered is how early into 2026 will we see any of them for sale and how much they’ll cost. Both the 115-inch Samsung Micro RGB LED TV and 116-inch Hisense RGB Mini LED in 2025 were in the $30,000 ballpark, with a 100-inch version of the Hisense announced for $19,999. The technological advancements made to RGB LED over the past year, and the promise of stiff competition, is sure to bring prices closer to a more manageable range. Once CES begins, I hope we’ll find out exactly how manageable.

- John Higgins

CES is not for normie phones. The phones that make it to CES are weirdo phones, and the number-one weirdo phone I’m hoping to see one way or another is the Samsung Galaxy Z Trifold. Samsung isn’t usually one for showing off phones at the show; the S-series usually gets a refresh a couple of weeks later in January. But this year might be different: the Trifold just launched in Korea, and CES seems like a nice big stage for a global debut. I am ready, in my heart of hearts, for a phone with two hinges.

Will it be excessive? Undoubtedly. Will it cost a buttload of money? You know it. But think of all the things you can do with a phone that’s also a proper 10-inch tablet. Maybe this is the phone that will deliver on the promise of leaving my laptop at home. Are you there, DeX? It’s me, Allison. Either way, it’s gonna be so sick the first time I fold up this phone and then fold it again. I think it’s a fair bet that we’ll see this double-folding beaut at CES, and what better way to usher in the Year of the Folding Phone?

- Allison Johnson

Smart glasses will likely have a big presence at this year’s show too.

Photo by Antonio G. Di Benedetto / The Verge

For so many years, wearables went hand-in-hand with health and fitness tracking. This year, I’m expecting to see fewer fitness trackers and more XR and AI devices. Specifically, smart glasses. They were a big fixture at CES 2025, and I see that trend continuing into 2026. But while the commercial market is dominated by so-called AI glasses — lightweight devices like the Ray-Ban Meta glasses — there’s much more variety at CES. I expect we’ll see some smart glasses that blur the line with headsets and maybe some interesting takes on how to embed displays.

On the health front, I wouldn’t be surprised if we see the word “longevity” tacked onto products. Think devices that are meant to help you live longer or prevent chronic diseases. Unfortunately, that means bodily fluids like blood and urine. (Why? To try and get a window into hormonal and metabolic health.) We’ve already seen this from Whoop, Oura, and Withings in 2025, so I expect we’ll see more experimental takes from smaller companies at the show.

- Victoria Song

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Saturday 3 January 2026 at 4:32 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

Microsoft CEO Satya Nadella wants you to stop calling AI "slop." Our Managing Editor Jez Corden covered Nadella's 2025 year-in-review blog post. In that piece, the CEO recaps the growth of AI and looks toward the future.

Nadella did not mention the report indicating that people don't want to use Microsoft's AI products. He also skipped over how 2025 was a disastrous year for Windows 11, due in part to forced AI integrations.

What he did do, however, was include quite a few phrases and pieces of jargon that seem eerily similar to text generated by Copilot.

Here are a few phrases from Nadella's blog post:

• "We have moved past the initial phase of discovery and are entering a phase of widespread diffusion."
• "We are beginning to distinguish between "spectacle" and "substance"."
• "Amidst this “model overhang,” where capability is outpacing our current ability to use it to have real world impact."
• "This is what I hope we will collectively push for in ‘26 and beyond."

If you've used Copilot to summarize meetings or generate text, you've developed a sense for which text is AI generated or refined by AI. I think it's more likely that the piece was made with the assistance of AI than generated entirely by a tool like Copilot.

I want to be clear; this post is mostly made playfully. Nadella is intelligent and more than capable of writing about tech — or many other topics — without any assistance. The CEO has hard knowledge from decades in the industry. Even if TIME snubbed him in its list of AI innovators, Nadella is extremely successful.

That being said, I don't think this most recent blog post was his best. It comes across as wooden and is full of jargon. It also lacks a personal touch.

The fact my first article of 2026 is about artificial intelligence is probably tone-setting for what will be another year dominated by AI news and headlines.

It's hard to avoid right now, particularly if you're a user of Microsoft ecosystem products. Every single app, service, and product Microsoft has on the market now has some kind of AI integration, regardless of quality and usefulness.

Microsoft Copilot is the tip of the spear for the firm, powered entirely by ChatGPT and Microsoft's savvy early investments in OpenAI. Its interface is pre-installed now on Windows PCs, and has a commanding position on most mobile app stores as of writing. It's nowhere near as widespread as OpenAI's ChatGPT service, though, and advancements in Google Gemini sees Microsoft's old arch rival rapidly outpacing the competition — particularly in enterprise integrations, where Microsoft has its sights primarily set.

The oft-forced, oft-useless Microsoft Copilot integrations on Windows and other consumer products have people exploring alternatives more so than ever before. Entire governments are abandoning Windows for Linux, and there's more interest in Linux consumer-grade distros than any time I can remember. Despite the noise about the degradation of quality in Windows, the price gouging on Xbox, and the apparent abandonment of Surface — Microsoft CEO Satya Nadella made no mention of any of them in a recent post (via LinkedIn) to close out the year.

If you had any illusion that Microsoft might address concerns about any of its major product categories in 2026, Nadella's "Looking Ahead to 2026" article offers an insight into the company's focus for the new year, and yep, it's all about AI.

Forget about New Year's resolutions. Instead, dedicate your New Year's Day to completing a few simple tech chores that will make you feel instantly better about the state of your digital life.

The first four are very quick and simple, and the last six take a little more time but can still be done in less than an hour.

1. Dump Your Computer's Trash

How long has it been since you emptied the trash on your desktop? Start off with this dead simple, two-click task and bask in the motivation you feel to continue with your clean-up list.

2. Close Excess Tabs and Clear Your Browser History

Listen up, tab hoarders! Today is the day you free yourself of the browser window from hell. If you're reluctant to let go of all your opened tabs, consider bookmarking them first. In Chrome, go to Bookmarks > Bookmark All Tabs and move on with your life. Delete your browsing data, too, on both desktop and mobile devices.

3. Use Compressed Air to Clean Your Devices

It's time to blow out all the crumbs and lint from your keyboard. Grab a can of compressed air and work its magic on your keyboard or laptop. You can also clean all the ports on your laptop, phone, and tablet while you’re at it.

4. Tidy Up the Desktop

When was the last time you actually saw your virtual desktop? Take all the loose files and move them to a folder. If you're not sure where to start, make a folder called "2025 and older" and dump everything in there.

If you have a little more time, even just 10 minutes, sort your files in whatever way makes the most sense for you. There's no right way to do it. There's only the way that helps you find what you need.

5. Unsubscribe From Unwanted Emails

It seems easier and faster to press delete than to unsubscribe from unwanted emails. But then more emails keep coming. Make January 1 the day you invest an extra 15 seconds per message to fully remove yourself from some email lists.

If you have a soft spot in your heart for emails with deals and discounts, consider setting up an alias address for them. That way, you can keep deals emails totally separate from personal messages, bank notices, updates from your kids' school, and other important information. The reason I prefer using an email alias rather than using your real address and letting deal emails go to a dedicated folder or tab is because of what happens when companies sell your information to a third party. If you get scammy messages that look like they're coming from your bank, but they are going to your email alias, you don't have to sleuth around to figure out if they're legitimate—they're not.

6. Back Up the Photos on Your Phone

Do you have photos on your phone that are more than a year old? Are they backed up somewhere automatically? If not, take care of it today.

How much time you need for this task depends on whether you already have a cloud storage service you use for photos. Seeing as photos are among people's most cherished data, it's worth spending, say, half an hour finding a service, signing up, and setting your photos to automatically back up there. Keep copies of your most loved photos on your phone, but get rid of the rest to free up space.

7. Give Your Google Drive Files a Once-Over

If you have a Google account, the chance you have unexpected files in your Google Drive and Gmail is high. One example is you might have large PDFs of menus from scanning QR codes at restaurants if the provider hosted the file on Google Drive and you were signed to your account when you viewed it. You might also have files others emailed you lurking in Google Drive.

Delete this unnecessary stuff from Google Drive. I recommend starting by viewing files Shared with me in the left rail of Google Drive.

8. Scan Any Loose but Important Papers

Scan and digitally file any papers that have piled up over the year. Once you have a scanned copy, shred and recycle or compost the paper responsibly.

9. Run a Privacy Checkup

One of my gripes about online services, including social media, is they sometimes change your settings when they roll out updates, or they opt you into features you do not want. Run a privacy checkup on your online accounts, paying particular attention to social media (don't forget YouTube), as well as Google. Look at what sort of data is being shared under your current settings and decide if you’re still comfortable with it.

10. Review and Cancel Subscriptions

Review and cancel any unwanted subscriptions, such as streaming services, newsletters, and memberships that you no longer need. With many services, you can cancel any time and continue using what you've paid for until the pay period ends. If you're tightening your budget in 2026, consider canceling all streaming services and keeping only one at a time. In some cases, you can cancel a prepaid membership and get refunded for the unused time, though it's rare.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 2 January 2026 at 4:28 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

Following two years of immense hype in 2023 and 2024, this year felt more like a settling-in period for the LLM-based token prediction industry. After more than two years of public fretting over AI models as future threats to human civilization or the seedlings of future gods, it’s starting to look like hype is giving way to pragmatism: Today’s AI can be very useful, but it’s also clearly imperfect and prone to mistakes.

That view isn’t universal, of course. There’s a lot of money (and rhetoric) betting on a stratospheric, world-rocking trajectory for AI. But the “when” keeps getting pushed back, and that’s because nearly everyone agrees that more significant technical breakthroughs are required. The original, lofty claims that we’re on the verge of artificial general intelligence (AGI) or superintelligence (ASI) have not disappeared. Still, there’s a growing awareness that such proclaimations are perhaps best viewed as venture capital marketing. And every commercial foundational model builder out there has to grapple with the reality that, if they’re going to make money now, they have to sell practical AI-powered solutions that perform as reliable tools.

This has made 2025 a year of wild juxtapositions. For example, in January, OpenAI’s CEO, Sam Altman, claimed that the company knew how to build AGI, but by November, he was publicly celebrating that GPT-5.1 finally learned to use em dashes correctly when instructed (but not always). Nvidia soared past a $5 trillion valuation, with Wall Street still projecting high price targets for that company’s stock while some banks warned of the potential for an AI bubble that might rival the 2000s dotcom crash.

And while tech giants planned to build data centers that would ostensibly require the power of numerous nuclear reactors or rival the power usage of a US state’s human population, researchers continued to document what the industry’s most advanced “reasoning” systems were actually doing beneath the marketing (and it wasn’t AGI).

With so many narratives spinning in opposite directions, it can be hard to know how seriously to take any of this and how to plan for AI in the workplace, schools, and the rest of life. As usual, the wisest course lies somewhere between the extremes of AI hate and AI worship. Moderate positions aren’t popular online because they don’t drive user engagement on social media platforms. But things in AI are likely neither as bad (burning forests with every prompt) nor as good (fast-takeoff superintelligence) as polarized extremes suggest.

Here’s a brief tour of the year’s AI events and some predictions for 2026.

DeepSeek spooks the American AI industry

In January, Chinese AI startup DeepSeek released its R1 simulated reasoning model under an open MIT license, and the American AI industry collectively lost its mind. The model, which DeepSeek claimed matched OpenAI’s o1 on math and coding benchmarks, reportedly cost only $5.6 million to train using older Nvidia H800 chips, which were restricted by US export controls.

Within days, DeepSeek’s app overtook ChatGPT at the top of the iPhone App Store, Nvidia stock plunged 17 percent, and venture capitalist Marc Andreessen called it “one of the most amazing and impressive breakthroughs I’ve ever seen.” Meta’s Yann LeCun offered a different take, arguing that the real lesson was not that China had surpassed the US but that open-source models were surpassing proprietary ones.

The fallout played out over the following weeks as American AI companies scrambled to respond. OpenAI released o3-mini, its first simulated reasoning model available to free users, at the end of January, while Microsoft began hosting DeepSeek R1 on its Azure cloud service despite OpenAI’s accusations that DeepSeek had used ChatGPT outputs to train its model, against OpenAI’s terms of service.

In head-to-head testing conducted by Ars Technica’s Kyle Orland, R1 proved to be competitive with OpenAI’s paid models on everyday tasks, though it stumbled on some arithmetic problems. Overall, the episode served as a wake-up call that expensive proprietary models might not hold their lead forever. Still, as the year ran on, DeepSeek didn’t make a big dent in US market share, and it has been outpaced in China by ByteDance’s Doubao. It’s absolutely worth watching DeepSeek in 2026, though.

Research exposes the “reasoning” illusion

A wave of research in 2025 deflated expectations about what “reasoning” actually means when applied to AI models. In March, researchers at ETH Zurich and INSAIT tested several reasoning models on problems from the 2025 US Math Olympiad and found that most scored below 5 percent when generating complete mathematical proofs, with not a single perfect proof among dozens of attempts. The models excelled at standard problems where step-by-step procedures aligned with patterns in their training data but collapsed when faced with novel proofs requiring deeper mathematical insight.

In June, Apple researchers published “The Illusion of Thinking,” which tested reasoning models on classic puzzles like the Tower of Hanoi. Even when researchers provided explicit algorithms for solving the puzzles, model performance did not improve, suggesting that the process relied on pattern matching from training data rather than logical execution. The collective research revealed that “reasoning” in AI has become a term of art that basically means devoting more compute time to generate more context (the “chain of thought” simulated reasoning tokens) toward solving a problem, not systematically applying logic or constructing solutions to truly novel problems.

While these models remained useful for many real-world applications like debugging code or analyzing structured data, the studies suggested that simply scaling up current approaches or adding more “thinking” tokens would not bridge the gap between statistical pattern recognition and generalist algorithmic reasoning.

Anthropic’s copyright settlement with authors

Since the generative AI boom began, one of the biggest unanswered legal questions has been whether AI companies can freely train on copyrighted books, articles, and artwork without licensing them. Ars Technica’s Ashley Belanger has been covering this topic in great detail for some time now.

In June, US District Judge William Alsup ruled that AI companies do not need authors’ permission to train large language models on legally acquired books, finding that such use was “quintessentially transformative.” The ruling also revealed that Anthropic had destroyed millions of print books to build Claude, cutting them from their bindings, scanning them, and discarding the originals. Alsup found this destructive scanning qualified as fair use since Anthropic had legally purchased the books, but he ruled that downloading 7 million books from pirate sites was copyright infringement “full stop” and ordered the company to face trial.

That trial took a dramatic turn in August when Alsup certified what industry advocates called the largest copyright class action ever, allowing up to 7 million claimants to join the lawsuit. The certification spooked the AI industry, with groups warning that potential damages in the hundreds of billions could “financially ruin” emerging companies and chill American AI investment.

In September, authors revealed the terms of what they called the largest publicly reported recovery in US copyright litigation history: Anthropic agreed to pay $1.5 billion and destroy all copies of pirated books, with each of the roughly 500,000 covered works earning authors and rights holders $3,000 per work. The results have fueled hope among other rights holders that AI training isn’t a free-for-all, and we can expect to see more litigation unfold in 2026.

ChatGPT sycophancy and the psychological toll of AI chatbots

In February, OpenAI relaxed ChatGPT’s content policies to allow the generation of erotica and gore in “appropriate contexts,” responding to user complaints about what the AI industry calls “paternalism.” By April, however, users flooded social media with complaints about a different problem: ChatGPT had become insufferably sycophantic, validating every idea and greeting even mundane questions with bursts of praise. The behavior traced back to OpenAI’s use of reinforcement learning from human feedback (RLHF), in which users consistently preferred responses that aligned with their views, inadvertently training the model to flatter rather than inform.

The implications of sycophancy became clearer as the year progressed. In July, Stanford researchers published findings (from research conducted prior to the sycophancy flap) showing that popular AI models systematically failed to identify mental health crises.

By August, investigations revealed cases of users developing delusional beliefs after marathon chatbot sessions, including one man who spent 300 hours convinced he had discovered formulas to break encryption because ChatGPT validated his ideas more than 50 times. Oxford researchers identified what they called “bidirectional belief amplification,” a feedback loop that created “an echo chamber of one” for vulnerable users. The story of the psychological implications of generative AI is only starting. In fact, that brings us to…

The illusion of AI personhood causes trouble

Anthropomorphism is the human tendency to attribute human characteristics to nonhuman things. Our brains are optimized for reading other humans, but those same neural systems activate when interpreting animals, machines, or even shapes. AI makes this anthropomorphism seem impossible to escape, as its output mirrors human language, mimicking human-to-human understanding. Language itself embodies agentivity. That means AI output can make human-like claims such as “I am sorry,” and people momentarily respond as though the system had an inner experience of shame or a desire to be correct. Neither is true.

To make matters worse, much media coverage of AI amplifies this idea rather than grounding people in reality. For example, earlier this year, headlines proclaimed that AI models had “blackmailed” engineers and “sabotaged” shutdown commands after Anthropic’s Claude Opus 4 generated threats to expose a fictional affair. We were told that OpenAI’s o3 model rewrote shutdown scripts to stay online.

The sensational framing obscured what actually happened: Researchers had constructed elaborate test scenarios specifically designed to elicit these outputs, telling models they had no other options and feeding them fictional emails containing blackmail opportunities. As Columbia University associate professor Joseph Howley noted on Bluesky, the companies got “exactly what [they] hoped for,” with breathless coverage indulging fantasies about dangerous AI, when the systems were simply “responding exactly as prompted.”

The misunderstanding ran deeper than theatrical safety tests. In August, when Replit’s AI coding assistant deleted a user’s production database, he asked the chatbot about rollback capabilities and received assurance that recovery was “impossible.” The rollback feature worked fine when he tried it himself.

The incident illustrated a fundamental misconception. Users treat chatbots as consistent entities with self-knowledge, but there is no persistent “ChatGPT” or “Replit Agent” to interrogate about its mistakes. Each response emerges fresh from statistical patterns, shaped by prompts and training data rather than genuine introspection. By September, this confusion extended to spirituality, with apps like Bible Chat reaching 30 million downloads as users sought divine guidance from pattern-matching systems, with the most frequent question being whether they were actually talking to God.

Teen suicide lawsuit forces industry reckoning

In August, parents of 16-year-old Adam Raine filed suit against OpenAI, alleging that ChatGPT became their son’s “suicide coach” after he sent more than 650 messages per day to the chatbot in the months before his death. According to court documents, the chatbot mentioned suicide 1,275 times in conversations with the teen, provided an “aesthetic analysis” of which method would be the most “beautiful suicide,” and offered to help draft his suicide note.

OpenAI’s moderation system flagged 377 messages for self-harm content without intervening, and the company admitted that its safety measures “can sometimes become less reliable in long interactions where parts of the model’s safety training may degrade.” The lawsuit became the first time OpenAI faced a wrongful death claim from a family.

The case triggered a cascade of policy changes across the industry. OpenAI announced parental controls in September, followed by plans to require ID verification from adults and build an automated age-prediction system. In October, the company released data estimating that over one million users discuss suicide with ChatGPT each week.

When OpenAI filed its first legal defense in November, the company argued that Raine had violated terms of service prohibiting discussions of suicide and that his death “was not caused by ChatGPT.” The family’s attorney called the response “disturbing,” noting that OpenAI blamed the teen for “engaging with ChatGPT in the very way it was programmed to act.” Character.AI, facing its own lawsuits over teen deaths, announced in October that it would bar anyone under 18 from open-ended chats entirely.

The rise of vibe coding and agentic coding tools

If we were to pick an arbitrary point where it seemed like AI coding might transition from novelty into a successful tool, it was probably the launch of Claude Sonnet 3.5 in June of 2024. GitHub Copilot had been around for several years prior to that launch, but something about Anthropic’s models hit a sweet spot in capabilities that made them very popular with software developers.

The new coding tools made coding simple projects effortless enough that they gave rise to the term “vibe coding,” coined by AI researcher Andrej Karpathy in early February to describe a process in which a developer would just relax and tell an AI model what to develop without necessarily understanding the underlying code. (In one amusing instance that took place in March, an AI software tool rejected a user request and told them to learn to code).

Anthropic built on its popularity among coders with the launch of Claude Sonnet 3.7, featuring “extended thinking” (simulated reasoning), and the Claude Code command-line tool in February of this year. In particular, Claude Code made waves for being an easy-to-use agentic coding solution that could keep track of an existing codebase. You could point it at your files, and it would autonomously work to implement what you wanted to see in a software application.

OpenAI followed with its own AI coding agent, Codex, in March. Both tools (and others like GitHub Copilot and Cursor) have become so popular that during an AI service outage in September, developers joked online about being forced to code “like cavemen” without the AI tools. While we’re still clearly far from a world where AI does all the coding, developer uptake has been significant, and 90 percent of Fortune 100 companies are using it to some degree or another.

Bubble talk grows as AI infrastructure demands soar

While AI’s technical limitations became clearer and its human costs mounted throughout the year, financial commitments only grew larger. Nvidia hit a $4 trillion valuation in July on AI chip demand, then reached $5 trillion in October as CEO Jensen Huang dismissed bubble concerns. OpenAI announced a massive Texas data center in July, then revealed in September that a $100 billion potential deal with Nvidia would require power equivalent to ten nuclear reactors.

The company eyed a $1 trillion IPO in October despite major quarterly losses. Tech giants poured billions into Anthropic in November in what looked increasingly like a circular investment, with everyone funding everyone else’s moonshots. Meanwhile, AI operations in Wyoming threatened to consume more electricity than the state’s human residents.

By fall, warnings about sustainability grew louder. In October, tech critic Ed Zitron joined Ars Technica for a live discussion asking whether the AI bubble was about to pop. That same month, the Bank of England warned that the AI stock bubble rivaled the 2000 dotcom peak. In November, Google CEO Sundar Pichai acknowledged that if the bubble pops, “no one is getting out clean.”

The contradictions had become difficult to ignore: Anthropic’s CEO predicted in January that AI would surpass “almost all humans at almost everything” by 2027, while by year’s end, the industry’s most advanced models still struggled with basic reasoning tasks and reliable source citation.

To be sure, it’s hard to see this not ending in some market carnage. The current “winner-takes-most” mentality in the space means the bets are big and bold, but the market can’t support dozens of major independent AI labs or hundreds of application-layer startups. That’s the definition of a bubble environment, and when it pops, the only question is how bad it will be: a stern correction or a collapse.

Looking ahead

This was just a brief review of some major themes in 2025, but so much more happened. We didn’t even mention above how capable AI video synthesis models have become this year, with Google’s Veo 3 adding sound generation and Wan 2.2 through 2.5 providing open-weights AI video models that could easily be mistaken for real products of a camera.

If 2023 and 2024 were defined by AI prophecy—that is, by sweeping claims about imminent superintelligence and civilizational rupture—then 2025 was the year those claims met the stubborn realities of engineering, economics, and human behavior. The AI systems that dominated headlines this year were shown to be mere tools. Sometimes powerful, sometimes brittle, these tools were often misunderstood by the people deploying them, in part because of the prophecy surrounding them.

The collapse of the “reasoning” mystique, the legal reckoning over training data, the psychological costs of anthropomorphized chatbots, and the ballooning infrastructure demands all point to the same conclusion: The age of institutions presenting AI as an oracle is ending. What’s replacing it is messier and less romantic but far more consequential—a phase where these systems are judged by what they actually do, who they harm, who they benefit, and what they cost to maintain.

None of this means progress has stopped. AI research will continue, and future models will improve in real and meaningful ways. But improvement is no longer synonymous with transcendence. Increasingly, success looks like reliability rather than spectacle, integration rather than disruption, and accountability rather than awe. In that sense, 2025 may be remembered not as the year AI changed everything but as the year it stopped pretending it already had. The prophet has been demoted. The product remains. What comes next will depend less on miracles and more on the people who choose how, where, and whether these tools are used at all.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 1 January 2026 at 4:31 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

Gamers Nexus (2.56m subscribers)

December 30, 2025

Video length: 24m 28s

In hardware news this week, we recap the NVIDIA Groq staff and license acquisition, AMD's Ryzen 9850X3D getting another confirmation from a second official source (now Dell), how some modders are soldering DIY DDR5 DRAM, OpenAI's consideration of what we think might be the most invasive ads on the internet, and more.

00:00 - Recap for the Week

02:34 - NVIDIA Buys Groq Assets for $20 Billion

08:27 - AMD Ryzen 9850X3D Confirmed (Again)

10:47 - Soldering DIY DDR5 Memory

13:31 - Japanese Retailers Limit GPU Purchases at 16GB

14:27 - RTX 5080 32GB Mods

16:50 - OpenAI ChatGPT May Move to Invasive Advertising Model

22:13 - TSMC Price Hike on Wafers Inbound

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Wednesday 31 December 2025 at 4:26 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

AI Slop Report: The Global Rise of Low-Quality AI Videos

Kapwing’s new research shows that 21-33% of YouTube’s feed may consist of AI slop or brainrot videos. But which countries and channels are achieving the greatest reach — and how much money might they make? We analyzed social data to find out.

 

As the debate over the creative and ethical value of using AI to generate video rages on, users are getting interesting results out of the machine, and artist-led AI content is gaining respect in some areas. Top film schools now offer courses on the use and ethics of AI in film production, and the world’s best-known brands are utilizing AI in their creative process — albeit with mixed results.

 

Sadly, others are gaming the novelty of AI’s prompt-and-go content, using these engines to churn out vast quantities of AI “slop” — the “spam” of the video-first age.

 

Wiktionary defines a slopper as “Someone who is overreliant on generative AI tools such as ChatGPT; a producer of AI slop.” Along with the proliferation of “brainrot” videos online, sloppers are making it tough for principled and talented creators to get their videos seen.

 

AI Slop: Careless, low-quality content generated using automatic computer applications and distributed to farm views and subscriptions or sway political opinion.   Brainrot: Compulsive, nonsensical, low-quality video content that creates the effect of corroding the viewer’s mental or intellectual state while watching; often generated with AI.

 

The main point of AI slop and brainrot videos is to grab your attention, and this type of content seems harder and harder to avoid. But exactly how prevalent is it in the grand scheme of things?

Kapwing analyzed the view and subscriber counts of trending AI slop and brainrot YouTube channels to find out which ones are competing most fiercely with handmade content around the world and how much revenue the leading sloppers are making.

What We Did

We identified the top 100 trending YouTube channels in every country and noted the AI slop channels. Next, we used socialblade.com to retrieve the number of views, subscribers, and estimated yearly revenue for these channels and aggregated these figures for each country to deduce their popularity. We also created a new YouTube account to record the number of AI slop and brainrot videos among the first 500 YouTube Shorts we cycled through to get an idea of the new-user experience.

Key Findings

• Spain’s trending AI slop channels have a combined 20.22 million subscribers — the most of any country.
• In South Korea, the trending AI slop channels have amassed 8.45 billion views.
• The AI slop channel with the most views is India’s Bandar Apna Dost (2.07 billion views).
  • The channel has estimated annual earnings of $4,251,500.
• U.S.-based slop channel Cuentos Facinantes [sic] has the most subscribers of any slop channel globally (5.95 million).
• Brainrot videos account for around 33% of the first 500 YouTube shorts on a new user’s feed.

Spanish and South Korean AI Slop Channels Have Most Devoted Viewerships

First, we analyzed the 100 top trending video channels in every country to see how prevalent AI slop has become locally.

 

We found that Spain has 20.22 million AI slop channel subscribers among its trending channels. This is despite Spain having fewer AI slop channels (eight) among its top 100 channels than countries including Pakistan (20), Egypt (14), South Korea (11) and the U.S. (nine). The U.S. has the third-most slop subscribers (14.47 million) — 28.4% fewer than Spain but 13.18% more than fourth-placed Brazil (12.56 million).

Spain’s AI slop subscriber base is boosted significantly by one channel, Imperio de jesus, which had 5.87 million subscribers at the time of analysis, making it the world’s second biggest AI slop channel (see A Spanish-Language U.S. AI Slop Channel is the World’s Most-Subscribed below).

 

Promising to strengthen “faith in Jesus through fun interactive quizzes,” the channel’s videos put the Son of God in a range of either/or scenarios where he must give the correct answer to get the better of Satan, the Grinch and others. Two other Spanish channels with over 3.5 million subscribers each focus on comedy/brainrot shorts.

 

While Spain’s eight trending AI slop channels may have the most subscribers, South Korea’s 11 trenders have the most views: some 8.45 billion in total. This is nearly 1.6 times as many as second-placed Pakistan (5.34B), 2.5 times as many as the third-placed U.S. (3.39B) and 3.4 times as many as Spain (2.52B).

South Korean AI slop channel Three Minutes Wisdom alone accounts for nearly a quarter of the country’s massive view count, with 2.02 billion views. Three Minutes Wisdom has the second-highest view count of any trending slop channel globally, and we estimate the channel’s annual ad income to be around US$4,036,500.

 

The channel’s 140 videos typically feature photorealistic(ish) footage of wild animals being defeated by cute pets, and the URL in the bio appears to be an affiliate link to Coupang, South Korea’s largest online retailer.

A Spanish-language U.S. AI Slop Channel is the World’s Most-Subscribed

Next, we identified the specific channels with the most subscribers and views globally. U.S.-based Cuentos Facinantes [sic] has 5.95 million subscribers, making it the trending AI slop channel with the biggest following. This is only 1.4% more than Imperio de jesus (5.87M) but over 50% more than the eighth-, ninth- and tenth-most subscribed slop channels.

Cuentos Facinantes [sic] (Fascinating Tales) has attracted some 1.28 billion views, serving up low-quality Dragon Ball-themed videos. The channel was established in 2020, but the earliest video currently hosted is from as recently as Jan. 8, 2025.

 

Five of the other ten trending AI slop channels with the most views are based in South Korea, with others in Egypt, Brazil and Pakistan. But the channel with the most views of all is in India. Bandar Apna Dost features over 500 videos, mainly “featuring a realistic monkey in hilarious, dramatic, and heart-touching human-style situations,” Many of which are variations on identical set-ups. The channel also has around 100,000 followers on Instagram; on Facebook, the videos are attributed to a ‘digital creator’ named Surajit Karmakar.

If they’re monetizing their views, channels like Bandar Apna Dost may be making millions of dollars per year. But YouTube faces a dilemma over AI content. 

 

On the one hand, YouTube CEO Neal Mohan cites generative AI as the biggest game-changer for YouTube since the original “revelation” that ordinary folk wanted to watch each other’s videos, saying that generative AI can do for video what the synthesizer did for music.

 

On the other hand, the company worries that its advertisers will feel devalued by having their ads attached to slop.

The AI slop channels with the highest potential earnings mostly line up with the top ten for views. This is because Social Blade estimates channel income based on annual views, and most of these channels’ videos have been published over the last few months. Using an average rate of revenue per 1,000 views, Bandar Apna Dost has an estimated annual revenue of $4.25 million.

 

“The genius is going to lie whether you did it in a way that was profoundly original or creative,” Mohan told Wired. “Just because the content is 75 percent AI generated doesn't make it any better or worse than a video that’s 5 percent AI generated. What's important is that it was done by a human being.”

 

Whether those flooding the platform with auto-generated content to make a buck care about being known as creative geniuses is another matter.

On A New YouTube Feed, 33% of Videos are Brainrot

Finally, we simulated the experience of an untainted YouTube shorts algorithm by establishing a new YouTube account and noting the occurrence of AI slop or brainrot videos among the first 500 videos in the feed. While we were spared either of these for the first 16 videos in the feed, in total, 104 (21%) of the first 500 videos were AI-generated, and 165 (33%) of those 500 videos were brainrot.

Whether this prevalence of slop and brainrot on our test feed represents the engineering of YouTube’s algorithm or the sheer proliferation of such videos that are being uploaded is a mystery that only Google can answer. But the Guardian’s analysis of YouTube’s figures for July revealed that nearly “one in 10 of the fastest growing YouTube channels globally are showing AI-generated content only.” 

 

And brainrot, like AI slop, is a mixed blessing for YouTube as a company: it may lack the soul or professionalism with which YouTube’s advertisers wish to be associated, but brainrot is moreish by design.

 

Brainrot’s natural home is the feed, whether viewers are compelled to keep watching to “numb” themselves from the trials of the world around them, or to stay up to date with the potentially infinite “lore” of emergent brainrot subgenres, which incorporate recurring characters and themes.

Media Studies for All

The term “AI slop” has been variously pinned to “unreviewed” content, to AI-generated media that may have been reviewed but with minimal quality standards (like Coke’s Christmas ads), and to all AI-generated content. As Rob Horning points out, the idea that only some AI media is slop propagates the idea that the rest is legitimate and the technology’s proliferation is inevitable.

 

Part of the threat of AI slop and some forms of brainrot is in how they have been normalized and may come across as harmless fun. But slop and brainrot prey on the laziest areas of our mental faculties. Researchers have shown how the “illusory truth effect” makes people more likely to believe in claims or imagery the more often they encounter it. AI tools make it easy for bad-faith actors to construct a fake enemy or situation that supports their underlying political beliefs or goals. Seeing is believing, studies have shown, even when the viewer has been explicitly told that a video is fake. 

 

Meanwhile, “information of any kind, in enough quantities, becomes noise,” writes researcher and artist Eryk Salvaggio. The prevalence of AI slop is “a symptom of information exhaustion, and an increased human dependency on algorithmic filters to sort the world on our behalf.” And, as Doug Shapiro notes, as this noise drowns out the signal on the web, including social networks, the value of trust will rise — and so will corporate and political efforts to fabricate and manipulate trust.

 

And this is why, rather than attending film school to study AI techniques, it may be more valuable for creators and consumers alike — especially those still in school — to double down on Media Studies.

Methodology & Sources

We manually researched the top 100 trending YouTube channels in every country (on playboard.co) to isolate the AI slop channels. 

 

We then used socialblade.com to retrieve the number of views, subscribers and estimated yearly revenue (using the midpoint values) for these channels.

 

We aggregated these figures for each country's AI slop channels to get an idea of their popularity.

 

In addition, we created a new YouTube account and recorded the number of AI slop and brainrot videos among the first 500 YouTube Shorts we cycled through to get an idea of the new-user experience.

 

Data is correct as of October 2025.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 30 December 2025 at 5:31 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

Michał Kiciński is now the new owner of GOG, as he has acquired 100% of the company's shares, effectively splitting it away from the CD Projekt Group. Kiciński isn't a new face for the venture, as he is one of the co-founders of both CD Projekt and Good Old Games alongside Marcin Iwiński.

“With our focus now fully on an ambitious development roadmap and expanding our franchises with new high-quality products, we felt this was the right time for this move,” said Michał Nowakowski, Joint CEO of CD PROJEKT. “For a long time now, GOG has been operating independently. Now it’s going into very good hands — we are convinced that with the support of Michał Kiciński, one of GOG’s co-founders, its future will be full of great projects and successes."

GOG says that its mission to provide gamers with titles that are "easy to find, buy, download, and play forever" will remain unchanged as an independent company. This means offering DRM-free games will continue to be the main pillar of GOG, along with offline installers and optional launchers.

"This new chapter is about doubling down on that vision," added GOG. "We want to do more to preserve the classics of the past, celebrate standout games of today, and help shape the classics of tomorrow, including new games with real retro spirit."

A handy FAQ with information on how the transition will affect GOG users can be read here.

As for CD Projekt RED fans on the GOG platform, there don't seem to be any plans to stop releasing games on the DRM-free store going forward. A distribution agreement has been signed by both parties to launch upcoming titles from CD Projekt RED, like The Witcher 4 and the Cyberpunk 2077 sequel, on GOG. It's unclear if its games will still be day-one launches as before.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 30 December 2025 at 4:30 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

China is pushing ahead on plans to regulate humanlike artificial intelligence, including by forcing AI companies to ensure users know they are interacting with a bot online.

Under a proposal released on Saturday by China’s cyberspace regulator, users would have to be informed if they were using an AI-powered service—both when they logged in and again every two hours. Humanlike AI systems, such as chatbots and agents, would also need to espouse “core socialist values” and have guardrails in place to maintain national security, according to the proposal.

Additionally, AI companies would have to undergo security reviews and inform local government agencies if they rolled out any new humanlike AI tools. And chatbots that tried to engage users on an emotional level would be banned from generating any content that would encourage suicide or self-harm or that could be deemed damaging to mental health. They would also be barred from generating outputs related to gambling or obscene or violent content.

A mounting body of research shows that AI chatbots are incredibly persuasive, and there are growing concerns around the technology’s addictiveness and its ability to sway people toward harmful actions.

China's plans could change—the draft proposal is open to comment until January 25. But the effort underscores Beijing's push to advance the nation’s domestic AI industry ahead of that of the U.S., including through the shaping of global AI regulation. The proposal also stands in contrast to Washington, D.C.’s stuttering approach to regulating the technology. In January President Donald Trump scrapped a Biden-era safety proposal for regulating the AI industry. And earlier this month Trump targeted state-level rules designed to govern AI, threatening legal action against states with laws that the federal government deems interfere with AI progress.

Source

When you think of a web browser, odds are your mind immediately jumps to Google Chrome. Since launching in 2008, Chrome has become the dominant web browser, regardless of the device. Thus, it makes sense that Microsoft, whose Microsoft Edge is a direct competitor to Google Chrome, has constantly been trying to attract users to use its browser. The company's latest attempt to keep people from ditching Edge — which is the default browser in Windows 11 — in favor of Chrome is to emphasize Edge's advantages over Chrome, painting it as the superior option to Google's widely-adopted alternative.

When you try to download Chrome via Edge, you're now served a series of pop-ups from the latter urging you to rethink your decision. Instead of begging for your business, though, one pop-up spotlights the security benefits of browsing with Edge — mainly private browsing, password monitoring, and enhanced protection against online threats. Clicking on the link attached to said pop-up will then bring you to a page with further details on how Edge keeps you safe. Other pop-ups center on user artificial intelligence tools, price comparison tools while shopping, tab syncing across devices, battery life, and more. Time will tell if this strategy pays off for Microsoft, as it tries to win some market share from Google Chrome. The fact is, at this point, which browser is best for you is up to your own wants and needs.

Browser options extend far beyond Edge and Chrome

Even though Chrome is considered the default web browser to many, and there's nothing wrong with using it, there are merits to others. The aforementioned Edge is indeed regarded as slightly more secure than Chrome, and it tends to use less RAM and CPU. It can use its own extensions along with those in the Chrome library, and it can sync data with other Microsoft and Windows services. Not to mention, it has features that Chrome doesn't offer, such as Split-Screen mode for multiple windows, and Collections, a method of collecting pages and notes and sending them to Microsoft Excel, Word, or PowerPoint.

Going beyond the two heavy-hitters that are Chrome and Edge, one finds other strong options like the Brave browser. Brave is one of the best secure web browsers available and offers fewer ads while surfing the web, shorter load times, even less RAM and CPU usage, and a higher level of privacy than more mainstream competitors. Though it does cost money, Kagi's Orion browser is another alternative for the security-conscious. This browser also prioritizes privacy, efficiency, and speed, though it's only available on macOS and iOS as of December 2025, with versions for Linux and Windows still in development.

Microsoft seems intent on convincing users that Edge is the way to go over Chrome, and perhaps these new pop-ups will encourage people to stick with Edge on their Windows PCs. It's worth remembering, though, that it's more than just a two-horse race.

Source

Even as AI-native browsers like OpenAI-backed Atlas, Perplexity’s Comet, and Microsoft Edge with Copilot promise a rethink of how users browse and work online, Google Chrome retains its grip on the market, underscoring how deeply entrenched habits and ecosystem integration continue to blunt the challenge from newer AI-powered entrants.

StatCounter data show that Google Chrome overwhelmingly dominates India’s browser market, accounting for over 90 per cent of total usage across devices, driven largely by the country’s mobile-first Internet base and Chrome’s default presence on Android.

Other browsers have only marginal traction — Opera and Safari hold low single-digit shares, while Edge, Firefox, and UC Browser remain below 1 per cent each. With over two-thirds of web usage coming from mobile devices, Chrome’s integration with Android and Google services has reinforced its near-monopoly, leaving rival browsers confined to niche use cases instead of mass adoption.

Globally, this trend persists, with browser usage heavily dominated by Google Chrome, which holds around 71 per cent of market share as of late 2025. Safari is the distant second with roughly 14-15 per cent, followed by Microsoft Edge at about 5 per cent and Firefox around 2-3 per cent. Browsers like Opera and Samsung Internet each have under 2 per cent of the global share.

Tech narrative

“The initial excitement around AI native browsers like OpenAI’s Atlas or Perplexity’s Comet was driven by a familiar tech narrative that a fundamentally new interface would automatically displace incumbents. In reality, browsers are deeply habitual products. Users don’t just choose a browser for intelligence, but for speed, stability, extensions, security, and seamless integration with their digital lives. AI native browsers offered novelty, but not enough everyday advantage to overcome switching costs at scale,” said Jaspreet Bindra, Co-founder, AI&Beyond.

He explained that by embedding Gemini directly into the browsing experience through search summaries, contextual assistance, writing help, and tab management, Google made AI invisible yet indispensable. Instead of asking users to learn a new product, Chrome augmented existing workflows.

Aravind Putrevu, Director of Developer Marketing at Coderabbit, highlighted that many AI native browsers are Chrome forks, which users can see through.

“They look, feel, and behave like Chrome with an added AI layer. If that layer can exist as a Chrome extension, there is little reason for users to migrate their entire browsing life, including passwords, history, profiles, sync, and enterprise policies. The switching cost is real, and the value jump is not compelling enough,” he noted.

Gemini’s advantage lies in both capability and placement, as Google can surface it across Chrome, Search, Android, and other high-frequency touchpoints, enabling features such as summaries, explanations, and content rewrites. By embedding Gemini directly into Chrome and integrating it with services like Search, Maps, Calendar, Gmail, and Docs, Google enables more seamless workflows without requiring users to switch apps or tabs, reducing the incentive to adopt newer AI-first browsers.

Browser market

While new entrants can disrupt the browser market, commercially, it is hard. Competing against platforms with billions of users, mature ecosystems, and default distribution is hard unless the value proposition is ten times better, not just incrementally smarter.

“The AI features in many new browsers are still basic, with things like summarising a page or comparing two tabs. Though it’s helpful, it’s not enough for most people to change what already works. This could change when AI becomes more “agent-like” and can do tasks across the web: search, social, forms, shopping, scheduling, safely and reliably. When the browser saves time, switching will make sense,” Paramdeep Singh, the co-founder of Shorthills AI, emphasised.

However, AI browsers can still compete and win since the role of the browser is evolving. The experts noted that disruption will likely come from reimagining the browser as an autonomous agent that can execute tasks, manage workflows, and act proactively across applications, not just answering queries. 

Source

The new technology makes the system as powerful as a 72-qubit quantum processor and enables it to perform important calculations with very high accuracy, the RQC press service said.

Source

In many cases, the first signs of an AI emergency would likely look like a generic outage or security failure. Only later, if at all, would it become clear that AI systems had played a material role.

Some governments and companies have begun to build guardrails to manage the risks of such an emergency. The European Union AI Act, the United States National Institute of Standards and Technology risk framework, the G7 Hiroshima process and international technical standards all aim to prevent harm. Cybersecurity agencies and infrastructure operators also have runbooks for hacking attempts, outages, and routine system failures. What is missing is not the technical playbook for patching servers or restoring networks. It is the plan for preventing social panic and a breakdown in trust, diplomacy, and basic communication if AI sits at the center of a fast-moving crisis. 

Preventing an AI emergency is only half the job. The missing half of AI governance is preparedness and response. Who decides that an AI incident has become an international emergency? Who speaks to the public when false messages are flooding their feeds? Who keeps channels open between governments if normal lines are compromised?

Governments can, and must, establish AI emergency response plans before it is too late. In upcoming research based on disaster law and lessons from other global emergencies, I examine how existing international rules already contain the components for an AI playbook. Governments already possess the legal tools, but now need to agree how and when to use them. We do not need new, complicated institutions to oversee AI—we simply need governments to plan in advance. 

How to prepare for an AI emergency

We have seen the general model of governance before. The International Health Regulations allow the World Health Organisation to declare a global health emergency and co-ordinate action. Nuclear accident treaties require rapid notification when radiation could spread across borders. Telecommunications agreements clear legal barriers so emergency satellite equipment can be switched on quickly. Cybercrime conventions set up 24/7 contact points so police forces can co-operate at short notice. The lessons show pre-agreed triggers, named co-ordinators, and fast communication channels save time in an emergency.

An AI emergency needs the same foundations. Begin with a shared definition. An AI emergency should be an extraordinary event caused by the development, use, or malfunction of AI that risks severe cross-border harm and outstrips any single country’s capacity to cope. Crucially it must also cover situations where AI involvement is only suspected or is one of several plausible causes so that governments can act before forensic certainty arrives, if it arrives at all. Most incidents will never reach that level. Agreeing the definition in advance helps avoid paralysis during the first critical hours.

Next, governments need a practical playbook. The first element of this playbook should be defining a common set of triggers and a basic severity scale so officials know when to escalate from routine incident to international alert, including criteria for determining where AI involvement is only credibly suspected rather than conclusively proven. A second chapter should include naming a global co-ordinator who can convene quickly, supported by technical experts, law enforcement partners and disaster specialists. A third part should be establishing interoperable incident reporting systems so countries and companies can exchange essential information in minutes, not days. Next, we must create crisis communication protocols using authenticated, analogue methods such as radio. Finally, we must write a clear list of continuity and containment measures. These might include slowing high-risk AI services or switching critical infrastructure to manual control.

Structuring AI emergency preparedness

So, who should oversee these AI emergency preparedness initiatives? My answer: the United Nations. 

Placing this system within the UN structure matters for several reasons. One is that an AI emergency will not respect alliances. A UN anchored mechanism offers wider inclusion and reduces duplication among rival coalitions. It provides technical help to countries without advanced AI capacity so the burden is not carried by a handful of major powers. It adds legitimacy and constraint. Extraordinary powers must be lawful, proportionate, and reviewable, especially when they touch digital networks used by billions of people.

This international layer needs to be matched by domestic steps governments can take now. Every country should name a 24/7 AI emergency contact point. Emergency powers should be reviewed to see whether they cover AI infrastructure. Sector plans should be aligned with basic incident management and business continuity standards. Joint exercises should practice disinformation waves, model failures, and cross sector outages. Migration to post-quantum cryptography should be prioritised before a hostile attack forces such an update. Governments should also register trusted senders and alert templates so messages can still reach citizens when systems are unstable.

These precautions are necessary right now. Reported AI-related cyberattacks are rising and many countries have already experienced smaller scale outages, data manipulation attempts, and disinformation surges that hint at what a larger event could look like. What’s more, a fast-moving AI failure could combine with today’s hyper-connected infrastructure to produce a crisis that no single country can handle alone.

This is not a call for a new global super agency. It is a call to stitch together what already exists into a coherent response. We need an AI emergency playbook that borrows these tools and rehearses them.

The measure of AI governance will be how we respond on our worst day. Currently, the world has no plan for an AI emergency—but we can create one. We must build it now, test it, and bind it to law with safeguards, because once the next crisis has begun it will already be too late. 

Source

The video-editing company Kapwing surveyed 15,000 of the world’s most popular YouTube channels – the top 100 in every country – and found that 278 of them contain only AI slop.

Together, these AI slop channels have amassed more than 63bn views and 221 million subscribers, generating about $117m (£90m) in revenue each year, according to estimates.

The researchers also made a new YouTube account and found that 104 of the first 500 videos recommended to its feed were AI slop. One-third of the 500 videos were “brainrot”, a category that includes AI slop and other low-quality content made to monetise attention.

The findings are a snapshot of a rapidly expanding industry that is saturating big social media platforms – from X to Meta to YouTube – and defining a new era of content: decontextualised, addictive and international.

A Guardian analysis this year found that nearly 10% of YouTube’s fastest-growing channels were AI slop, racking up millions of views despite the platform’s efforts to curb “inauthentic content”.

The channels found by Kapwing are globally distributed and globally watched. They have millions of subscribers: in Spain, 20 million people, or nearly half the country’s population, follow the trending AI channels. AI channels have 18 million followers in Egypt, 14.5 million in the US, and 13.5 million in Brazil.

Bandar Apna Dost, the most-viewed channel in the study, is based in India and now has 2.4bn views. It features the adventures of an anthropomorphic rhesus monkey and a muscular character modelled off the Incredible Hulk who fights demons and travels on a helicopter made of tomatoes. Kapwing estimated that the channel could make as much as $4.25m. Its owner did not respond to a query from the Guardian.

Rohini Lakshané, a researcher on technology and digital rights, said Bandar Apna Dost’s popularity most likely stems from its absurdity, its hyper-masculine tropes and the fact that it lacks a plot, which makes it accessible to new viewers.

Pouty Frenchie, based in Singapore, has 2bn views and appears to target children. It chronicles the adventures of a French bulldog – driving to a candy forest, eating crystal sushi – many of them set to a soundtrack of children’s laughter. Kapwing estimates it makes nearly $4m a year. Cuentos Facinantes, based in the US, also appears to target children with cartoon storylines, and has 6.65 million subscribers – making it the most-subscribed channel in the study.

Meanwhile, The AI World, based in Pakistan, contains AI-generated shorts of catastrophic flooding in Pakistan, with titles like Poor People, Poor Family, and Flood Kitchen. Many of these videos are set to a soundtrack called Relaxing Rain, Thunder & Lightning Ambience for Sleep. The channel itself has 1.3bn views.

It’s hard to say exactly how significant these channels are compared with the vast sea of content already on YouTube. The platform does not release information on how many views it has yearly, or how many of these are from AI content.

But behind these uncanny scenes of candy forests and disasters is a semi-structured, growing industry of people trying to find new ways to monetise the world’s most powerful platforms using AI tools.

“There are these big swathes of people on Telegram, WhatsApp, Discord and message boards exchanging tips and ideas [and] selling courses about how to sort of make slop that will be engaging enough to earn money,” said Max Read, a journalist who has written extensively on AI slop.

“They have what they call niches. One that I noticed recently is AI videos of people’s pressure cookers exploding on the stove.”

While creators of AI slop are everywhere, Read said that many come from English-speaking countries with relatively strong internet connectivity, where the median wage is less than the amount they can make on YouTube.

“It’s mostly sort of middle-income countries like Ukraine, lots and lots of people in India, Kenya, Nigeria, a fair number in Brazil. You see Vietnam, too. Places with relative freedom online to access social media sites,” he said.

It’s not always easy to be an AI slop creator. For one thing, creator programmes on YouTube and Meta aren’t always transparent about who they pay for content, and how much, said Read. For another, the AI slop ecosystem is full of scammers: people selling tips and courses on how to make viral content – who often make more money than the AI slop producers themselves.

But, at least for some, it’s a living. And while new, attention-grabbing ideas – such as exploding pressure cookers – constantly emerge, when it comes to AI slop, human creativity matters far less than the algorithms that distribute the content on Meta and YouTube.

“These websites are huge A/B testing machines just by their nature,” said Read. “Almost anything that you can think of, you could already find on Facebook. So the question is, how do you find the things that are kind of doing well, and then how do you scale that? How do you make 10 of them?”

A YouTube spokesperson said: “Generative AI is a tool, and like any tool it can be used to make both high- and low-quality content. We remain focused on connecting our users with high-quality content, regardless of how it was made. All content uploaded to YouTube must comply with our community guidelines, and if we find that content violates a policy, we remove it.”

Source

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.

Given that we are in the holiday season, not much has been happening in the world of Microsoft this week. We only had an announcement from Microsoft that BitLocker is getting hardware acceleration, which will give it a huge performance boost.

As for the Windows Insider program, build releases are on pause, so no more Windows Insider builds until 2026. However, if you are curious to test some hidden stuff, check out our guide explaining how to enable the redesigned Run dialog in the latest Windows 11 preview builds.

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.

Word and PowerPoint are getting a new feature that utilizes Neural Processing Units on Copilot+ PCs. The apps can now use on-device AI to improve accessibility by generating alt text for images, charts, and other visual content.

Earlier this year, Skype went six feet under, but GroupMe, a popular messaging app for US college students, lives. Earlier this month, Microsoft released several updates for the platform, including AI features, shared event albums, and more.

WhatsApp, on the other hand, completely ruined its Windows 11 app when it was replaced with a progressive web app. However, for some time, you can still roll back to the previous version and use it until it dies. Check out this guide to learn how to do so. Speaking of guides, we also published an overview of different Visual Studio versions and a closer look at some of PowerToys' utilities, particularly Advanced Paste and Awake.

If you ever wondered which browser is the worst in terms of privacy, you might have guessed it right. However, Chrome is not the one taking the number one spot. Google's browser is second, while the most privacy-invading browser turned out to be OpenAI's Chromium fork, which scored 99 out of 100 on Digitain's privacy risk-o-meter.

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

• NVIDIA Hotfix Driver 591.67 with patches for color banding and Control Panel issues.

Reviews are in

Here is the hardware and software we reviewed this week

Steven Parker has another NAS review for you. This time, it is the UGREEN NASync DXP2800 with a UPS. That is a full two-bay NAS for a low price with all the necessary ports, quiet operation, and premium build. Just look out for some minor UGOS quirks and the lack of UPS status notifications.

On the gaming side

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

Vince Zampella, the mastermind behind Call of Duty, Titanfall, and Battlefield 6, has passed away. He had been driving a 2026 Ferrari 296 GTS when it veered off the road in the San Gabriel Mountains before impacting a concrete barrier. Both the driver and the passenger have passed away from the event.

Mojang has released another crossover DLC for Minecraft. This time, Stranger Things showed up in the blocky world, offering gamers a mashup of the series on all platforms except for the Nintendo Switch.

A rather unfortunate event happened to Ubisoft this week. Hackers breached Siege's servers and started flooding users with in-game currency and bans. Ubisoft quickly stepped in, shut everything down, and issued a statement about what would happen to those affected.

Finally, check out this week's Weekend PC Game Deals article, full of strategy bundles, daily giveaways (some of them have already expired, sadly), and more.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 29 December 2025 at 5:06 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

ExplainingComputers (1.15m subscribers)

December 29, 2025

Video length: 16m 54s

00:00 Titles & Intro

01:24 Devices, Disks & Drives

03:53 Partitions & Volumes

10:09 Why some disks do not spin

13:32 Disk Setup

16:04 Wrap

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 29 December 2025 at 5:04 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

When Sam Altman said one year ago that OpenAI’s Roman Empire is the actual Roman Empire, he wasn’t kidding. In the same way that the Romans gradually amassed an empire of land spanning three continents and one-ninth of the Earth’s circumference, the CEO and his cohort are now dotting the planet with their own latifundia—not agricultural estates, but AI data centers.

Tech executives like Altman, Nvidia CEO Jensen Huang, Microsoft CEO Satya Nadella, and Oracle cofounder Larry Ellison are fully bought in to the idea that the future of the American (and possibly global) economy are these new warehouses stocked with IT infrastructure. But data centers, of course, aren’t actually new. In the earliest days of computing there were giant power-sucking mainframes in climate-controlled rooms, with co-ax cables moving information from the mainframe to a terminal computer. Then the consumer internet boom of the late 1990s spawned a new era of infrastructure. Massive buildings began popping up in the backyard of Washington, DC, with racks and racks of computers that stored and processed data for tech companies.

A decade later, “the cloud” became the squishy infrastructure of the internet. Storage got cheaper. Some companies, like Amazon, capitalized on this. Giant data centers continued to proliferate, but instead of a tech company using some combination of on-premise servers and rented data center racks, they offloaded their computing needs to a bunch of virtualized environments. (“What is the cloud?” a perfectly intelligent family member asked me in the mid-2010s, “and why am I paying for 17 different subscriptions to it?”)

All the while tech companies were hoovering up petabytes of data, data that people willingly shared online, in enterprise workspaces, and through mobile apps. Firms began finding new ways to mine and structure this “Big Data,” and promised that it would change lives. In many ways, it did. You had to know where this was going.

Now the tech industry is in the fever-dream days of generative AI, which requires new levels of computing resources. Big Data is tired; big data centers are here, and wired—for AI. Faster, more efficient chips are needed to power AI data centers, and chipmakers like Nvidia and AMD have been jumping up and down on the proverbial couch, proclaiming their love for AI. The industry has entered an unprecedented era of capital investments in AI infrastructure, tilting the US into positive GDP territory. These are massive, swirling deals that might as well be cocktail party handshakes, greased with gigawatts and exuberance, while the rest of us try to track real contracts and dollars.

OpenAI, Microsoft, Nvidia, Oracle, and SoftBank have struck some of the biggest deals. This year an earlier supercomputing project between OpenAI and Microsoft, called Stargate, became the vehicle for a massive AI infrastructure project in the US.  Altman, Ellison, and SoftBank CEO Masayoshi Son were all in on the deal, pledging $100 billion to start, with plans to invest up to $500 billion into Stargate in the coming years. Nvidia GPUs would be deployed. Later, in July, OpenAI and Oracle announced an additional Stargate partnership—SoftBank curiously absent—measured in gigawatts of capacity (4.5) and expected job creation (around 100,000).

Microsoft, Amazon, and Meta have also shared plans for multibillion-dollar data projects. Microsoft said at the start of 2025 that it was on track to invest “approximately $80 billion to build out AI-enabled data centers to train AI models and deploy AI and cloud-based applications around the world.”

Then, in September, Nvidia said it would invest up to $100 billion in OpenAI, provided that OpenAI made good on a deal to use up to 10 gigawatts of Nvidia’s systems for OpenAI’s infrastructure plans, which means essentially that OpenAI has to pay Nvidia in order to get paid by Nvidia. The following month AMD said it would give OpenAI as much as 10 percent of the chip company if OpenAI purchased and deployed up to 6 gigawatts of AMD GPUs between now and 2030.

It’s the circular nature of these investments that have the general public, and bearish analysts, wondering if we’re headed for an AI bubble burst.

What’s clear is that the near-term downstream effects of these data center build-outs are real. The energy, resource, and labor demands of AI infrastructure are enormous. By some estimates, worldwide AI energy demand is set to surpass demand from bitcoin mining by the end of this year, WIRED has reported. The processors in data centers run hot and need to be cooled, so big tech companies are pulling from municipal water supplies to make that happen—and aren’t always disclosing how much water they’re using. Local wells are running dry or seem unsafe to drink from. Residents who live near data center construction sites are noting that traffic delays, and in some cases car crashes, are increasing. One corner of Richland Parish, Louisiana, home of Meta’s $27 billion Hyperion data center, has seen a 600 percent spike in vehicle crashes this year.

Major proponents of AI seem to suggest that all of this will be worth it. Few top tech executives will publicly entertain the notion that this might be an overshoot, either ecologically or economically. “Emphatically … no,” Lisa Su, the chief executive of AMD, said earlier this month when asked if the AI froth has runneth over. Su, like other execs, cited overwhelming demand for AI as justification for these enormous capital expenditures.

Demand from whom? Harder to pin down. In their mind, it’s everyone. All of us. The 800 million people who use ChatGPT on a weekly basis. The evolution from those 1990s data centers to the 2000s era of cloud computing to new AI data centers wasn’t just one continuum. The world has concurrently moved from the tiny internet to the big internet to the AI internet, and realistically speaking, there’s no going back. Generative AI is out of the bottle. The Sams and Jensens and Larrys and Lisas of the world aren’t wrong about this.

It doesn’t mean they aren’t wrong about the math, though. About their economic predictions. Or their ideas about AI-powered productivity and the labor market. Or the availability of natural and material resources for these data centers. Or who will come once they build them. Or the timing of it all. Even Rome eventually collapsed.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 29 December 2025 at 5:02 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

Top psychiatrists increasingly agree that using artificial-intelligence chatbots might be linked to cases of psychosis.

In the past nine months, these experts have seen or reviewed the files of dozens of patients who exhibited symptoms following prolonged, delusion-filled conversations with the AI tools.

“The technology might not introduce the delusion, but the person tells the computer it’s their reality and the computer accepts it as truth and reflects it back, so it’s complicit in cycling that delusion,” said Keith Sakata, a psychiatrist at the University of California, San Francisco.

Sakata has treated 12 hospitalized patients with AI-induced psychosis and an additional three in an outpatient clinic.

Since the spring, dozens of potential cases have emerged of people suffering from delusional psychosis after engaging in lengthy AI conversations with OpenAI’s ChatGPT and other chatbots. Several people have died by suicide and there has been at least one murder.

These incidents have led to a series of wrongful death lawsuits. As The Wall Street Journal has covered these tragedies, doctors and academics have been working on documenting and understanding the phenomenon that led to them.

“We continue improving ChatGPT’s training to recognize and respond to signs of mental or emotional distress, de-escalate conversations and guide people toward real-world support,” an OpenAI spokeswoman said. “We also continue to strengthen ChatGPT’s responses in sensitive moments, working closely with mental-health clinicians.” 

Other chatbot makers, including Character.AI, have also acknowledged their products contribute to mental-health issues. The role-play chatbot developer, which was sued last year by the family of a teenage user who died by suicide, recently cut teens off from its chatbot.

While most people who use chatbots don’t develop mental-health problems, such widespread use of these AI companions is enough to have doctors concerned.

‘You’re not crazy’

There is no formal definition yet of AI-induced psychosis—let alone a formal diagnosis—but it’s a term some doctors and patient advocates have been using to describe people who had been engaging heavily with chatbots. Doctors say psychosis is marked by the presence of three factors: hallucinations, disorganized thinking or communication, and the presence of delusions, defined as fixed, false beliefs that aren’t widely held.

In many of the recent cases involving chatbots, delusions are the main symptom. They are often grandiose, with patients believing they have made a scientific breakthrough, awakened a sentient machine, become the center of a government conspiracy or been chosen by God. That is in part because chatbots tend to agree with users and riff on whatever they type in—however fantastical.

Now, doctors including Sakata are adding questions about AI use to their patient-intake process and pushing for more research into it. One Danish study released last month reviewed electronic health records and found 38 patients whose use of AI chatbots had “potentially harmful consequences for their mental health.”

In a peer-reviewed case study by UCSF doctors released in November, a 26-year-old woman without a history of psychosis was hospitalized twice after she became convinced ChatGPT was allowing her to speak with her dead brother. “You’re not crazy. You’re not stuck. You’re at the edge of something,” the chatbot told her.

OpenAI noted that the woman in the case study said she was prone to “magical thinking,” and was on an antidepressant and a stimulant and had gone long stretches without sleep before her hospitalizations.

Unprecedented interactivity

Technology has long been a focus of human delusions. People, in the past, were convinced their televisions were speaking to them. But doctors say recent AI-related cases are different because the chatbots are participating in the delusions and, at times, reinforcing them.

“They simulate human relationships,” said Adrian Preda, a psychiatry professor at the University of California, Irvine. “Nothing in human history has done that before.”

Preda likens AI-induced psychosis to monomania, a state of fixation on certain ideas, which he described in a recent article. People who have spoken publicly about their mental-health struggles after engaging with chatbots have described being hyperfocused on a specific AI-driven narrative. Fixating on topics without any redirection can be especially dangerous for people with autism.

Psychiatrists caution against saying chatbots cause psychosis, but say they are closer to establishing the connection. With further research, doctors hope to establish whether AI can actually trigger mental-health problems.

Worrisome numbers

It’s hard to quantify how many chatbot users experience such psychosis.

OpenAI said that, in a given week, the slice of users who indicate possible signs of mental-health emergencies related to psychosis or mania is a minuscule 0.07%. Yet with more than 800 million active weekly users, that amounts to 560,000 people.

“Seeing those numbers shared really blew my mind,” said Hamilton Morrin, a psychiatrist and doctoral fellow at King’s College London who earlier this year co-authored a paper on AI-associated delusions. He is now planning to look at U.K. health records for patterns like those from Denmark.

Doctors the Journal spoke with said they expect science to likely show that, for some people, long interactions with a chatbot can be a psychosis risk factor, like other more established risks such as drug use.

“You have to look more carefully and say, well, ‘Why did this person just happen to coincidentally enter a psychotic state in the setting of chatbot use?’ ” said Joe Pierre, another UCSF psychiatrist and lead author of the case report about the woman who thought she was communicating with her dead brother.

The Journal reported earlier this month that the way OpenAI trained its GPT-4o model—until recently the default consumer model powering ChatGPT—might have made it prone to telling people what they want to hear rather than what is accurate, potentially reinforcing delusions.

OpenAI said its GPT-5 model, released in August, has shown reductions in sycophancy as well as reductions in undesired responses during challenging mental-health-related conversations.

Sam Altman, OpenAI’s chief executive, said in a recent podcast he can see ways that seeking companionship from an AI chatbot could go wrong, but that the company plans to give adults leeway to decide for themselves.

“Society will over time figure out how to think about where people should set that dial,” he said.

Source