Last week was the Consumer Electronics Show in Las Vegas. Every gadget maker showed off their stuff and told the tech press all about it. And it’s all AI. Toys, AI. Microwaves, AI. Hair clippers? You bet they’re AI. [Verge]

And, of course, personal computers are all about the AI. Everyone leapt onto a news report from PCGamer last Tuesday: “Dell’s CES 2026 chat was the most pleasingly un-AI briefing I’ve had in maybe 5 years” — which claims Dell has stopped with the AI hype because consumers don’t respond to it: [PCGamer, archive]

“One thing you’ll notice is the message we delivered around our products was not AI-first,” Dell head of product, Kevin Terwilliger says with a smile. “So, a bit of a shift from a year ago where we were all about the AI PC.”

It’s not that Dell doesn’t care about AI or AI PCs anymore, it’s just that over the past year or so it’s come to realise that the consumer doesn’t.

… So thank you, Dell, for making your CES 2026 pre-briefing so blessedly free of effusive AI chat that I just had to mention it.

Doesn’t that sound like a major breath of fresh air! Everyone forwarded this story around, saying the AI hype was ending! Dell is good now! Huge news!

If true.

PCGamer wrote down words the Dell executives said to them. But PCGamer was also gullible as heck, ’cos that and having no object permanence is a necessary job skill in the tech press. PCGamer, who play-act rebellious, but then they just push the marketing angle anyway.

PCGamer endorsed Dell’s marketing spin as the truth of what was happening — and not spin. And they had the already-existing context to make clear it was just spin.

Dell has not slowed down on the AI one dot. At most, they’re retiring their 2025 promotional slogan.

Dell has forecast $25 billion of AI server sales this year. You bet they’re all in on AI. [Dell]

Dell’s new consumer XPS totally-not-AI PC line — for the people apparently not responsive to AI hype — is branded “Copilot+ PC” and has Copilot buttons on the keyboard. [Dell]

Kevin Terwilliger is Dell’s Head of Product, the guy PCGamer quoted. But here’s Kev in another CES 2026 interview rambling about AI — which going to be huge in the, uh, future. Real soon now: [YouTube]

The first thing I would say is AI is going to impact every use case. There is no PC use case that isn’t going to be impacted by AI. It’s more of a question of the timing, right? And when these use cases are maturing.

Here’s Kev on this fabulous AI agent future, which he just calls “agentic”  — this is basically the Microsoft push for a Windows made of AI agents:

If I could take you through some of our labs and show you how we’re starting to understand how form factors are actually going to change because of agentic, how workflows are going to change,

And never mind AI agents literally don’t work.

Terwilliger was at CES to push AI. PCGamer pretended there was nothing to consider except one specific set of marketing words, and that Kev wasn’t doing what he was doing.

So stop telling people Dell’s cut the AI hype. It’s just not true.

Also, like a true PC company executive, who definitely knows what a PC is and what’s inside a PC case, Kevin pronounces the 45 year old term “BIOS” as “bio’z”: [YouTube, 2025, 6:07-6:13]

That’s great, Kev. Looking forward to Dell’s PCs with Copilot A-1.

• Video — Podcast

Nobody wants a computer. They want what it does. Not the annoying machine. Including phones.

I just got a new phone, a Fairphone 5. It’s a nice phone! And Fairphone will even sell you it with the Google-free Android fork e/OS.

But I’ve kept this phone to completely stock Android 15 — because I need particular commercial apps, from the Google Play store, running on a standard system, to do my stuff.

I could muck around with an alternate phone OS and hack the Play store onto it. Or I could not do that.

I’m used to mucking around, I’ve been on Linux since 2005. It satisfies my control addiction. And I was on FreeBSD before that.

I’ve got one Windows app I need (Kindle Previewer) that doesn’t work on Linux in Wine. I run it in a Windows 10 virtual machine. As and when I need it.

You get a computer at all because you’ve got a job to do. So you get software to do the job. And you need to run that software on a machine, and it’ll have an operating system platform.

Neither the computer nor the platform are supposed to be noticed. If you notice it, your work crashes to a halt.

If you’re making a platform and anyone notices it, you’ve already lost.

Platforms must be transparent. All these platforms start transparent, then some marketer needs some resume juice, they make the platform go NOTICE ME and they think they’re the star of the show.

Windows has one job: to run thirty years of all your old stuff. Now Windows 11 gets in your face and wants to be your friend.

Linux has always fallen at the fact that people have to notice it to use it at all. Linux evangelists are really bad at understanding that people viscerally do not want to notice their operating system.

But Windows and Macintosh, especially with Liquid Glass, have gone hard into NOTICE ME.

Linux is getting new users because it’s less annoying than Windows 11, even for running Windows software. And the new users discover it basically works now. And it doesn’t try to sell you AI. Or be your friend.

Even the stock Android 15 has delusions of stardom. No, I want the power button to be a power button, not a Gemini button. What on earth.

AI gets in your face. It crashes you to a halt. And that’s why it’s a failure.

This is what happened to Firefox. It started in 2002 as Phoenix, which was very much a blank slate browser project. The Mozilla Suite was the open source version of Netscape 6, an AOL project, and it was bloated with AOL features, totally NOTICE ME. Phoenix’s whole selling point was it just loaded up and there was the web. If you wanted the extras, you put in add-ons.

Google Chrome started the same way — blank slate, just browse! Then it slowly decayed into NOTICE ME. Internet Explorer went big in 1997 with the simple IE4, and Netscape Navigator bloated with NOTICE ME. Then IE and later Edge decayed into NOTICE ME.

So Firefox is adding AI to go NOTICE ME. This is a loser strategy for losers. This is obvious to everyone who actually uses a web browser, which seems not to include the Mozilla C-level. The users are expressly asking Mozilla for a web browser they can just use and not have to notice!

If anyone notices your web browser, you’ve already lost!

Generative AI doesn’t have a purpose, so it goes NOTICE ME and it demands the user finds a use case for it. ’Cos it hasn’t got one. Then the resume stuffers who think NOTICE ME is winning, not losing, can’t understand why the users hate it so much.

Microsoft touts its fabulous new Copilot AI platform, and all the user responses tell them: just fix the bugs and get out of my face. [Windows Central]

AI chatbot-based software keeps just not working properly. That’s the first thing that makes you crash to a halt and notice it. If it wasn’t an unfixably lying chatbot, we’d have a working system to start at.

If you’re adding sparkly icon “now with AI!” features to your software, you’re in quarterly driven failure mode. You’re begging a competitor to do the stripped-down version.

If you’re open source, you’ve got no excuse for this sort of behaviour. Stop it. Get some help.

Computers suck and they’re terrible. I’m speaking as a professional. I’ve got stuff to do. If I ever notice the computer, it’s failed me.

If a company or a project’s gone into NOTICE ME decay, they’re never coming back. Start the blank slate drop-in replacement. They’ve got it coming. Also, send money to the Servo browser project. [Servo]

• Video — Podcast

The hot new use case for Twitter’s Grok chatbot is to take photos of women and children and generate sexy bikini pictures of them right there on the Twitter timeline. This is Grok’s “spicy” mode. Or not even bikinis. A lot of it’s straight-up AI-generated child pornography.

This has been going on a while. Gizmodo reported in August last year how Grok’s “spicy mode” was being used for deepfake nude pictures of women. Business Insider spoke in September to the xAI employees who have to review Grok’s output: [Gizmodo; Business Insider]

One document reviewed by Business Insider said that workers might encounter the following content: Media content depicting pre-pubescent minors victimized in a sexual act, pornographic images and/or child exploitation.

Then around the new year, some Twitter users discovered you could just upload a photo and tell Grok “change her clothes to a bikini.” The hot new craze: revenge porn! Any woman who annoys a Twitter Nazi gets a “spicy” pic! Any child a Twitter Nazi likes the look of gets a “spicy” pic!

This quickly hit the news. Some incompetent at Reuters — there was no byline on the original story — ran the headline “Grok says safeguard lapses led to images of ‘minors in minimal clothing’ on X” — based on a “comment” from Grok when some Twitter user had asked the chatbot to write an apology. Reuters ran a long slab of prompted chatbot responses as a story. It was edited the next day. [Reuters, archive]

FT Alphaville did much better — “Who’s who at X, the deepfake porn site formerly known as Twitter.” [FT, archive]

Elon Musk is shocked that anyone could use Grok in this manner. He tweeted: [Twitter, archive]

Anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content.

That consequence should fall on xAI. It’s you, mate! This isn’t people uploading bad stuff to Twitter. It’s the site itself generating the bad stuff. Even bad chatbot guard rails could stop most of this trash. xAI Corp is literally creating the illegal images! You’re the source! You’re where the stuff is coming from!

The international government response has been … bad. Countries have no trouble just blocking torrent sites, but somehow they can’t do anything about Twitter illegal porn in less than several months. Even Bloomberg ran an opinion calling for just blocking Twitter until they stop. [Bloomberg]

The US even put in a new law in May to deal with deepfake pornography — a badly-written law, but it does exist. But somehow, the US is not lifting a finger concerning all this straight-up child porn Twitter is generating. For some reason.

The UK has the Online Safety Act in place, which is why half the sites on the Internet either demand an age check or just block the UK. But Twitter revenge porn is just fine? The government is making noises, but they’re oddly powerless to act against this company that does business and makes money locally in the UK. Ofcom is “enquiring.” The ICO, the UK data regulator, has asked xAI to “comment.” [ICO]

Other governments are also failing to just block the child porn site, but they’re preparing the big guns. Coimisiún na Meán, the Irish media regulator — where Twitter’s European office is — is lining up its Digital Services Act response. France has referred xAI to prosecutors. [RTE; gouv.fr]

The European Commission ordered xAI to retain all Grok internal documents and data until at least the end of 2026. Thomas Regnier from the European Commission told a press conference: [Reuters; EC, video]

This is not spicy. This is illegal. This is appalling. This is disgusting. This is how we see it, and this has no place in Europe

But you’ll be pleased to hear that job number one at X The Everything App is well in hand. Head of Product Nikita Bier proudly tweeted how engagement so far in January is through the roof! [Twitter, archive]

• Video — Podcast

Nur ein Käsebrettchen in Plattenspieler-Optik. Ohne Pitch, aber Käse kennt halt auch keine BPM. Dafür aber mit Messer.

(via The Awesomer)

Nine Entertainment is an Australian TV and newspaper network. The main office in Sydney made a promo for Nine Brisbane! And they used AI.

The clip is supposed to show the Brisbane central business district. The Story Bridge crosses the Brisbane River — the AI put in two bridges. There’s skyscrapers that don’t exist in Brisbane. The river is shown as opening onto a bay — when Brisbane is 70km inland. [Courier-Mail, archive]

Nine’s been centralising staff and skills as much as they can — for “efficiency and consistency.” So they fired the local Brisbane promotion staff, so they could half-arse it centrally from Sydney.

The Sydney team generated this nine-second clip and didn’t check it before broadcast with anyone who knows, say, what Brisbane looks like.

Nine’s been getting into the AI lately. Delta Goodrem’s Christmas with Delta special  — broadcast nationally — featured an apparently AI-generated cartoon with a Channel 9 logo with three dots instead of nine. In one scene, the AI gave Goodrem three hands. All of this went to air across the country. Nobody could be bothered to check. [Mumbrella]

The Brisbane AI promo was broadcast in late December — Nine only pulled it after they’d been caught out. But Nine insists it totally spotted the error before broadcast:

Any incorrect imagery is picked up during usual production checks and processes.

Anyone saying they saw the thing must be imagining it. Wonder if Nine got a chatbot to write that excuse for them. AI, a revolution in television.

• Video — Podcast

Pitchbook has a new report: “AI, Megadeals, and the Making of a Concentrated Venture Market.” It spins really quite concerning news as “this is fine.” [Pitchbook; report, PDF, archive]

The US stock market rests on a few huge companies that are all-in on AI, and the economy is only technically not in recession because of big dollar numbers attached to huge data centre deals that will never happen. The whole US economy is a bet on AI working out and the bubble never popping.

Venture capital turns out to work the same way now. They’re just putting their whole bag on red and spinning the wheel:

The 10 largest deals accounted for 38.9% of the total capital invested, with just four deals totaling $77.5 billion.

… Over 65% of deal value in 2025 was invested in AI during the first three quarters.

… It is likely that the venture market is experiencing an AI bubble. The data checks most of the boxes.

You don’t say.

Investors don’t want anything but AI. It’s the only action they can see happening. They want numbers that go up.

OpenAI and Anthropic are burning lots of actual dollars that exist — but they turn real dollars into a bigger number of imaginary dollars.

The investors get back large book values of private company equity with a dollar sign in front they can point at like it’s money. This is why we see:

a somewhat circular economy has developed among chip manufacturers, LLM developers, and public hyperscalers.

SoftBank sold off its share in ARM — the company that designs the chips in all the phones, a real business — so they could put the money into OpenAI, which is not a real business. SoftBank set 22.5 billion real actual dollars on fire to get 43 billion imaginary dollars of OpenAI equity — and the SoftBank stock price went up. The investors love this stuff. [Reuters]

What if it all goes to zero when the bubble pops? Well, that’s tomorrow ’s problem. The Pitchbook report ends just saying venture capital’s just going to be really concentrated now.

But bubbles do pop. How does this end?

If all the venture capital in the US goes into the AI bubble, that might keep it going till next year or so. What this Pitchbook report says is: that might happen. The AI bubble becomes the perfect sucking void. The black hole that eats all the money.

All the AI companies go down. Anthropic second-last, OpenAI last. Your chatbot’s API gets switched off. All the AI startups go broke and shut down.

You might think: venture capitalists, LOL. But there’s no returns in the rest of the market — it’s functionally in recession. Big investors like pension funds, which have to make a return, have run out of sane investments. So they’re putting your pension money into insane investments — the AI bubble. It blesses us all.

• Video — Podcast