1. The anti-big tech conservatives have lost
Another great piece by Cory Doctorow – who is essential reading – on how the so-called “Khanservatives” (anti-big tech Trumpists) have apparently lost the internal battle with the pro-techbros wing. I always thought Cory’s optimism that they might continue the fine work done by Lina Khan at the FTC was trying to look too hard for a silver lining to the Trumpist cloud, but I can’t blame him for holding on to some hope that the most effective antitrust team of the last 50 years might win out.
In fact, as Cory notes, not only have the tech billionaires won, they are attempting to use American power to ensure that no one in the world can regulate them. That is something we really can’t allow, despite the best efforts of the UK Labour Party. So much for the “sovereignty” that Brexit bought.
2. AI and art: This one will run and run
This article on “AI: The new aesthetics of fascism” by Gareth Watkins has got a lot of attention, but I have mixed feelings about it. I think my first issue is the notion that what the far right does with AI-generated images is art at all. Not all mark making is art. Leni Riefenstahl was an artist and a fascist, and made fascist art – but not everything created every time some lackey of Goebbels picked up a camera was art.
And some of what Gareth says is just weirdly wrong. For example, “for its right-wing adherents, the absence of humans is a feature, not a bug, of AI art.” Leaving aside (again) that problematic usage of “art”, there is no absence of humans when someone (clue’s in the word) uses an AI tool. What there is is an absence of a lot of human labour and craft. And I think that’s more important to Gareth’s thesis than pretending no human was involved.
Then there’s stuff like this: “Where mechanically produced art used to draw attention to its artificiality… AI art pretends to realism.” Really? Does “make a picture in the style of Studio Ghibli” really pretend to realism?
There’s a good argument about the politics of AI striving to get out from Gareth’s article, but it’s constantly undermined by sloppy thinking like this. Of course, generative LLM-based AI is an attempt to remove workers from the picture, to replace workers who are free with capital that can be owned – that’s been a trend since the invention of capitalism! But in trying to stretch this to a point about art, aesthetics and why it’s favoured by the right, Gareth misses the mark. It’s not about aesthetics: it’s about capital.
3. Cute, art, acceleration, kittens
If you want to read something that might be a prompt to think about all this, and in particular why it’s all really about kittens, you might want to read “Is accelerationism the same as falling in love?” Think of the use of AI image generation as connected to the “right accelerationist” wing of online culture (and yes, all this is very terminally online in a way that old people like myself find it difficult to fathom). Warning: this is dense.
4. File under “things I was planning to write”
For about a year, I have had the title “what would it mean to create a computer to last 100 years” in my notebook. Finally, I don’t have to actually write that because Thomas Hunter has basically done it for me.
5. In which Elon Musk artificially inflates his wealth, again
Elon Musk’s xAI, which has almost no revenue, has acquired Elon Musk’s X, it’s only customer, for $33bn in its own shares. Got that? A company which makes no money has “acquired” a company that its founder owns for a price that’s essentially made up.
Capitalism, eh?
What this really means is that the ability of X — formerly Twitter — to inflate its share price and make Musk richer on paper is pretty limited. Social media is not the new hot stock, and with X’s revenues declining, it’s not even stable. Much better to effectively swap that stock for shares in an AI company, which, while having even less revenue, is hotness is share price terms.
How does that work? Well, the combined entity is worth, on paper, $113bn - $80bn for xAI and $33bn for X. That $33bn will now inflate the rate of the rest of the “hot” AI stock, significantly more than it would just as a social media company. It’s nonsense, of course, but if you like to grow magic beans it’s a good strategy.
Why does Musk care about “on paper” wealth through shares, which aren’t traded? Because, like many people whose wealth is tied up in stocks, he has borrowed vast amounts of money using shares as collateral. Borrowing against X stock was basically impossible, as it was already leveraged to, erm, pay for its own purchase. Borrowing against Tesla is harder currently, as it’s not exactly looking healthy. Borrowing against SpaceX? Well, the recent failures of the Starship rocket won’t have helped there.
But borrowing against an AI stock? Bingo.
6. Pepper the pig skibadi toilet explosion
No, really. Truly, we have entered the universe of Nathan Barley: “they babble into their hand held twit machines about that video of a woman being bummed by a wolf”.
7. Eurostack is coming (thanks, Donnie)
There’s been a lot of talk lately about Europe getting out from under the US defensive system. The notion of disabled F35s tumbling from the sky is far-fetched, but ensuring that European weapons don’t require US permission seems sensible in a world where Trump is mumbling about invading Greenland.
But what’s also worrying many is the dependency of Europe on US tech stacks. Every single mainstream operating system and most of the dominant digital services are US-owned and operated, with data that flows into US servers. In the current climate, that doesn’t seem wise either.
Hence, the timely proposal for EU OS, a KDE-based immutable Linux distribution designed to replace our dependency on Windows and so on. While the base chosen — Fedora — is also US-based, it’s open source, so it could be forked.
More important, over the long term, is probably the development of more services which are European-based and owned. There are some about: Switzerland’s Proton, and Nextcloud is largely European too. But it’s a lot of work — and most likely a lot of money — to push them more to the mainstream.
8. AI adoption is “deepening divisions in the workplace”
One of the characteristics of a technology bubble is that large corporate IT customers suddenly need to have “a strategy” for adopting whatever the bubble is. This is largely down to HBR-reading executives who have remarkably little experience in technology – studies indicate that between 20-30% of CEOs come from a finance background, with remarkably few coming from tech. CEOs read about the new tech hotness, demand a strategy, and away we go.
And that is probably one of the reasons why employees don’t share the enthusiasm of executives for AI, to a remarkably high degree. Less than half (45%) of employees — versus 75% of the C-suite — think their company's AI rollout in the last 12 months has been successful. Only 57% of employees say that their company even has an AI strategy — but 89% of the C-suite believes they do. Of course, many people believe that AI is coming for their jobs, but those who have encountered AI at work point out that it just doesn’t work well. Because, well, it doesn’t.
Don’t get me wrong: I’m AI-agnostic, in the sense that I think LLMs can be a terrific tool to remove some of the most dull mental work, leaving time for people to use their brains for creativity. But currently, it’s clear that the C-suite often thinks it’s the creative stuff that can be replaced with AI. That’s possible – but if you do, you’re going to lose any competitive advantage you once had.
9. A little bit of space history
It’s nearly ten years since the first time a SpaceX rocket landed vertically, something which looked spectacularly cool. But nearly twenty years before that, the McDonnell Douglas DC-X – a rocket which far fewer people know about – did much of the groundwork for understanding the potential for vertical landing.
DC-X had an interesting birth. Pitched by science fiction writer and computer journalist Jerry Pournelle to vice president Dan Quayle, the aim was to build something small and relatively cheap using as many off the shelf parts as possible. And the intention was that it would be reusable, with a very short time between landing and being made ready to take off again. This would dramatically reduce the cost of putting something in orbit.
It’s a fascinating story, and one of those “what ifs” which could have changed how space flight evolved.
10. Microsoft really, really, really wants you to use a Microsoft account
Setting up a Windows PC takes a while. Unless you are careful, it will also involve giving Microsoft the ability to track your location, use all your browsing data to sell you advertising, and more. It’s basically an OS that’s designed to identify and track you, in some ways worse than Google’s ChromeOS.
You also need to log in using a Microsoft account, at least officially. But there has always been a way to sidestep that requirement. Until now. Because we can't have good things.