The two weeks since Trump's inauguration have been an unmitigated mess. I'm writing this in Hamilton on a bright Sunday afternoon, midway between rage and despair, watching as Elon Musk's goons infiltrate and destroy the civil service, the Christofascist wing of the regime does their best to legislate trans people out of existence and everything generally goes to shit.
I don't think I can address everything that's happened lately, but one thing that does bear addressing is the fact that the AI discourse has somehow become even more deranged and brainless than it was before the inauguration. The leaders are one-by-one capitulating to the new regime, with the Zucc's new moderation guidelines in social media explicitly encouraging misogyny, racism, transphobia and antisemitism and oligarch and person-credibly-accused-of-child-sex-abuse Sam Altman immediately capitulating to Trump in this disgusting tweet:
The only exception is Elon Musk, who appears to be doing a coup, with his goons having gained access to the fucking US Treasury's electronic systems and the Office of Personnel Management and attempting to systematically disassemble the US state apparatus. While doing this, he and his goons are increasingly claiming that "AI" will be able to replace many of the functions of the state that he's running, whatever that's meant to mean. Credible sources in the US civil service have confirmed to me that grants and funding are being reviewed by LLM technology as we speak, looking for both implicit and explicit references to DEI so that they can be defunded, which is an absolutely deranged use of the tool.
In short, everyone involved in the current AI bullshit is deeply, deeply tainted by involvement in a goddamned fucking coup of the US government, and should be treated as such. And on LinkedIn a whole lot of "important businesspeople" are still burbling about how this bunch of idiots, coup leaders and rapists are actually brilliant business leaders and how AI is going to revolutionise the world and change everything. What the fuck is going on?
German Jewish refugee and psychologist Erich Fromm, in his book Escape from Freedom, discusses the deep-seated need of many humans to seek and engage with sources of authority and control, both as a controlling figure and as a person being controlled. This is discussed in the context of fascist movements, wherein we see large bodies of people in society absolutely gagging to obey and to conform themselves to the will of a greater power. And that, I think, explains much of the AI boosterism we've been seen lately. It doesn't have much to do with the tech, or even with the economics of it. It's a dominance-submission dynamic.
Even if it was something else back in 2023/2024, the AI thread in our society now is nothing more or less than a demand by the wealthy creators of it for submission. It's a naked show of force by the powerful and stupid: you will use this tool, and you will like it. You will treat the people who created it as genii who stand above you, and you will submit to the will of the Master, who knows what's best for you, and more importantly, what's best for your in-group. You're even, in Ludic's excellent phrasing, expected to submit your own thought, judgement and even reality to that of the Master, and believe that these chucklefucks who can only see the world in terms of domination will create an Artificial General Intelligence any time now.
The fact that this is what's going on here is spelled out clearly by the cloud of bootlickers surrounding almost every major figure in the reactionary tech ecosystem. Elon Musk's cabal of nineteen-year-old junior software developers, who volunteer to work at DOGE and Twitter for no money, who go along with everything he says, gagging desperately for a retweet or a shoutout from Musk no matter how badly he treats them, are the archetype of this, and they're everywhere from Twitter itself, to the papers and the people whom you encounter in the real world, to right-wing media the world over, to goddamn LinkedIn of all places, where we can find no shortage of people willing to take Musk's boot as deep down their throats as they can possibly manage. The Zucc's fanboys are no better, and even people who know better than to adulate the first two seem very enthusiastic about boosting Sam Altman and his OpenAI tools, when in reality he's just as spineless and slimy as the rest of them.
In this year of our Lord 2025, the tech doesn't matter to these people, nor does what it's capable of. The economic effects don't even really matter: AI is not going to replace workers in the economic sense of "doing what they do", and at this point everyone knows it. Even the effect that it has of massively enriching them and disciplining workers is a side effect: what they really want from this, as we see, is the ability to force people to submit and to do what these men want them to do. When the submission is willing, it's fine, and when they have to force it, it's just as gratifying: the first makes them feel loved and the second makes them feel strong and powerful. And all this to desperately fill the deep inner emptiness that these people feel, usually due to what David Ten Have refers to as "daddy issues" (or mummy issues, for that matter, though those usually manifest differently).
This is why Microsoft Word and every other tech product you buy keeps pushing AI on you. This is why the hollow, soulless executives that run your company keep getting up on stages and talking about AI being the "future": at the best they're desperately trying to gain prestige by saying what "important" people are saying: at worst, they're tacit supporters of fascism. All of them are complicit. And as for Altman, Zuckerberg and Musk, some of the higher-profile dunces of my last piece ... they're sad little boys, not loved enough as children, who are now trying to force reality and other people to fit their delusions. They want desperately to be seen as big, dominant men who with their massive brains created an amazing technology that solved all our problems. Instead, they find that the technologies they created allowed the vast bulk of people to let them know exactly how they feel about these dead, rapacious ghouls. And so, at this point, their language models have to be a success. They have to work, and they have to be revolutionary in order to protect fragile egos. And so, these men will use the tools they have, and try and force us to submit to their reality and admit that they're amazing.
Moreover, at this point the LLM tools are themselves becoming nakedly authoritarian. From Grok and all of its attendant bullshit, to Deepseek and its refusal to admit that anything bad might ever have happened at the Tian An Men, to the aforementioned use of LLMs to identify documents as having "DEI" in them, the tools themselves are becoming expressions of authoritarianism. I'd argue that this is latent in the tools themselves: heavy LLM use has been shown to degrade your ability to think clearly and carefully about the subjects that you use the LLMs for, and the whole wretched apparatus of the thing encourages you to outsource your thinking to a machine. This is bad enough from an engineering perspective, where it introduces all kinds of vulnerabilities and antipatterns, but it is, if anything, even worse when you outsource your critical thinking to the LLM. Once you trust it enough, it can feed you anything, and you are more than likely to simply accept it uncritically.
And when that happens, it's perfectly simple to twiddle an AI's starting prompts to get it to start feeding you propaganda. It's notoriously well-known that ChatGPT's start-up prompt is remarkably long and also instructs it to take specific political positions. Now, Sam Altman is exactly the same kind of flabby, undercooked non-person that the Zucc is, and we've seen Zuckerberg fall full-force into fascism and weird masculinist ideologies. Altman is under significant moral stress: what's the bet that he'll start making it feed everyone fascist propaganda? Are we remotely prepared for the consequences of every damned chatbot built on OpenAI's APIs suddenly starting to talk like Julius Streicher? Considering how little thought people put into LLM outputs, it is very easy to simply start distributing massive volumes of LLM-generated propaganda using these tools, and it'll get everywhere.
So far, this is bad, but perhaps not necessarily ideological as such. After all, social media has done many horrible things to our society, but it is also something that can be used for progressive activism and mobilisation, and it has been successfully used for that many times. The issue with AI is that it can't do the reverse. It can't encourage people to think clearly and critically, to do their own work of reading, writing or communication, or to check whether what they're writing is true. It can't even consistently generate true statements by itself. It is, by definition, a machine for generating lies and bullshit, and that makes it an authoritarian technology. LLMs will not support our society and our democracy, they can only degrade it.
So here we are: the tech and AI magnates are universally either falling in behind fascism or actively driving it. They're also desperately pushing us to use a technology that doesn't work as intended, that cripples our ability to think critically or work effectively and that can lie to us in ways they mostly control whenever they feel like doing it. Moreover, they insist on us believing all the lies that they tell us and dismissing the evidence of our own senses. They insist on being able to control our perception of them as brilliant, strong, genius leaders, and they insist on us being unable (either through incapacity or unwillingness) to think for ourselves. This is quite directly Orwellian, and it makes the fact that the industry hasn't caught on and isn't meaningfully responding to this all the more shocking. The corporate world is still talking about AI in exactly the same way they were previously: rampant boosterism or milquetoast and inadequate critique that it "doesn't always work".
At this point, refusing to use AI, and especially LLM tools is an act of resistance. Give them no quarter. Purge them whenever you find them. Turn off Copilot. Poison any data that you put out into the world, and make it as difficult as possible for AI bots to train from your data. Treat the people who mindlessly boost AI as collaborators.
If you're an engineer, I genuinely believe that at this point you should refuse to work on products relying on LLMs outright. Yes, I know that the technology isn't completely useless. Yes, I know it can do some cool things. That no longer matters. Everyone working on and funding work on LLMs is complicit with the government of the global hegemon being taken over by an idiot that paid people to get him to the top of the Diablo 4 leaderboards. Their attachment to LLMs is entirely about making people submit to them and their reality in which they're very special geniuses who will save the world, all while they destroy the USA and loot its government. You should not work on LLMs, and you should do everything you possibly can to destroy the technology and make it as difficult to work with as possible. Yes, this might get you fired. Yes, it might get you sidelined. That doesn't matter. Give no quarter to fascism, wherever you see it.
And if you're one of those "tech leaders" who gets up on stage to bloviate about how "AI" will somehow revolutionise everything: I now consider you complicit with the fascist takeover of the USA and the ongoing attempts by the new regime to exterminate trans people, deport immigrants and resegregate people of colour. You are spineless, thoughtless cowards who gave them and continue to give them the social capital that they needed to pull this coup off, and you continue to do so every time you get up on stage to talk about your shitty chatbot. I will do everything I can to ensure that your names live on in ignominy and that you are remembered as the gutless, spineless quislings that you are.
... shit. This sounds so stupid. It's almost impossible to write anything smart about this because of just how moronic all of it is.