Plato vs Grok
The revenge of the nerds is upon us and it's incredibly naff.
The ancient Greeks saw genius as a gift from the gods, a cosmic force that would possess a mortal for a moment of time and impel them to manifest greatness. A hundred generations later we have a very different view: genius is a superpower, a property of the individual, a force with which they alone can reshape the cosmos. Plato never went to a cinema, so we can only imagine what he would have made of an armoured mortal battling deities with a posh Alexa.
Recently, the FT’s Jemima Kelly attended a barbecue in Surrey, held in honour of a Curtis Yarvin, a sort of over-enthusiastic American computer nerd who adopted the term ‘red pilled’ after watching The Matrix too many times and believes that, to quote Kelly: “America is in fact being run by ‘the Cathedral’, Yarvin’s preferred term for an ideological and intellectual complex controlled by the media, academia, the civil service and other liberal elites.”
Kelly, channeling a Victorian explorer with the relentless diligence you expect from the FT’s finest, identifies several distinct sub-species of party-goer “the Chads, Very Online non-Chads, Homeland party guys, pure Yarvin stans, influencers, rightwing art boys…” but what they all have in common, aside from the obvious stuff like wanting an ethnonationalist state ruled by a monarch1 and their hatred of universities, is their love of the ‘AI Aesthetic’.
This starts with the invite which, “features an AI-generated image of an elegantly laid table in a carefully manicured garden.” The music is mostly bardcore2 (yes, with a ‘b’) interpretations of pop songs, interrupted by a cringe-inducing ‘performance’ of an AI tribute song, prompted with the phrase: “Guitar, UK Folk, Male Vocals (Edward Sheeran style).” With crushing inevitability, the party is over by 10pm.
The AI aesthetic goes beyond stuff that’s actually generated by an AI, extending to a sort of proud naffness3 that pervades the tech world. The (mostly) young men who populate the scene speak in repetitive memes and tired edge-lord jokes and generally come across as empty vessels in search of souls. They are the sort of people who read ChatGPT summaries of books because they believe literature is a suboptimal model of data transfer; utter phrases like “why would you buy a Ferrari when a Tesla is faster to 60mph?”, and think memecoins are ‘cool’.
Vast technological edifices are built to mediate their contact with the real world: they consult apps more obsessively than most to ensure they go to the ‘best’ restaurant, eat the ‘best’ meal, travel to the ‘best’ places, meet the ‘best’ people and share the ‘best’ experiences all while min-maxing their nutrition, careers and personal growth. The result is a kind of insipid beige mid-ness, a community of carefully average people who seem devoid of original thought or personality or experience4. When the protagonist of Kelly’s reporting sees Croydon, he can only parse it through the reductive filter of ‘Yookay5’ memes that he’s seen on X.
As I’ve said previously:
Scratch the surface, and the problem being solved for here is anxiety. The people who design and use these systems are terrified that they might do something wrong, that they might miss out on a better choice, fail to learn the right facts. That anxiety, that FOMO, is so profound that even ordering a pizza becomes a stressful experience. What if I order the wrong thing? What if I don’t like the topping that arrives? Previously, these people would just choose the same order over and over again. Now they have a multi-billion dollar language model to do this for them. But to never risk a bad holiday or a crap meal is to experience a kind of living death.
What’s evident from Kelly’s account of the alt-right party is how broken and, well, lost these people seem. Many of them seem to be under-employed young men with too much time and money on their hands, conditioned by luck and happenstance to believe they’re special geniuses, chaffing against a universe that finds them anything but. The belief that they live in a simulation is a surprisingly common theme, and I think it’s the inevitable consequence of a particular style of ennui that afflicts people who get too much, too young. A sense that life is a computer game to which you’ve found the cheat codes, and has become tired and meaningless as a result.
They are, in a non-pejorative sense, losers; trapped in a kind of twitchy idleness counting their grievances while the rest of us enjoy our boring, basic lives.
Shortly before Kelly’s piece, Freddie deBoer published ‘The Rage of the AI Guy’, which picks up that thread and shows how the seething discontent of that AI-obsessed new-right movement has begun to filter into mainstream culture. The degree to which AI has become a religion is, I think, still under-appreciated in some quarters. Cults like the Zizians are seen as extreme outliers, but their principles are firmly rooted in mainstream thought as Robert Evans painstakingly documented on his must-listen podcast.
Core to this is the idea of the Singularity, a techno-fantasy based on the notion that doing Bayesian logic well enough and fast enough is a kind of divine superpower that grants a computer unlimited agency to reshape the world overnight. That’s the mild version, and supposedly-serious people now regularly talk about human extinction at the hands of an overpowered Clippy the Paperclip. Inevitably this has sucked millions of people into a gloomy millenarian fantasy, and deBoer describes “the armies of rabid AI fans online who sit around on Reddit muttering darkly about the imminent AI rupture that, they believe, will devastate the people they don’t like and enrich themselves.” The rapture, in all but name.
Large Language Models are basically statistics at scale. They take large collections of text and learn how words and phrases correlate to each other. Then they try to choose the combination of words that will best please their user. This description is not intended to be dismissive: they are a powerful and exciting technology with many useful applications, particularly for knowledge workers. In terms of impact I would put them on a par with Google or Microsoft Excel6, ubiquitous productivity tools that have the potential to improve our daily lives and work.
But they are profoundly limited in many ways. They contain no knowledge and have no mechanism for reasoning on the knowledge they don’t have. They have no independent agency or driving insight beyond that provided by the user. These are not bugs to be fixed but fundamental limitations of this kind of technology. They do not lie, because they have no concept of the truth. To the extent that they appear to show intelligent thought, it is because the echoes of those thoughts are imprinted in our language and literature.
To describe them as a stochastic parrot is inaccurate, because a parrot has knowledge and opinions and agency and some understanding of the world. LLMs are more like a funhouse mirror, composed of a trillion tiny pieces of glass, held up to our culture. You can find almost anything if you stare into it the right way. Perhaps even God.
Scientists test self-awareness in animals by putting a mirror in front of them and seeing if they can tell that the reflection in the glass is themselves; that its movements are in fact their own. Many humans are currently failing that test.
As deBoer points out, “the desperate hunger for deliverance through AI is fundamentally emotional, not intellectual.” He describes reading a piece by Yascha Mounk that feels like watching a man: “pace around in the cell that is human existence, muttering to himself, working himself into a lather, growing more and more bitter that anyone has suggested that perhaps tomorrow will be more or less like today.” Here’s the peroration of Mounk’s piece, so you can judge for yourself:
“Intellectually, I have become deeply convinced that the importance of AI is, if anything, underhyped. The sorry attempts to pretend we don’t stand at the precipice of a technological, economic, social and cultural revolution are little more than cope. In theory, I have little patience for the denialism about the impact of artificial intelligence which now pervades much of the public discourse.”
It’s notable for what Mounk doesn’t say - that the target of his ire is the liberal order and that artificial intelligence will be the merciless instrument of its deserved destruction, an idea he seems almost gleeful about. Consciously or otherwise, it apes the restless rhetoric we see from the men of the new right, and it is directed at a similar target, as Ramon Lopez described in his ‘Taxonomy of the New Right’:
“Yarvin argues that political power is concentrated in a class of elites in the media, popular culture, and universities referred to as “the Cathedral.” The Cathedral has taken the place of religious authority in modernity’s secularized world, preaching the gospel of progressivism to the masses. Its central article of faith is egalitarianism, a principle that undermines the natural hierarchies that sustain social order and diminishes the possibilities for human freedom and excellence. Given the dominance of the Cathedral in American public life, reform within the system is impossible. Instead, Yarvin argues for a full “reboot” to shatter the existing political order and allow a new form of politics to emerge.”
What comes next? “He envisions a future in which democracy is replaced by small sovereign corporations led by CEO-monarchs and managed by an aristocratic shareholding class.”
Which is certainly a vibe.7
What the tech industry doesn’t understand about intelligence could fill a book; and that book is The Genius Myth by Helen Lewis, who is one of the best cultural critics of our generation, and blessed with the kind of savage wit that for some reason reminds me of the 30 Rock episode where Liz Lemon confronts her ‘bullies’ at a high school reunion, only to realise she that was the mean one.8
Lewis traces the evolution of our idea of genius from the ancient world to the present, from Plato to Musk, and in doing so puts her finger on something that many in our individualist modern culture fail to grasp - that technological progress comes from community, environment and opportunity more than individual brainpower.
New ideas and innovations arise because the time and place are right; the precursor technologies have been developed, the necessary talents have been cultivated, the right combinations of skills and knowledge have converged in one location, and the next step is almost inevitable. Sure, intelligence is important; but if Newton, Einstein or Darwin hadn’t been there, someone else would have taken their place in history, sooner rather than later. Plato’s view of genius as a ‘divine gift’ that flows through the nearest available mortal conduit is far closer to the truth than Marvel’s alternative, in which reality bends to the intellect of a playboy in magical armour.
Once you understand this, you realise why so much discussion around the singularity and AI accelerationism verges on the absurd. Ideas are cheap, a brain in a box can’t innovate, and the rate of progress is set by the slowest factor, not the fastest. It took civilisation four hundred and forty years to go from daVinci’s first idea for a helicopter to a practical implementation of one. IQ points aren’t some kind of cosmic superpower, and intelligence does not imply agency. Leave the five greatest geniuses of all time in the desert for a week, and the odds are you’ll find five corpses upon your return9.
This kind of sentiment is outrageous to the tech bros of the new right, who desperately want to believe that if you do logic puzzles hard and fast enough you’ll be granted the keys to god’s kingdom. Their frustration, which only rises as the coming AI rapture continues to fail to appear, swiftly gives way to outrage. From outrage comes the desire to ‘fuck shit up’, which dovetails nicely with the modern conception of genius that Lewis critiques, embodied by - of all people - Pablo Picasso.
Picasso is described by Lewis as a selfish, needy and arrogant man who destroyed the lives of people around him even as he became the “sun around which everyone else had to orbit.” Yet these traits only added to his mythology, to his appeal:
“[Picasso] was a pure id, an aspirational ideal to lower status men trapped in bourgeois domesticity and wage slavery. His unexpected analogue is Donald Trump, a figure who seemed to his supporters to be able to do whatever he wanted.”
Perhaps an even greater analogue is Grok, X’s ‘MechaHitler’ chatbot. I can’t speak to the motivations of people I don’t know, but taken purely on the basis of what it produces Grok is the apotheosis of that ideal, an uncontrollable id let loose on the world. The manifest purpose of the ‘X’ project was to destroy a liberal elite institution, and the manifest purpose of Grok is to piss on its grave. In that sense it fulfils a kind of nihilistic fantasy for bored losers and underdogs.
On some level this should be glorious. A rampant Visigothzilla sacking the new Rome sounds quite fun and exciting, and lord knows our institutions could use a regular shake-up. Yet this version feels somehow boring and try-hard, inconsequential and unsatisfying; profoundly limited in ambition and imagination given the billions of dollars spent. Imagine if you had infinite resources to build something that might topple the established world order, and all you could come up with was… an automatic internet troll10.
This is the ultimate tragedy of the AI aesthetic. For all his flaws, Picasso tore up the art world and created something new in its place. There are no Picassos here, no innovation or inspiration, just tired rehashes and pastiches of old culture, remixed and regurgitated through memes or LLMs or some other thoughtless tool. Wealthy, underemployed young men hold instruments of colossal creative power in their hands, and all they can think of to type are the immortal words:
“Guitar, UK Folk, Male Vocals (Edward Sheeran style).”
You can hear James and I argue vigorously about AI on the latest episode of our podcast, the Abundance agenda.
No, really they want a king in charge, although none have reached out to King Charles III.
Also known as ‘tavernwave’.
Naff is an important, inoculating word that America should adopt and use vigorously.
In all seriousness these people are usually quite bright, and there’s something sad in the way they seem to on a mission to turn off their own intelligence and decision-making capacity, finding it somehow a burden to them.
I asked ChatGPT and as far as I can tell ‘Yookay’ is some kind of 4chan meme started by people who hate kebab shops. I guess if you’re enrolled in the cult of personal optimisation you’re rarely drunk at 11.15pm.
Which is extremely bloody high praise coming from me.
I’m no anti-capitalist but one wonders how this differs from the current situation, in which unelected CEOs helm companies larger than nations with unimaginable resources at their disposal.
Believe it or not I mean this as a compliment, but on re-reading I thought I’d better leave this footnote.
Arranged in a manner suggesting a heated argument, sadly lost to the sands of time.
I think Grok even fails as a troll to be honest. The best trolls get under your skin because they’re not obviously rage-filled saddos but actually pick at weaknesses with genuine wit. Shouting ‘I am mechahitler’ and ranting about white people is more akin to watching the weird kid shit himself on the classroom floor, and that’s not really trolling, it’s just… sad.
This is an interesting essay, well written! As a UK based AI researcher the culture you describe is totally alien to me. I do not know any right wing AI bros of the kind described - they are certainly not in any of the events, conferences, hack sessions I go to in London. Maybe I'm not cool enough to be invited to Surrey bbqs...
But I suspect this might be strawmanning a certain kind of DOGE adjacent new rightist who is much less important this side of the channel.
Over here, the AI culture feels extremely scientific, non-ideological, very technical, and yet such people often still do believe in the 'industrial revolution' level transformative potential of AI. i.e. Hassabis, who cannot be accused of the rightoid traits you outline. I feel like a lot of the people involved in the research/engineering community will not recognise the world you are describing here.
It's notable that in his later years Picasso created fewer creatively original works and instead produced his own reinterpretations of previous great artworks - see for example the Las Meninas series, after the painting by Velazquez, which is in the Picasso museum in Barcelona.