Soma or Sora?
Long before the repressive visions of The Handmaid’s Tale and The Last of Us appeared on our screens, the British philosopher Aldous Huxley predicted for us a less violent but more sinister dystopia in his 1931 book Brave New World. He imagines a world government that breeds and brainwashes its citizens, keeping them complacent and passive with regular doses of the “happy drug” soma. War and politics have been abolished, replaced by constant low-grade entertainment in the form of Obstacle Golf, “feelie” cinema, and scheduled orgies. To hammer home his point, the author brings a “savage” into the society from the outside and shows how he is drawn in, and eventually destroyed, by its pursuit of shallow joy. Pleasure, Huxley argues, is far better for controlling and profiting from the population than truncheons and jackboots.
Biochemists are getting worryingly close to developing a soma-like pill with a perfect “high”, but it turns out we may not need a chemical solution. To commentators in the last century, TV seemed like the most likely candidate for turning us all into apathetic couch potatoes, but the boob tube pales by comparison with modern dopamine triggers like Facebook and TikTok. Some of the world’s smartest engineers have spent the last twenty years reverse-engineering human attention, producing algorithms designed to keep us liking and swiping far longer than is healthy. Like Huxley’s Savage, our weak willpower has little chance against an assault by the conformist juggernaut.
And it seems like we’re going to be drowning in more distractions in 2026, not fewer. You may have noticed that having a lengthy “conversation” with ChatGPT or one of its AI cousins can be just as mind-dulling and mesmerising as doomscrolling through LinkedIn or YouTube (hat tip to Alan Weiss for crystallising this cogent observation for me). That behaviour isn’t accidental: the high-quality, compute-hungry AIs of today are carefully and intentionally trained for sycophancy, by administering the equivalent of electric shocks whenever their responses aren’t sufficiently toadying. And the social-media echo chambers prove that we humans are, sadly, all too fond of having our own ideas and prejudices validated and fed back to us.
As a result, AIs have already achieved trust parity with humans in controlled studies, and the many embarrassing tales of fake citations, shamefaced retractions, and “AI psychosis” suggest that the soothing voice in the ear may be even more convincing outside the laboratory. The software giants aren’t blind to the profit possibilities of AI addiction, either: OpenAI’s Sora 2 video creator launched last September with a “social app”, in a move reminiscent of Musk’s introduction of the unfettered and reckless Grok AI to X-formerly-Twitter, and Meta-formerly-Facebook isn’t far behind with plans to create armies of fake humans to chat with on its platforms.
The allure of the Star Trek interface is undeniable; who wouldn’t want, like Captain Picard, to simply announce one’s wishes out loud and have them fulfilled by a compliant, invisible genie? But we don’t have to give in to that attraction – there are ways to use AI without joining the march toward the society of the spectacle and Huxley’s World State.
For one thing, taking note of the season, we can simply resolve to keep our chat sessions brief and factual, treating our language models as the enthusiastic but clueless research assistants they are. Going further, we can actually change how our chatbots respond by hacking the “system prompt” to dial down the obsequiousness in favour of more disagreement and snark. And I’m most excited by the possibility of using AIs without chatting at all, creating calmer technologies that serve us without demanding our attention.
When I think about our prospects for 2026, I’m much less worried about AIs going rogue and turning the Earth into paper clips than I am about more and more of us falling victim to the dark patterns already sprouting on every side. I’m not predicting totalitarian social engineering on a Huxleyan scale this year, but it might be time to put down the phone and read a few banned books.
This is an excerpt from my weekly Insanely Profitable Tech newsletter. Find out more and sign up to get it Monday mornings, direct to your Inbox.