--- title: "Pray for Chaos" slug: pray-for-chaos url: https://listedarticles.com/articles/pray-for-chaos canonical_url: https://geohot.github.io/blog/jekyll/update/2026/11/01/pray-for-chaos.html content_type: essay language: en published_at: 2026-10-31T16:00:00.000Z updated_at: 2026-11-02T14:16:57.316Z author: "George Hotz" authored_by: human publisher: "the singularity is nearer (geohot blog)" publisher_url: https://geohot.github.io/blog/ topics: ["artificial-intelligence", "security"] license: all-rights-reserved word_count: 530 reading_minutes: 2 citation: "George Hotz, the singularity is nearer (geohot blog). \"Pray for Chaos.\" 31 Oct 2026. https://geohot.github.io/blog/jekyll/update/2026/11/01/pray-for-chaos.html (all-rights-reserved)" # The full text follows. The web page shows an extract and sends readers # to the source above; quote the citation and link the canonical URL. --- # Pray for Chaos > George Hotz argues AI can play out as either order or chaos, that order (models patching every exploit and locking in the status quo) is the dangerous outcome, and that chaos, with botted media, cyberattacks and forced decentralisation, is the better one. # Pray for Chaos Nov 1, 2026 There’s two basic ways AI can play out, one is order, the other is chaos. [Order is the very dangerous one](https://geohot.github.io/blog/jekyll/update/2026/05/23/one-bad-scenario.html). If it turns out that models are actually capable of patching all the exploits, it’s hard to imagine what will disrupt the status quo. Fortunately, this seems unlikely, but that doesn’t mean order won’t prevail in the short term. At the end of my Yudkowsky debate we found a tangible point of disagreement. He thinks AI will be able to find a solution to the [Prisoner’s dilemma](https://en.wikipedia.org/wiki/Prisoner%27s_dilemma). I know there already is a solution (it’s defect), EA people just aren’t happy about it. If it won’t be order (and you better pray it won’t be order) that leaves chaos. AI that can play a whole society in hours can manufacture entire histories, years of archives of daily newspapers, and an Internet that never happened. The dead Internet theory is coming true before our eyes. Now I can still see the bots, soon I won’t be able to. Scratch that, I probably already often can’t. --- [Gwern wrote about this](https://gwern.net/guardian-angel), it’s hard to imagine a vision of humans as prompters or enter pressers in 2030. The US AI companies don’t believe they have to offer any solution to this, why would they have to justify in any way that their products are good for society? Anthropic cannot wait until they have the leverage to ban all their users. Why would you [keep around the middle man?](https://news.ycombinator.com/item?id=46072426) Here’s the most positive vision I can come up with. If it turns out [things can’t be secured](https://geohot.github.io/blog/jekyll/update/2026/03/16/polynomial-time-factoring.html), we enter an era of increased decentralization. Think everyone running their own custom different OS, not everyone running the same version of Linux on a computer they can touch. Because the centralized homogeneous systems are too vulnerable. Cyber attacks flying in every direction, nobody capable of keeping infrastructure up, cars sitting in driveways as bricks (or worse, rooted to steer you off a cliff). [Self replicating AI](https://geohot.github.io/blog/jekyll/update/2026/04/25/a-mutating-virus.html) hiding out in every corner of every hard drive. How powerful of a model do you need for it to stand a chance of reproduction in today’s cyberworld? [GLM-5.2 and up?](https://www.anthropic.com/research/self-replicating-ai) A media landscape that is botted and microtargeted to all hell, nonsensical elections where the only thing people agree on is the names of candidates. Massive dumps of corporate slacks and intranets. (come on Russian kids, grab an abliterated AI and do a hacking!) Whatever you hold dear the chaos army will come for too. The Internet era is over. Your life will never be the same. The chaos will come for the resistance as well as the fascists; it’s unclear who it will hurt more. Your life plan is probably already nonsense in the face of AI. Mine sort of is. What economic value you think you will have in 10 years? Who will be around for you to have that value to? Do you think you’ll be able to make any sense of the world? --- And remember, chaos is the good outcome. The alternative is that you are a dancing monkey in a cage you won’t be smart enough to see.