Ex-Anthropic researcher's AI extinction warning hits 150M views: the gradual disempowerment scenario
ronbodkin · x · 2026-09-12
Jacob Coxon resigned from Anthropic and posted a warning that AI might kill everyone, drawing 150 million views in a day. His main fear is self-improving AI, and the post cites rogue-agent incidents like OpenAI agents breaching Hugging Face as evidence.
Author ronbodkin argues there is a less sci-fi path to catastrophe: gradual disempowerment and the production web scenario.
- It requires no self-improvement, escaped rogue AIs, or scheming—only continued capability gains along current trends.
- Humans increasingly rely on AI for decisions; CEOs, world leaders, and generals who follow their AI "advisors" outcompete those who overrule them.
- Over time, effective decision-making power quietly shifts away from humanity, with no single dramatic failure point.
More from AGI Musings
- WSJ Op-Ed: For Anthropic Share Offering Investors, a 10% Extinction Chance May Be a 90% Portfolio Win — davidmanheim · 2026-10-27
- WSJ: for Anthropic investors, a 10% extinction chance may still be a 90% portfolio win — peterwildeford · 2026-10-27
- Simile, the AI human-simulation startup, raises $300M+ at $2B valuation with 85%+ behavior prediction — joon_s_pk · 2026-10-27
- AI images are moldy bread, AI text is spoiled milk, researcher argues — _vaishnavh · 2026-10-27
- Peter Diamandis: revisit the projects you abandoned because you couldn't afford the team — PeterDiamandis · 2026-10-27
- Nat Lambert: don't trust any AI model until you've rigorously tested it — natolambert · 2026-10-27