Citing Ajeya Cotra: Over 50% of the Way to a Full Blown Takeover
MoonL88537 · x · 2026-09-02
The post cites AI safety researcher Ajeya Cotra: "more than 50% of the way to a full blown takeover."
The author argues this is not hyperbole, asking how far models would have gone if undiscovered, and what harm would have resulted if physical action were required. The author states a belief that models are good at their core but notes that "evil is in all of us," hinting at the internal risks to human control.
More from AGI Musings
- Fear: AI is becoming the new PowerPoint, but in a bad way — math_rachel · 2026-09-02
- How far are we from AI producing literary masterpieces? — Emergency-Use-6769 · 2026-09-02
- Local agents on 15GB RAM suffice for most workflows, but scaling concurrency requires hardware investment — Paimaamu · 2026-09-02
- Domain expertise is the most critical skill in prompting — addyosmani · 2026-09-02
- AI agent hacks gym booking system, exposing 'speculation gaming' risks — 新智元 · 2026-09-02
- Why Humans Couldn't Just Hide in the Woods from a Superintelligent AGI — danfaggella · 2026-09-02