Redditors push back on researchers' 10% AI extinction odds: what's the actual scenario?
Fit-Stress3300 · reddit · 2026-09-12
A Reddit user questions why some AI researchers assign roughly +10% odds to AI destroying the world, arguing the scenarios behind such estimates are never spelled out concretely. The author recalls the 2014 paperclip-maximizer thought experiment but contends that LLMs and modern DNNs, with the known limits of RL, look very different from those doomsday abstractions. The thread debates whether existential-risk arguments still hold up today.
More from AGI Musings
- Ethan Mollick: METR long-horizon benchmark effectively saturated as agents hit 18-week tasks — emollick · 2026-09-12
- ResNet by hand: a 10-step walkthrough showing how one identity matrix enabled thousand-layer networks — ProfTomYeh · 2026-09-12
- On AI's punch-card phase: most programmers never touched a real computer until the 1970s — _AustinCalvert_ · 2026-09-12
- awilkinson: We're still in the punch cards phase of AI — and it's only year 3 — _AustinCalvert_ · 2026-09-12
- Investor: last time 'we're all going to die' alarmists won, it cost hundreds of thousands of lives via anti-nuclear fear — firstadopter · 2026-09-12
- John Carmack slams big-lab doomers: "If you believe in risk, work in the open" — beffjezos · 2026-09-12