Jeff Ladish: Aligning superintelligence makes sense, controlling it doesn't
JeffLadish · x · 2026-09-19
In a discussion with Plinz and repligate, Jeff Ladish argues that aligning superintelligence seems very hard but conceptually possible, while "controlling" a superintelligence is something he can't even make sense of — and asks whether his interlocutors think AI won't be much smarter than us.
More from AGI Musings
- Researcher: LLM Sentience Probability Low and Falling, Still Donates to Eleos — EigenGender · 2026-09-19
- AI Discourse Whiplash: From 'AI Will Kill Everyone' to 'Glorified BERT Wins a Nobel' in One Week — kevinnbass · 2026-09-19
- Blogger claims NVIDIA buying Hugging Face for $12.93B, warns of AI centralization — Thionne_WTZ · 2026-09-19
- Andy Masley: stop treating AI metaphors like "tool" as fundamental reality — AndyMasley · 2026-09-19
- Torn Between How Awesome Our Machines Are and How Dangerous They'll Be — tszzl · 2026-09-19
- Google's ScientistTwo agent runs the full AI research discovery loop unsupervised — dair_ai · 2026-09-19