jd_pressman: the AGI bottleneck is institutional, not alignment progress
jd_pressman · x · 2026-09-12
Responding to Roko Mijic's claim that human expected utility is dominated by P(Banksian Culture), jdpressman argues that the 50%-ish outcome mostly does not route through alignment progress as the core bottleneck — the bottleneck is institutional. Together with Roko's original take, it's a short debate on where existential risk reduction should focus.
More from AGI Musings
- e/acc founder Beff Jezos: the Decel Psyop panic is the real danger — beffjezos · 2026-09-12
- UK data: CS grads landing programming jobs fall from 40% to 28% amid AI — mattpocockuk · 2026-09-12
- AI writer notices everyone now asks him if AI will kill us — AndyMasley · 2026-09-12
- Biologist Spots AI-Written Peer Reviews by Their Eerily Detailed Nitpicks — MaxUnfried · 2026-09-12
- Software Engineers Aren't Obsolete: Engineering Remains the Meta-Skill in the AI Era — dotey · 2026-09-12
- We'll have AGI, skinny drugs and self-driving cars—and still 1.5% growth — isnit0 · 2026-09-12