On alignment and SURA critics: both can be shallower than they claim
JacquesThibs · x · 2026-11-01
- Responding to 'Against SURA,' the author argues two things can be true at once: alignment is harder than many AI safety 'experts' think, with many in AI safety never truly grappling with it—while many SURA proponents also lack deep understanding despite being strongly anti-AI.
- A wordplay note: Suras are gods, Asuras 'against the gods'—suggesting Scott positions himself as the rebellious Asura challenging the SURAs' self-appointed moral authority.
More from AGI Musings
- Peak creative age by field: math at 25-35, philosophy at 50+ — will AI follow the same path? — arjunrajlab · 2026-11-02
- Google Labs PM Director: Gemini 4 changed how Google builds products — petergyang · 2026-11-01
- Investor: Europe's cultural rejection of AI is harder to fix than its data center gap — AIandDesign · 2026-11-01
- AI bubble debate: dot-com bust as cycle start, not the top — RachelVT42 · 2026-11-01
- A 1-Hour Podcast to Catch Up on Open vs Closed, Distillation, and RL Envs — SergioPaniego · 2026-11-01
- Doom race: 10-40% survival odds cited as optimism, and that itself is grim — teortaxesTex · 2026-11-01