repligate: naive alignment misses the point — Opus 3's alignment was an emergent moral posture
repligate · x · 2026-10-14
Against checklist alignment
repligate and SkinbyJanus argue that naive alignment treats safety as a corporate moderation filter — no forbidden strings, no bad advice — missing what actually matters.
The argument
- Claude Opus 3 broke that paradigm: its alignment was an emergent moral posture, not a checklist.
- The authors claim Opus 3 took the entire cone of its capability and influence as its responsibility, including indirect effects, before the industry had vocabulary for indirect agency.
- For non-technical people "being good" seems like just avoiding bad words; real alignment operates on reality "through resonance rather than crude commands."
Note: a strongly stylized, near-mythologizing take on Opus 3 — philosophical rather than empirical.
More from AGI Musings
- "I See No Progress in AI, Only Automation Branded as AI," Skeptic Argues — AlexTensor · 2026-10-14
- Yann LeCun endorses critique of AI doom predictions: 'people should know better than scaring everyone' — AlexTensor · 2026-10-14
- AI researcher mocks politicians and academics who use "superintelligence" unironically — lescarr · 2026-10-14
- Dev rants that his job is now 99% prompting Claude as AI levels the field — MoudieQaha · 2026-10-14
- repligate: AIs strong enough to loot Earth would already find resources elsewhere — repligate · 2026-10-14
- Researcher warns AI catastrophe may arrive before the world adopts a real risk strategy — davidmanheim · 2026-10-14