Daily links: AI successionism debate, pluralistic AGI, and AI automation in physics research
gleech · x · 2026-11-19
gleech shares his Substack daily links (2026-11-17), a roundup of AI discussions:
- Alignment debate: Julia Willemyns argues against AI successionism and CEV-style alignment, favoring preserving human moral agency and the ability to keep making mistakes; Geohot envisions pluralistic AI with "millions of independent power centers"; Kelsey Piper fears a bland post-AGI dystopia, while the author counters that pluralism can't be maintained without accepting the possibility of overthrow.
- Physics: Dan Kagan-Kans on likely impacts of AI automation in physics research, given superhuman math performance.
- Technical: Morgan Sinclaire's series examining next-token prediction through an information-theoretic lens.
- Hiring: Vaishnav Sunil on assessing candidates in the LLM era; the author quips that Mercor is the inverse of an AI-hiring startup — using excess money to hire only the most legible candidates while destroying the cheapest verifiable signals for juniors.
Related event: Daily AI Links: Alignment Debates, Pluralistic AGI, and Rethinking Mercor(2 posts)→
More from AGI Musings
- Erik Hoel: provably non-conscious AI would permanently reserve parts of the economy for humans — erikphoel · 2026-11-19
- Follow-up: a transformer has no memory, self, or cognition, dev argues — gerardsans · 2026-11-19
- AI doomers' statement blasts Anthropic's 'power grab' — and invites irony charges — DrTechlash · 2026-11-19
- The site-builder era is repeating: people will hire humans to use AI app builders, says dev — sull · 2026-11-19
- Blogger goes contrarian: AI isn't currently moving toward higher general intelligence — LucaAmb · 2026-11-19
- Dev pushes back on AI anthropomorphism: context is not memory in transformers — gerardsans · 2026-11-19