Automated research still needs a million research-years, but simulation advantages are real, researcher argues
gleech · x · 2026-10-13
A thread between researchers gleech and krishnanrohit on the feasibility of automated research and alignment. Gleech lists simulation's huge advantages over real science: a complete closed causal graph, no measurement error or cost, full output distributions, a stationary target, exact gradients with a complete locally linear model, causal evidence without IRB blockers, and known training data.
Yet he adds that it will still likely take "a million research-years" with heavy serial dependencies smearing it across calendar time, and describes automated alignment as "the act of a desperate species" — a bet made under time pressure.
More from AGI Musings
- Pediatrician: The Biggest Threat to Kids Is Lost Connection, Not AI — nordicinst · 2026-10-13
- Leave the Blank Page Blank: Offloading First Ideas to AI Skips the Thinking That Matters — DrKavner · 2026-10-13
- Richard Hamming's final 1995 course, filmed at 80, is now fully free online — thisdudelikesAI · 2026-10-13
- VC: AI slop cannons are killing the one-liner influencer playbook — StewartalsopIII · 2026-10-13
- Michael Jordan's new essay declares superintelligence is science fiction — gleech · 2026-10-13
- 51% of AI researchers put >10% on extinction-level outcomes, data pushback on skeptics — gleech · 2026-10-13