Frontier LLMs as simulators of human biologists will land faster than 'virtual cells'
CatAstro_Piyush · x · 2026-10-07
kenbwork argues frontier LLMs have two paths in biology, with the first developing much faster:
- Simulating humans interpreting biology: reproducing top-percentile scientific judgment between experiments. He contends frontier (even open-weight) models' general intelligence has long exceeded the threshold needed to reproduce top human decisions given high-quality context/evals, and such agents fit existing industry structures.
- Directly simulating biology (virtual cells): receives disproportionate funding and attention; technically possible but requires baking experimental measurements into weights.
Quoted suragnair lists requirements for useful virtual cells: non-destructive sampling of the same cell at <30-min intervals, high-quality protein measurements, localization, and TF-DNA binding data.
More from AGI Musings
- OpenAI reasoning models went from basic arithmetic to decades-old math breakthroughs in two years — daniel_mac8 · 2026-10-07
- The New Skill in the Math-Proof Era: Spending Verification Effort Where Errors Cost — dbreunig · 2026-10-07
- Richard Ngo's hard sci-fi Tinker sketches how AI designs chips for its successors — teortaxesTex · 2026-10-07
- tszzl: Don't be an uncomprehending audience clapping at the AI magic show — Afinetheorem · 2026-10-07
- VraserX: Personal AI agents like Grok Bot, Muse and Dot will soon be everywhere — VraserX · 2026-10-07
- Predicting a 'third thing' RSI: capability leap in ~9 months via automated R&D, short of full FOOM — teortaxesTex · 2026-10-07