Bitter lesson vs biologists: LLMs may beat molecular ML in near-term bio breakthroughs
charlieharris01 · x · 2026-10-26
A route debate on whether AI can crack biology. The quoted post highlights a sharp divide between the AI camp (scaling laws, bitter lesson) and biologists, who find a "bitter lesson for biology"近乎 ridiculous and think reliable human toxicology simulation is 5-10 years away. Sam Sinai adds discordant points: near-term disease-solving gains will come mostly from LLMs' usefulness rather than molecular ML; bitter-lesson proponents for bio models are right but pathologically early—data is scarce and the wrong kind; scaling models on passively accumulated data is cost-ineffective since today's bio model outputs are GPT-2-level, and a bioinformatician with linear regression can often beat them, yet few benchmark against that.
Related event: Investor Essay Sparks Debate Between AI and Biology Camps on Drug Discovery(3 posts)→
More from AGI Musings
- Posting has power-crept into a basic skill; attention is the only scarce resource — tokenbender · 2026-10-26
- Dunking on those who believe in AI doom yet trust hardware verification plans — jessi_cata · 2026-10-26
- AI adoption up to 78% but developer trust down to 33%, Stack Overflow surveys show — agh_murad · 2026-10-26
- Nation-state vulns now have shorter half-lives, raising crisis instability — jeremiecharris · 2026-10-26
- Sci-fi plot nearing reality: AI tunes film scenes to viewers' biometric reactions — IgorBrigadir · 2026-10-26
- Professor catches student cheating with AI who left the prompts in the paper — venturetwins · 2026-10-26