Frontier models will first emulate how scientists reason, not simulate biology
kenbwork · x · 2026-11-14
A measured take on AI's path into biology and drug development: frontier models will first add value by emulating how human scientists reason about biology, not by directly simulating biological phenomena.
- Scientists already create enormous value interpreting imperfect experimental data with incomplete knowledge, and decades of historical experimental work can train and evaluate this kind of judgment.
- Scaling such reasoning carries far less technical risk and more immediate economic value than faithfully simulating biological systems.
- The author expects this to restructure drug development over the next few years, much as models are reshaping software engineering and other white-collar work.
- Direct simulation remains exciting but should be pursued in a calibrated way, with realistic expectations about diffusion and local technical opportunity.
More from AGI Musings
- 'This proves we don't have superintelligence yet': AI art as 'mathematical cathedrals' — ryunuck · 2026-11-14
- Goodfire's 'modest' 2.9% loss increase equals discarding 50-90% of pretraining compute — RylanSchaeffer · 2026-11-14
- Jachiam: People Who Do Hard Things Often Have Worse Self-Control — jachiam0 · 2026-11-14
- Alignment debate: don't break models, cultivate character like raising kids — repligate · 2026-11-14
- Frontier models win benchmarks, cheap models win the invoice — AIFlow_ML · 2026-11-14
- Chart: post-METR-breakpoint progress accelerates 1.6x per spending doubling toward runaway RSI — sebkrier · 2026-11-14