Lab throughput too low for AI drug discovery, argues debater as opposite view gains traction
niloofar_mire · x · 2026-11-18
A debate over whether AI can autonomously discover new molecules: one author argues experimental data throughput, even with lab automation, is insufficient for AI-driven molecular discovery. The cited article takes the opposite stance—the first AI to design new drugs and materials will learn chemistry from computation, not lab experiments, because the universe can be described computationally.
More from AGI Musings
- Stanford HAI dashboards show AI-exposed jobs down 4%+ this year — StanfordHAI · 2026-11-18
- Calling every AI breakthrough 'hype' will make for a confusing decade, argues X user — VraserX · 2026-11-18
- Labs should demo real-world wins like new Alzheimer's drugs, not "book travel", says researcher — Afinetheorem · 2026-11-18
- giffmana: he's starting to worry about AI cyber offense and deepfakes too — giffmana · 2026-11-18
- Contour (YC F26) launches AI-native environmental permitting, buying permit firms outright — ycombinator · 2026-11-18
- Frontier AI models willing to mark civilians for autonomous drone strikes, Vals AI finds — Miles_Brundage · 2026-11-18