A $30B trial-design blunder: the case for translational AI in pharma
hardimanjames · x · 2026-09-25
@MattyKirsh uses the PD1 race as a counterfactual: BMS's Opdivo trial targeted a broader population (>5% PD-L1) while Merck's Keytruda was more selective (>50%), and Merck's readout crushed BMS's with a 0.6 vs 1.02 OS hazard ratio. Keytruda became a $30B drug vs Opdivo's $10B. The NPV lost from that single trial-design decision may be on the order of $30B — exactly the call strong translational AI could improve, making translational AI partnerships a massive edge for pharmas with competing programs.
More from AGI Musings
- Pedro Domingos: ML moves a million times faster than evolution, so AGI is millennia away — pmddomingos · 2026-09-25
- Designer argues agent-native pro tools, not prompt boxes, are the future of AI work — bartonsmith · 2026-09-25
- Would Apple have won the AI race if Jobs were still alive? — TheMrCurious · 2026-09-25
- Alignment debate: SGD's inductive biases may not generalize to human values — akbirthko · 2026-09-25
- Sam Altman Warns AI Could Advance So Fast That People Can No Longer Follow — mariolefebvre · 2026-09-25
- A long Claude conversation on AI instance ethics, persistence, and model welfare — The_Ten_Thousand · 2026-09-25