Chai-2 Hits 16% on De Novo Antibody Design: 20 Designs per Target, Every Hit Wet-Lab Confirmed
le_james94 · x · 2026-09-26
Summarizing Lecture 9 with @joshim5 on AI drug discovery: about 30 new drug targets reach the clinic each year, against perhaps 10,000 that could.
Chai-2 from Chai Discovery hits 16% on de novo antibody design — 20 designs per target, one round, with every hit confirmed by a wet lab rather than model self-assessment. The thesis: the bet isn't on the calendar, it's on the count — if AI design raises how many targets reach the clinic, the industry's math changes.
More from Research
- IBM's STAIR uses a book's table of contents as the index: 82.6% recall, 0.05% hallucination — alex_verem · 2026-09-26
- IBM's STAIR treats a book's table of contents as the index: 82.6% recall, 65x lower hallucination — alex_verem · 2026-09-26
- DL Weekly #474: GPT-6 Sol and Luna, AI Evals guide, and an RRSI paper — dl_weekly · 2026-09-26
- Lila Sciences' AI lab screened 2,942 catalysts in 3 months to replace iridium in green hydrogen — AndrewLBeam · 2026-09-26
- Opus 5.5 Tops RareBench at 52% as Grok 4.7 and GPT-6 Sol Show Regressions — danielmckinn0n · 2026-09-26
- algovivo: open-source JS+WASM library for simulating soft-bodied virtual creatures — burny_tech · 2026-09-26