Anthropic's protein design PR dissected: open-source models did most of the lifting
teortaxesTex · x · 2026-08-19
Commentators dissected Anthropic's new protein design results, sharing a "no-PR" version: Claude orchestrated a suite of open-source protein design models — PXDesign, RFdiffusion, Genie, BoltzGen — using a 30k-token expert prompt and 12,500 H100-hours, designing binders against 14 of 15 targets with 22–35% hit rates versus a 10–15% baseline.
The core critique: those open-source models came from the Baker lab, Columbia, MIT and ByteDance Seed, many already wet-lab validated before Claude touched them, and some report similar numbers on their own. The orchestration is genuinely impressive, but the ceiling is set by these generators — all sharing a single PDB-shaped distribution — and the narrative overstates Claude's contribution.
More from Models
- Brett Adcock Says Hark Has "Sci-Fi Movie Hardware" in the Lab, AGI Needs More Multimodality — ChrisGPT · 2026-08-19
- DeepSeek's demand destruction strategy hasn't worked yet — zephyr_z9 · 2026-08-19
- Qwen3.8 27B Breaks Size-Intelligence Curve, Resets Consumer Hardware Expectations — smtabatabaie · 2026-08-19
- Analysis Suggests OpenAI's 'Astra' Model Unaffected by RL Slowdown — haider1 · 2026-08-19
- Inkling-Small hits Agent Arena: matches Inkling performance at half the price — simonguozirui · 2026-08-19
- LFM2.5 Models and Embedding Now Available on OpenRouter — maximelabonne · 2026-08-19