Latent Space (Swyx)
7/1/2026
The original headline is about diffusion research in drug discovery, not LLMs. Let me rewrite it following the rules.
Original: π¬ The Coolest Diffusion Research Isn't in LLMs β Evan Feinberg & Sergey Edunov, Genesis Molecular AI
Short summary
Meta's Llama lead Evan Feinberg and Sergey Edunov founded Genesis Molecular AI to apply diffusion research to drug discovery. Their PEARL model achieved zero-shot protein structure prediction wins, suggesting major potential when molecular co-folding accuracy improves.
- β’Llama lead shifted from LLMs to molecular AI drug discovery
- β’PEARL model shows zero-shot success on protein folding (OpenBind)
- β’Diffusion models may unlock breakthroughs in biotech beyond language models
Generated with AI, which can make mistakes.
Is this a good recommendation for you?


