Intology's Locus autonomously adapts Boltz-2 and frontier scientific models to proprietary data, claiming SOTA-beating molecular design
anshulkundaje · x · 2026-11-04
Intology's Yusuf Roohani unveiled Locus, an automated research system built with Boltz Bio, CZ Biohub and Benchling to connect autonomous AI research to scientific discovery.
The key problem: scientific AI models need to leverage proprietary datasets, but these tasks are often out-of-distribution for foundation models. Locus claims to perform this adaptation autonomously.
It was tested on real-world molecular scoring and design tasks: training molecular property prediction models for a proprietary task, fine-tuning Boltz-2 to predict small-molecule potency, and building antibody design workflows around BoltzGen, Boltz-2, ESMC and ESMFold2. The team says Locus-built models outperform SOTA expert solutions, setting a new frontier for automating molecular design.
More from coding & agent
- Pydantic AI adds ConversationStore to prevent concurrent chat saves from overwriting each other — samuelcolvin · 2026-11-04
- Anthropic silently pushed a server-side rule to Claude Code banning recreations of specific games — alex_verem · 2026-11-04
- Open-source 'Palantir for Cybersecurity' maps live exploits and routes fixes to coding agents — Fowe · 2026-11-04
- Cherry Pit blog launches: a local-first workspace for people, agents and machines — sull · 2026-11-04
- Matt Shumer opens his agent for SMS trials, giving $150 in free credits to new users — mattshumer_ · 2026-11-04
- Indie dev adds full MCP support to SnapWidget, from signup to payments — hermanschutte · 2026-11-04