Benchmarking AI agents in biology takes center stage at Genentech
kenbwork · x · 2026-10-23
The author visited Genentech's campus to share principles of biology task design and how benchmark shapes change across multi-omics, biosecurity, and therapeutics. Engaged scientists took these problems seriously; benchmarks are positioned to become a first-class tool in this new ecosystem. Slides in the image show verifiability-violation examples for cell-type assignment tasks.
More from Research
- SlopCodeBench: Top Model Scores Just 16% on Long-Horizon Coding Agent Benchmark — ajratner · 2026-10-23
- Charter v0.3: a formally verified language for institutional rules with AI constitutional prompts — QuixiAI · 2026-10-23
- QuixiAI hails eFLINT, will redesign Charter as a superset of the formal rule language — QuixiAI · 2026-10-23
- ArchitectureIQ Benchmarks LLMs' Training Intuition: GPT-5.6 Sol Hits 76.4% vs 33% Chance — ZimingLiu11 · 2026-10-23
- Claude Links Groundwater Pulse Pumping to the Sleeping Brain's Waste Clearance — abstrusejoker · 2026-10-23
- 345K trajectories reveal LLM agent personalities: Claude most planful, GPT-5.5 least energetic — mdredze · 2026-10-23