@VraserX: Scientists at Lila initially thought their AI workflow had gone wrong. The materials it selected looked unpromising. Th…
Summary
Scientists at Lila initially doubted their AI workflow's material selections, but continued testing revealed a catalyst that survived over 1,000 hours in acidic lab conditions, highlighting the value of exploring unconventional ideas in research.
View Cached Full Text
Cached at: 09/28/26, 03:34 PM
Scientists at Lila initially thought their AI workflow had gone wrong. The materials it selected looked unpromising.
They kept testing. Lila now reports that its leading catalyst survived over 1,000 hours in acidic lab conditions.
I’d give every research lab a budget for testing ideas its senior people think won’t work.
How much would you set aside for the possibility that your best experts are wrong?
Similar Articles
@VraserX: The human advantage inside frontier AI labs is increasingly described as research taste. Choosing the right problem. Kn…
The human advantage in frontier AI labs is increasingly described as research taste, which involves choosing the right problem and knowing which results to trust. If AI models develop this ability, it could transform the research loop completely.
AI helps design new materials that work in the real world
MIT researchers introduced a framework called CrysVCD that enhances the chemical stability of AI-generated materials by enforcing valence rules, making them more suitable for real-world applications like computer chips and rockets.
@rohanpaul_ai: New Anthropic research shows AI agents may look brilliant at code, but in biology they can fail before the science star…
Anthropic research reveals that AI agents struggle with biology databases, producing highly variable answers for the same query (e.g., Ebola sequence counts ranging from 5 to 106 vs. expected 266), but adding a repeatable retrieval tool significantly improves consistency and accuracy.
Stanford/Princeton AI4S unveils LabOS² -the agentic AI system that spanned from dry-lab planning to wet-lab execution, using physical AI to assist scientists - now is capable of performing fully autonomous cell culture workflows.
Stanford and Princeton's AI4S initiative unveils LabOS², an agentic AI system capable of fully autonomous cell culture workflows spanning from dry-lab planning to wet-lab execution.
@robertnishihara: This talk is uploaded now! https://youtube.com/watch?v=Sv0wfbPp5LI…
Lila Sciences is developing an integrated AI-driven platform to automate scientific experiments, using AI models and robotics to tackle unverifiable hypotheses in science, with a vision to create scaling laws for scientific intelligence.