AI agents still struggle to find new laws
A new benchmark tests whether AI agents can discover scientific laws from scratch, revealing they still fail at the hardest step.
A new benchmark tests whether AI agents can discover scientific laws from scratch, revealing they still fail at the hardest step.