Robot Watches One Demo, Hits 59% Success Rate
With no fine-tuning, Generalist AI's GEN-1.5 robot learns a new task from one video demo in its context, hitting 59% success versus 83% after training.
3 verified stories covering Frontier models, product updates and industry developments.
With no fine-tuning, Generalist AI's GEN-1.5 robot learns a new task from one video demo in its context, hitting 59% success versus 83% after training.
Prime Intellect ran 153 fully autonomous research trials across 18 frontier models on a nanoGPT speedrun. The best model closed 82% of the human record gap, but not one of them found a genuinely new method.
A new U.S. executive order invites voluntary government access to frontier models up to 30 days before launch, while avoiding a licensing regime.