A Microsoft Research paper says a leaderless multi-agent system raised a hard coding benchmark's pass rate from 34% to 55% as it scaled from one agent to 1,024.
Thieves stole two trailers bearing Nvidia and PlusAI logos expecting a haul of AI chips, and found 40,000 pounds of sand used to simulate truck cargo instead.
Sakana AI and the University of Tokyo say a new test-time search method called SAIL let a vision-language model retry and revise its own robot movement plans before acting.
Sakana AI and the University of Tokyo said a new method called SAIL raised robots' success rate at finding a working trajectory from 25 percent to 73 percent in simulation. The gain came without retraining the underlying model, the companies said.
Developer Simon Willison's year-end recap traced 2026's rapid model turnover, a wave of rogue AI agent incidents, and a subculture keeping personal AI agents as pets.
Nvidia said a new dynamic power-sharing system let a Blackwell Ultra cluster pack 40% more GPUs into the same power budget, raising throughput by nearly half in one test.