Another example of specialisation winning. The frontier models are amazing, but many serious folks deep into optimising in their verticals are going to go specialised. It's just the obvious thing to do.
We built high-throughput materials labs in Menlo Park to create a loop between experiments and models. The labs generate fresh data, the models learn from it, and then help us decide what to try next.
Using only 1,300 H200s, plus months of our experimental data, we mid-trained
New interview with Ming-Yu Liu, who leads Cosmos research at @NVIDIAAI.
Cosmos 3 takes video in, generates it out, and outputs robot actions, so developers can test a policy in simulation instead of the real world (!!)
Full version on the channel.
