Post #1734 231 Sep 11, 2026, 09:10 UTC Модели показывают видос действия и робот повторяет его, без изменения весовhttps://youtu.be/cqwKceUSZ5Q YouTube Robots Just Had Their GPT-3 Moment In-context learning for robots has been a long-anticipated capability, as it could allow robots to quickly adapt to new tasks without additional training, and Gen-1.5 might have just demonstrated that it is possible for robots at complex tasks. Gen-1.5… 🔥 3 🤯 1