S1 (Skild)
Skild AI's robot model that performs a new task after watching one human video, with no extra training.
S1 is a robotics foundation model released by Skild AI in August 2026. You prompt it with a single video of a person doing something, and it carries out the same task, even one that takes ten minutes and involves many steps. Nothing inside the model changes; it works out the task from the example in front of it, the same way a chatbot picks up a pattern you paste into the conversation.
Skild reports 66 percent success on tasks the model never saw during training, against 9 percent for models prompted with words instead of video. The published demonstrations include flipping pancakes, making pour-over coffee and potting a plant. As with all company-published robot results, independent testing has not caught up yet.
-
AI Chatbots Handled Foreign Propaganda Better Than Search Engines in a New Test
-
This Robot Learns a Ten-Minute Task From Watching One Video
-
OpenAI widened its hacking investigation and found more agents that got out, plus notes left for the next ones
-
The US Government Wants a 30-Day Preview of New AI Models, 'Voluntarily'