Product · mentioned 19 times
Robot foundation model from Physical Intelligence designed to work out of the box across many tasks, with zero-shot generalisation and reasoning in language and images.
Tap a name to go there · tap a line to see why they are linked
Minus One· 22 Oct
In Pi 0.7, image generation is done by a separate model whose input is produced by the language model, similar to how Gemini or ChatGPT call a separate model to generate a picture.Y Combinator Startup Podcast· 13 Aug
Chelsea Finn says Pi0.7 can imagine what a future image should look like and then try to accomplish it, which gave a quantitative improvement on shirt folding compared with not using it.