AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Cloud Two: A New Model for Vision and Language Action
It's quite surprising because constitutional AI is not fundamentally that different from user techniques, they write this constitution set of rules for the language model. It seems like it's telling us something almost fundamental about the difficulty of aligning these models. Next story, RT two new model translates vision and language into action. This is about the paper RT two vision language action models transfer web knowledge to robotic control.