Comment on NSF and Nvidia award Ai2 $152M to support building an open AI ecosystemparentComments−ivape1yI don't think we can say that until we hear how Genie3 and Veo3 were trained. My hunch is that the next-gen multi-modal models that combine world, video, text, and image models can only be trained on the best chips.
Comments
I don't think we can say that until we hear how Genie3 and Veo3 were trained. My hunch is that the next-gen multi-modal models that combine world, video, text, and image models can only be trained on the best chips.