It must be a pretty exciting time for you. I can't imagine how much of a difference it would make when vision models like this can be run at the edge in real t

It must be a pretty exciting time for you. I can't imagine how much of a difference it would make when vision models like this can be run at the edge in real time for blind users, including proper triage of relevant scene information narration. Wild to think about the long tail of accessibility and how that's going to be wildly different with the improvements to generative AI.