Hacker News new | ask | show | jobs
by Ey7NFZ3P0nzAe 197 days ago
Well, behind "models" not "langual models".

Of course models purely made for image stuff will completely wipe it out. The vision language models are useful for their generalist capabilities