On the "user" side, almost everything is covered. On the dataset side as well IMHO. The parts in the middle of the pipeline such as inference services it's not clear yet — but if there are crimes either side inference services would be under pressure to do something.

→ View original post on X — @alexjc