Hacker News new | ask | show | jobs
by buescher 779 days ago
I think it has to do with who trained them, and how, and not the technology itself. Just wait until the capability of training on large corpuses is more widely available…