What's the Half That's Missing?

M m d and all these things, they're terrible generatie models. Part of ies say they're missing the replacement of a kurnel by anetnetwork. The thing that breaks kind of the girls of dimention, and why nera networks workin calmantion on spaces is current methods just kind of suck. And another thing that's missing, coming back to the original point, is good bench marks for it becas you should only use uncertainty or diversity,. Only when you come a long striam task. People have been using antrope machine learning for many, many, many decades, specially in the basian community, this is all wrong.