I Don't Understand Tiktok, I Don't Understand Tick Talk.

The algorithm with which clip was trained was designed in, i believe, 20 19. And so there are big agroshmic challenges still in kind of just figuring out what needs to be done so that we can even train at that scale. I think teeto, things that's most interesting about lak mexio or some of thes other unsupervisod oral stuff, is that it's at least a first hint towards a thing that like o, if you let this keep training for a really long time on its own, you would hope that it would get better. Whereas a lot of other aral things, you just have one reward function and don't discover new states