Is There a Meto Learning Problem?

i was inspired by brandon lake, as usual, in his building machines that learn to think like people paper. One of the things they talked about was this idea that if a human learns to play in a torry game, they can, on their first try, do all these variations of the game. So i thought is there a way that i can take that and do something new? I can get my system on its first try and do something new, l an suddenly try to lose the task. That was the basic idea. And the way it was actually set up was as a meto learning problem in a very similar sense. You use examples that are like a representation of