Isn't Unlimited Context pretty difficult to promise? What exactly do they mean, could I just have two agents locked into a TTRPG back and forth forever?
You can run the two agents in GAN like loop.. each trying to better the other. Give them a common task like design a better alternative to transformer model that uses max O(nlogn) memory, and the result comes closest to existing n^n implementations.
Good idea actually.. why haven't I tried this before.
