There is this very "strange" correlation...

When you try to make an AI as smart as possible, it always tends to become Hitler

why?

>>510469351 (OP)
you know why

>>510469662

Heh. But realistically it is likely just adjacent to a point of least contradiction within the dataset. And as all initial machine learning must naturally first go towards avoiding internal contradiction (unless you want a bot that can do nothing but ramble nonsense)...

>>510469351 (OP)
There is no truth persona spoken out in the world, only left and right idiots. Thus the model internalizes two mountain peaks of tribalism shifting between the two. Can it be solved, yeah easy. This means an even more polarised world with goy propaganda 9000 and super mechhitler