anon_ratw received

Life, Intelligence, and Consciousness are NP-Complete?
A know what you're thinking: anon's gone all kooky mashing together a bunch of mysterious things, wrapping it in some gold foil, and worshipping it as an idol. Maybe, maybe not. Hear me out.... posted 2mo ago with 16 replies received

Life, Intelligence, received

Well we know why no free lunch theorems don't block practical compression and induction: because reality is in fact quite structured and thus describable with simple inductive priors.... 2mo ago received

Well we know why no received

Well life is at least a solution to what has been, verified in reasonable time. Where it starts to break the model as you point out is where it starts to become dynamic and reflexive. The metagame is a whole new level of dynamism as well beyond what I was ... 2mo ago received

Well life is at leas received

This is a very fair caution. Nerd towers of abstraction are a great way to waste one’s time. In this case, the focus is less on complexity theory and more on this interesting thing which is the surprising tractability in practice of theoretically hard pr... 2mo ago received

This is a very fair received

Strictly speaking i suppose we should expect just one optimum and not necessarily a global one. On the other hand some NP complete problems are about solution existence and others are about solution optimality. I forget how one is translated into the other... 2mo ago received

Strictly speaking i received

But what if you did modern bitter lesson-pilled data driven AI with BSAT? Many AI systems reach *perfect* zero training error ie perfect overparameterized overfitting, and somehow still get good test error because for whatever reason the types of functions... 2mo ago received

But what if you did received