openai saying eth is false is a big deal, and i absolutely agree. the implications could be huge, hundreds of computer science papers were built on eth being true
openai claims eth is false, with an algorithm that runs in sub-exponential time for 3sat (NP-complete!). that's surprisingly faster than brute force. and it's not just sat, the same "assuming eth" wall sat under vertex cover, dominating set and hamiltonian cycle, the math behind network monitoring / route planning
also the proof even comes with an actual algorithm this time. so what’s next, who turns it into something practical?
The world is changing fast. Humans seem to very quickly become the bottleneck in scientific and technological progress.
The way coding has changed and the way scientific breakthroughs are changing (the exponential time hypothesis got disproven by AI!!!) is astounding.
Excited by what is to come. So optimistic about how things can turn out. It depends on us to utilise this new power in the right way.
@choblin29 i’m impressed once again that proofs of such tremendous complexity don’t even require 10000 agents running in parallel anymore (and maybe less data too, if the story about sketch is to be believed)