Comment by transformi

is it possible to solve NP-hard problems with transformers/LLM?