Same. We've all fooled ourselves into believing that an LLM / stochastic process was finally solved based on a good result. But the sample size is always to low

Same. We've all fooled ourselves into believing that an LLM / stochastic process was finally solved based on a good result. But the sample size is always to low to be meaningful.