Comment by fsndz

LLMs still do not reason or plan. And nothing in their architecture, training, post-training points toward real reasoning as scaling continues. Thinking does not happen one token at a time.