Comment by hammyhavoc

Why would they progress significantly than where they are now? An LLM is an LLM. More tokens doesn't mean better capabilities, in fact, quite the opposite seems to be the case, and suggests smaller models aimed at specific tasks are the "future" of it.