well you are assuming that there's not going to be any new progress and that we are going to be stuck with whatever LLM version we have currently
well you are assuming that there's not going to be any new progress and that we are going to be stuck with whatever LLM version we have currently
No. I assume that
* we're already close to the ceiling of LLM capabilities: LLM providers probably already consumed everything they could. Plus there have been no large improvements for a while, only small and incremental. I believe Mythos-too-dangerous-to-release is a marketing bullshit, until proved otherwise.
* people generate overengineered slop at light speeds. If that is the training base for future models, I doubt they're going to improve significantly, rather quality is going to stagnate at best.