Comment by speedgoose

In one hand I think we should react quickly, one another hand maybe we should let people talk a bit more and wait for a bubble crash and better LLM inference hardware.