Random thought: If we are seeing an explosion in AI research and models are going to become exponentially better, that probably means we will hit a wall much sooner than expected in hardware doing inference. Does that mean, assuming also that there will always be people solving
