SF Compute CEO @evanjconrad breaks down why auto research kills the AI talent moat: "A lot of people haven't woken up to auto research as just an early thing that probably will take off. My impression is that you're gonna get the models themselves that will train the models. Having the talent to do it doesn't really seem like the limiting factor in the way that it used to be" "Everyone is gonna train models because training models will be something that the agent does in the background, or…