Seems like every AI podcast can be summarized by "we don't really know what will happen, and we don't know what it means for a machine to be able to think"
This was painful to watch. Dwarkesh was worse than a clueless interviewer.
I didn't have this impression at all. Dwarkesh came across as knowledgeable enough about the subject and with enough confidence and energy to ask meaningful, challenging questions of an expert. There was some real and productive disagreement and I think this made for a more interesting podcast than most.
RL researcher says LLMs won't work, LLM researchers say RL won't scale. Meanwhile, the rest of us are just shipping products with whatever actually works today. This is like watching database academics argue about ACID properties while MongoDB just took over the world.
The db example is bad because most people in db and data world agree mongodb is not a great choice anymore. In fact mongodb today has actually adopted a lot of the characteristics of a sql db.