It seems like the other part of this is the ability of LLMs to rationalize even answers that are not rational (he says piling BS on BS). That has previously bee

It seems like the other part of this is the ability of LLMs to rationalize even answers that are not rational (he says piling BS on BS). That has previously been a very human (and quite distracting/confusing) activity and it's very effective on most people most of the time.