Rex,

> I suppose whether or not what LLMs can already do has any claim to fall under the "understanding" label (and/or "semantics") eventually boils down to a choice of what we want to call "understanding" <

Yes, but let’s not play semantic games, which is one of the maddening things about some people who work with AI. By understanding, throughout this discussion, I meant the human variety: conscious awareness of how to solve a problem or how things stand, as in “I understand the Pythagorean theorem,” which is different from “here’s an algorithm that demonstrates the Pythagorean theorem.

> Without going into detail, it is in applying both p-zombies and the Chinese Room to this question that I find that they're essentially the same. <

Sorry, I still don’t get it, but that’s okay.

> I am predisposed to use a experimentalist-centric view of understanding. If I were to go into a lab and look for correlates of understanding <

Understood ;-) but experimentalism has limits. Just because two systems behave similarly it doesn’t mean that the underlying mechanisms are the same. Hence the more comprehensive approach I have been articulating: experiment away, but also take into account what we know from neuroscience, developmental biology, evolutionary biology, and so forth. The more multifaceted the approach the more it becomes clear, I think, that AI’s “intelligence” has nothing to do with the human variety.

> We could call it something other than "understanding", but I wouldn't see the point <

The point is to make distinctions where there are differences. That’s both good science and good philosophy.

> It's not pure behaviorism because it's not just checking that the output looks superficially like we expect, but that the mechanism is also running in the way it (almost surely) must if it is the thing we're hoping to study. <

But you have been objecting all along that we don’t really know the mechanisms. So this still looks like behaviorism to me.

Anyway, we may be getting close to having explored this aspect as much as we can for now. Thanks for the engaging discussion!