Right. A flying machine doesn’t need to understand anything to fly. It’s not even clear what it would mean for it to do so, or how it would fly any differently if it did.

Same with the AI machines.

Understanding is not something that any machine or person does. Understanding is a compact label applied to people’s behavior by an observer that allows the observer to predict future behavior. It’s not a process in itself.

And yes, we apply this label to ourselves. Much of what we do is only available to consciousness post-hoc, and is available to be described just the same as the behavior of someone else.

  > Understanding is not something that any machine or person does.

I can model the wind and air flow across the surface and design airfoils.

I can interpret the mathematical symbols into real physical meaning.

I can adapt these equations to novel settings or even fictitious ones.

I can analyze them counterfactually; not just making predictions but also telling you why those predictions are accurate, what their inaccuracies are (such as which variables and measurements are more precise), and I can tell you what all those things mean.

I can describe and derive the limits of the equations and models, discussing where they do and don't work. Including in the fictional settings.

I can do this at an emergent macroscopic level and I can do it at a fine grain molecular or even atomic level. I can even derive the emergent macroscopic behavior from the more fine grain analysis and tell you the limits of each model.

I can also respond that Bernoulli's equation is not an accurate description of why an airfoil works, even when prompted with those words[0].

These are characteristics that lead people to believe I understand the physics of fluid mechanics and flight. They correlate strongly with the ability to recall information from textbooks, but the actions aren't strictly the ability to recall and search over a memory database. Do these things prove that I understand? No, but we deal with what we got even if it is imperfect.

It is not just the ability to perform a task, it includes the ability to explain it. The more depth I am able to the greater understanding people attribute. While this correlates with task performance it is not the same. Even Ramanujan had to work hard to understand even if he was somehow able to divine great equations without it.

You're right that these descriptions are not the thing itself either. No one is claiming the map is the territory here. That's not the argument being made. Understanding the map is a very different thing than conflating the map and the territory. It is also a different thing than just being able to read it.

[0] https://x.com/BethMayBarnes/status/1953504663531388985