upvote
And there are entire categories of phenomenon that don't arise or are imperceptible in a simple context. Chaos is an emergent phenomenon that requires a surprisingly deep knowledge of a situation to predict where it will arise.

The way I understand the problem with explaining things in simple terms is this: We usually introduce quantum mechanics students to a toy model -the infinite square well- to build up intuition. However, depending on the teacher and the students, it may well take years for those students to recognize and understand that the boundary conditions we apply to literally every other QM calculation are violated by the infinite square well, and while we ultimately recover the expected conclusion via successive approximations, it takes a much deeper understanding of the material to understand why the "simple" explanation is actually wrong. That is, while it turns out that the solution to the infinite-square well is, at the limit, identical to a very deep square well, showing that is actually quite a bit harder.

The simple explanation causes incorrect intution about the situation.

reply
I'm not sure how you measure this in practice, when one way that explanations are simplified is to externalize a lot of information by reference to some specific terms of art (jargon, really). Where is the boundary for measuring the content, when we always have the implicit compression dictionaries of the speaker and listener to contend with?

This is also where the ironic part of simple explanations arises. Whether the explanation serves any purpose depends on the alignment of the speaker and listener background knowledge. An explanation between two experts versed in the topic may seem coherent, concise, and meaningful while it fails to teach the concept to someone new.

Sadly, it also works between cranks, who can agree on misleading or incorrect explanations. Or, sometimes novice listeners can believe they've learned, but they learn a different crank concept than what the expert believes was being states. A good teacher needs to be aware of these potential problems, attempt to ascertain the background knowledge of the listener, and adjust the explanation for that background rather than their own teaching knowledge. And followup testing attempts to ascertain whether the new concept got delivered or not.

Now that we have novices asking LLMs for explanations, I find this all quite troubling in a practical sense.

reply