I like to think that a “simple” explanation doesn’t necessarily have to be a short one. 3blue1brown is a good example of that. It took multiple videos for him to explain LLMs, but it’s simple and relatively easy to understand.
Actually, Feynman himself is a good example of that as well. His explanations are rarely short, but very simple. He has a series of YouTube videos where he explains all kinds of concepts like fire, but it still takes him at least 15 minutes to explain it.
So I’d argue that thinking it needs to be condensed into one or two sentences is getting the meaning of the expression wrong.
A thing being more or less complex refers to the amount of information needed to encode its description.
"Simple" therefore means "reduced in information content".
The topic being explained has a certain minimal information content, and if you reduce beyond that, you are telling a fiction which bears analogies to the topic.
A simpler description is not necessarily short; rather, it is a description of something simpler.
When descriptions (of a given complexity, being held constant) are short, that can make them harder to follow, because encodings have to be clever in order to achieve compression, and that has a complexity of its own.
Fully agree. Some of the best textbooks I’ve had are verbose and try to explain the same concept using different words several times.
I thought a lot of those Feynman videos on YouTube were AI fabrications. How do you know which ones are real?
That's my excuse for being long, verbose, and rambley about everything :)
My favorite Feynman video (which the OP references but completely misunderstands) is this one: https://m.youtube.com/watch?v=MO0r930Sn_8
It's not so much an explanation of magnetism as a meta-discursive about explanations, and why a "simple" explanation is often not really possible. (In this case because our notion of what is weird and requiring of an explanation ("things pushing things at a distance") and what is normal and not requiring one ("the glass doesn't fall through the table") aren't scientifically grounded).
I think it's precisely the situation seanhunter is describing in their example