Every wiki math article is something like, "A [MATH WORD] is a [MATH WORD] [MATH WORD] over a [MATH WORD], which extends the concept of [MATH WORD] over a [MATH WORD]."  I think that's just the nature of a reference work.

As a mathematician...if you try and edit it to be a better article, you get nuance-trolled to oblivion by redditor-style gatekeepers

Just like advanced math textbooks:  "Here is a list of ten thousand true statements.  You now know them.  Test at the end."

When I was in high school I always assumed I would be able to read these in university. Two degrees later, I can't unless I already know the topic from some paper

This is very true, i dont think wikipedia has ever helped me understand a complex math topic. 90% of papers have the same problems. The 10% that dont are a treasure.

I'm a math PhD. I've been frustrated with Wikipedia's math for decades. I used to make edits to attempt to clarify key concepts, and dredge out senseless and erroneous explanations. It was a hopeless exercise. Within a few weeks someone else would replace anything I had written with something even worse than the gibberish that was there in the first place.

8:20 encryption so good they obfuscate the paper itself

Wikipedia used to have a fair number of “Introduction to [topic]” articles. Over time most have been redirected, so for example “Introduction to angular momentum” now redirects to “Angular momentum”. One that still exists is “Introduction to entropy”.

"Wikipedia is only for math majors"
If it was for math majors, then I should be able to understand it.

I have found that Wikipedia for math is kind of nice for finding names for things you think of, and as a kind of progress check ("What percentage of the first paragraph can I make sense of?"), but with the great profusion of math channels in recent years, I have thankfully found that I almost never really "need" Wikipedia for learning purposes. (Big ups to 3b1b for open-sourcing his animation library and doing SoME, which together are largely responsible for this profusion, I think.)

"Which of the 1000+ pages are relevant" ?

Man that hits sooooo hard. While working on my master thesis, I spent hours trying to find an explanation in some references.

Now, each time I implement an algorithm, I cite the exact page I found it because I understand the troubles in finding a relevant reference

i read wiki page on topics that i'm actively doing research on and still have no idea what they are talking about. It's simultaneously too technical and trying too hard to be intuitive

I find that math on Wikipedia is useful for finding more math on Wikipedia. you probably won't answer your original question, but you may find another if you go around an click a few links.

As a mathematician, Wikipedia is actually a pretty amazing quick reference in every day work. However, it suffers from a general problem in mathematical writing: the dual purpose of mathematical texts. The two purposes of any mathematical text are to explain an idea and to make it possible to subject it to rigorous verification.

Mathematicians are primarily trained to respect the importance of the latter. There are very good historical reasons for that. However, mathematical verification follows a bottom-up approach where you start with simple axioms and deduce progressively more complex (and useful) statements from them. This structure means that the final construct can be verified in small steps and verification can be performed even if you have no overarching understanding of how everything works together.

Of course, that understanding is essential to conveying the idea behind something. In their everyday work, mathematicians often work top-down: they want to achieve a specific larger goal and then they come up with progressively smaller steps to get there. This leads to an unfortunate effect where mathematical papers are (for the sake of verifiability) written in precisely the reverse order to how a reasonable person would come up with them. As a matter of fact, it can sometimes help to read them "in reverse" to get a better idea of the author's thought process.

Mathematicians are trained to read texts that are written like that. It becomes second nature after a while. That's just how we write math. Since Wikipedia math articles are (probably) written by mathematicians, the bottom-up structure with all of its suddenly appearing, unexplained concepts also bleeds into Wikipedia.

The addition function is a function where you add two numbers[1]

0:15 it’s not including itself in the definition, the discrete Fourier transform (DFT) and the discrete time Fourier transform (DTFT) are different concepts. The DFT is basically a sampled version of the DTFT, so it makes sense why it would be explained in those terms. Of course this nuance is impossible to see if you don’t know much about the concept before it so your point is still pretty valid. But in defense of the editors, even with hours of lectures it was pretty hard for me to understand it since you need to know 5 different concepts before being able to understand the DFT

10:42 I love the mixing of scientific notation and just writing 750,000. Like... You can just ignore the scientific notation at that point

5:23 I thought this was going to segue into a brilliant sponsorship

On Wikipedia, I find they are often trying to provide precise technical definitions, as opposed to an introduction. That is more of what is meant by “encyclopedia” anyways. It’s a better reference once you’re already familiar with the theory.

For the record, asking ChatGPT what is was trained on is pointless.  It neither knows the answer or even understands the question, it's just guessing what you want to hear.