OpenAI, @wtgowers , and more(?) report that there's a better configuration of points for producing unit distances. In fact, an entire family! There's this blog post that comes up first when I search: https://openai.com/index/model-disproves-discrete-geometry-conjecture/

They helpfully picture the point configuration that is no longer champion (the best set of points humans could find with as many pairs as possible being unit distance apart). As far as I can see, they don't include a picture of any of the new point configurations!

Internet. I have a task. Somebody please plot the new configuration and show me a picture. It would make me very happy.
#maths #unitdistance #erdos #mathConjecture #mathart

Y should I Sine the contract by myself when my X can Cosine and pay the Tan grand on my car loan

Stolen from a friend, who knows who they stole it from. Mnumonic to remember that sin gives the Y coord, cos gives the X coord, of points on the unit circle.
#precalc #teaching #maths #trig

I have just learned about the delightful mathematical research postcards! https://secure.math.ubc.ca/Links/mrp/

#maths #math #research #postcards #mathart

Recent interactions with the editing & styling at a journal (the folks in charge of the physical layout of the text and figures) has me very confused about their incentives. I guess they do not (or at least should not) read for the actual mathematics. So what incentives are they working under?

It seems like they also aren't incentivized to make the final product look... good? Aesthetically pleasing? They returned a submission to us with wildly variable figure sizes, some of the math in the figures was too small to read, and others were triple the size of surrounding text. Figures referenced only in a single example would appear 4 pages before, or 3 pages after the reference.

Latex was supposed to communicate author intent; I feel like it got lost in translation.