Working on an article for SIAM News covering Courtney Paquette's plenary at AN25. Something for the precalc/calc students: horizontal asymp. here show you that more effort (flops) training a model yields no improvement.

Another data sci overtones one could sneak into the existing math curriculum.

you know what you have to do

Continuing an old joke,

I did win the teacher/scholar (and not the other way around) award this past year. Too bad it will be on pause because of the budget next year.

hell yeah cc: @chadtopaz.bsky.social

("...change the letter order to dodge the censors")

why the 1945-1960 bump

(not an endorsement)

I tried coloring by difference from median value, but doesn't show much extra.

(node sizes are log-scale in the number of stones with that number)

#AdventOfCode day 11. Turns out the number of unique stones is (surprisingly) small. So you can animate the flow of stones by state on a directed graph.

Day 8 of #AdventOfCode . Antinodes looking like a Jackson Pollock. @matplotlib.bsky.social

Day 4 #AdventOfCode :) (Since the pattern was *so hard* I did use regex for part one.)

(not doing it in R) but I felt if I used regex I'd take out some of the value for myself 😬 the end result was a small mountain of code for a low-tech solution.

Though honestly... I don't know how to handle a conditional with the "do()" and "don't()". Clever to strip out segments!

Did first day of #AdventOfCode in python... then thought, why not brush off the old fortran (plus the more modern fortran stdlib).

Love getting these in my firefox recommended www.youtube.com/watch?v=NJaS...

whoops meant to reply here - not sure if you were being facetious but here you go

THAT is certainly something.

My hot take is that Daylight Savings Time doesn't go far enough.

Clock time should be continuously adjusted relative to your latitude and time of year such that sunrise is always at 6AM.

Email apps can handle the conversions.

Don't ask me about extreme latitudes.

github.com/maminian/day...