Simon Wood, the GOAT of generalized additive models & creator of the mgcv #rstats package, has an Annual Review of Statistics essay on GAMs, available open access #statssky #mlsky

www.annualreviews.org/content/jour...

Thought I'd say a little more about some probprog-related stuff I'm building in Rust...
🧵

A friend of mine, I recently found out, does really fantastic memes about economics and econometrics. The initial ones are in Russian, but I convinced her to translate them into english so I could share them with a broader audience, I think many of you will appreciate them!

keen to read this one!

arxiv.org/abs/2504.12190
'Creating non-reversible rejection-free samplers by rebalancing skew-balanced Markov jump processes'
- Erik Jansson, Moritz Schauer, Ruben Seyer, Akash Sharma

a few days ago I tried to get chat ghibli to write some boilerplate roxygen docstrings so that I could finally defeat the linter, and instead it repeatedly just tried to delete random other parts of the file.

im not saying this stuff is useless, but its clearly not “generally” intelligent …

woah @cavesofqud.com merch unexpectedly fire, sad to see they sold out of my size

a familiar point nicely put, by Phil Dawid

Turing v cool. I love Stan, but I think it sometimes suffers not being embedded in a larger language.

A while back @gdalle.bsky.social made an HMM library, and it was stupidly easy to just toss his code into a Turing model to get efficient HMMs in Turing for free.

gist.github.com/JasonPekos/8...

me at work, macrodata refining (picking gp hyperparameters)

“There are uncountably many reals” factoid actually just a statistical error. Diagonal Georg, who lives in paradise & instantiates over 10,000 numbers a day, is an outlier who cannot be counted.