it’s a great day for transphobes to eat shit

email clients need to start supporting LaTeX send tweet

over time + agnostic to the response scale --- I'll zap out some LaTeX and send your way when I get back to the house

have you thought about modeling approval using an ordered logistic likelihood? would need separate likelihood statements per response (eg, approve/disapprove vs strongly approve/approve/disapprove/strongly disapprove), but that would let you compare strength of approval/disapproval in a generic way

pymc is a really great ppl --- it's super pythonic, so you can do things like have your model code literally be a part of a class and use standard python dependency tooling. The one downside imo is that you can't save a pre-compiled model as a separate exe for later use w/o compilation

AHAHAHAHAHA

*wheezing. Deep breaths*

AHAHAHAHAHA

three very good things today
- electricians fixed some faulty wiring in the basement near a gas line so my house is no longer at risk of blowing up
- got two small PRs merged and opened up a big important one
- my wife & I planted a new garden bed after workbefore it started raining

How to Estimate a Correlation, and What It Means for Science | Computational Psychology Introduction In Part 1 of this series, we showed that six different methods for estimating individual means—James-Stein, classical true score estimation, empirical Bayes, ridge regression, and hierarc...

his is such a good write up (and also validates the "just be bayesian" approach [you get all the shrinkage and extensions to differing number of observations per subject for free!])
haines-lab.com/post/how-to-...

I now have a primitive interactive browser interface to the workflow graph. I can get this to support an interactive design mode I think.

I just want to say, and let me be very clear so that no one misunderstands: javascript f'ing sucks.

working on a project w/one of my siblings that includes:

Lord of the Rings -> Impossible Odds
Lord of the Rings -> GAY

fit go crazy

cover of the book "Bayesian Workflow" by Gelman, Vehtari, et al. Coming out later this year, in the summer probably.

I would have preferred to have the "draw the rest of the owl" meme on the cover, but this will do. Seems like it is on schedule, and we'll leave some typos so you know we didn't write it with AI.

one of the all time gamechanger great eps^^^

"Ify Nwadiwe is the best Um, Actually contestant because he has the most wins"

this phrase sent me down a spiraling path of math that I'll yawp about tomorrow

pretty sure we'd know if the president was dead from the insider trading on polymarket

a very good day today
- got a lot of yard work done
- lil afternoon get together with friends
- back porch coding while listening to Zach Bryan play at TU

just one of those glass-half-full days

if you understand this please go outside

silly goose mode is maximally engaged

they interrupted survivor for this

BIG THINGS a-goin' on in like 15min

they are calling it the scotus cuck chair

I AM HOLDING A LOVELY TRANS FLAG AND DECLARING THUS: HOORAY FOR MY NICE TRANS FRIENDS!!

people talking about statistics and getting everything wrong is a free perk of the 9-5

screaming, crying, throwing up

(need to convince people to stop using an average of averages)

in a fever dream