thank fucking god www.koco.com/article/ryan...

Tell me something you do when you code that other people would tell you that you shouldn't do. Tell me the rules you break! I'll go first: I work in untitled files in the wrong project directories all the time. Like, all the time. Yes, I do tend to lose things 😂 #databs #rstats #python

wonderful, emotional, shambles, etcetera

the next project that’ll consume my consciousness for a few months (was telling my wife about it in the kitchen and needed to get it written down ~somewhere~ before it leaked out of my brain)

do you think iger read elliott's article www.businessinsider.com/jimmy-kimmel...

Against my better instincts, I have written some notes on how human probability judgements work and what you should expect from surveys that ask people to guess what proportion of the population is transgender. I hope never to speak of this matter again

I would argue this is actually about Bayesian inference

Senate WAR/WARP estimates (with confidence intervals) now up on Strength In Numbers: www.gelliottmorris.com/p/the-senato...

instead of python 3.14 they shoulda called it pi-thon

It's very gratifying to write with a focus on statistical rigor and earnestly attempting to get the methodology right while being able to open-source the codebase. If you dig this sort of thing, you should subscribe to Elliott's substack --- multiple publications per week! www.gelliottmorris.com

MC Stan is here! Follow for the latest Stan news, and tag if you want us to repost your posts about new papers, packages, courses, etc. about Stan

The work we do takes resources. Whenever you see someone say, "The D outperformed by 22 pts," that data came from us. We can only survive if people know we exist. When influential writers pass our work off as their own, that's very damaging. You can stand up for the little guy by subscribing today

Really insightful post from Julie Tibshirani (spotted in LinkedIn, can't find on Bsky) reflecting on #rstats 's unique governance structure and what can be learned for other languages jtibs.substack.com/p/if-all-the...

80% of the reason I’m submitting an abstract to this is to have an excuse to visit Boston

Several groups, including @campaignlegal.org, @aclumo.bsky.social, and @aclu.org, have filed a lawsuit in state court challenging the Missouri GOP's new congressional map for violating the state constitution

rain_from_the_ground.mp3

isn’t actually over, currently in shambles

“I’d buy a Steinway” I say, unaware of the monkey paw curling

Never ask a man his age, a woman her salary, or GPT-5 whether a seahorse emoji exists

good work gmail (it is, in fact, neither a Monday nor a Father's Day)

very nervous about my psyche as I get close to the end of Clair obscur

I had to take leaves today so we’re moving from summer to fall earrings

A manor filled to the brim with just ridiculously lavish pianos

okay another point of comparison --- it's way easier to define an arbitrary number of independent gaussian process draws given the same covariance matrix in stan than it is in pymc (this is something niche but that I do relatively often)

this is a nice watch — interesting look into the language but also the stress that comes with doing public-facing work

the only thing I dislike prepping for in dnd is a shopping sesh (world's tiniest violin, I will survive)

Results so stunningly clear they inspired this classic xkcd (xkcd.com/2400/):

Ope there’s the third manning brother

had another show tonight! felt good to be back up on stage

Bayesianism is ZuTara while Frequentism is KatAang.

I have had an incredibly intense personal / professional schedule over the past month and I am very much looking forward to closing my laptop and just vibing with my plants in the backyard for the next two weeks

I’m especially proud of this article I wrote about Gaussian Processes for the Recast blog! 🥳 GPs are super interesting, but it’s not easy to wrap your head around them at first 🤔 This is a medium level (more intuition than math) introduction to GPs for time series. getrecast.com/gaussian-pro...

also not good for us-based students financially. I was able to afford the University of Tulsa while graduating with only federal loans because a _huge_ portion of my academic scholarship was subsidized by the ~25% international students who (generally) paid full price

some initial thoughts after plunking around in pymc: - wow, hilbert space GPs are really fast - models are a lot more compact than their stan equivalent - tensors are hell - holy shit, you end up spending so much time debugging backend problems last one might be a skill issue; still a cool package

found out some good news today ! wahoo !

every day I wake up, open blue sky dot com, and find another reason to move to the netherlands

testing out better functions for modeling "shocks" to random walk processes; each line is a simulated random walk, highlighted line is the draw with the largest single-step change

I finished Act 1 of Clair Obscur last night and I am still emotionally reeling

oh the joy of the double-whammy: getting sent a notebook in which the first line is os.chdir('path/that/only/works/on/original/users/machine/')

the people yearn for the mines (multi member districts with proportional representation)

good use of polling youtu.be/FxqrVoec20U?...

there's kind of a fun etymology associated with looking through other people's code. like seeing four spaces in a stan model means that the user is probably interfacing from python; or the fact that the recast docs reference "depvar" is probably an artifact from the founder's time using Robyn

whenever I feel adrift I think "more bookshelves in the house would fix me" and this is probably not addressing the root cause of what's going on but is also probably True

interactivity in html tables was the one thing keeping me tied to gt in some fashion — this is super cool ! (:

I will add: tiktokers pick up on this stuff BECAUSE it’s being reported first in places like @the-downballot.com or @boltsmag.org. That’s why people should support independent outlets that are paying attention to these local elections.

Words/phrases Dems should never use: • "approaching statistical significance" • "the proof is trivial and left as an exercise for the reader" • "convenience sample" • "instrumental variables approach" (but with intent to treat exception) • "well if anything the bias would be downward, wouldn't it?"

bought a crop top, feeling grand

@galendruke.bsky.social brings up specification uncertainty in the last 3 minutes of the pod — all the more reason to quantify how sure we are about stuff!

nobody is talking about uncertainty in these WAR models. that’s why i was happy to discuss the contributions of strength in numbers contributor @markjrieke.bsky.social in this debate