"if you want to be open-minded, thereβs no good reason to start with work promoting a politically loaded, much-debunked idea and published on a propaganda site."
--Andrew Gelman
@danpsimpson.bsky.social
Lover, artist, statistician.
"if you want to be open-minded, thereβs no good reason to start with work promoting a politically loaded, much-debunked idea and published on a propaganda site."
--Andrew Gelman
Weird. Youβre a much better writer than
04.08.2025 11:22 β π 1 π 0 π¬ 0 π 0I regret to inform you that she has, yet again, written more words than are necessary on a topic you don't care about
03.08.2025 02:02 β π 46 π 13 π¬ 2 π 1Naked Gun is extremely stupid and Iβm very happy about that.
02.08.2025 21:59 β π 3 π 0 π¬ 0 π 0WBN releasing this and making me have these feelings in the middle of my work day? Rude.
01.08.2025 18:53 β π 2 π 0 π¬ 0 π 0I donβt know why but while Wes is just not an English name, Streeting is so English he might as well be named Britain McBrittface
31.07.2025 03:20 β π 0 π 0 π¬ 0 π 0If you liked this thread, and you're like, "I bet that guy would make an awesome coworker," I'm currently on the job market - primarily looking for data scientist/quantitative researcher roles, but I'm flexible!
30.07.2025 21:05 β π 39 π 10 π¬ 2 π 1:(
31.07.2025 01:10 β π 3423 π 442 π¬ 22 π 4*Screams in @worldsbeyondnumber.bsky.social*
30.07.2025 03:30 β π 0 π 0 π¬ 0 π 0I just read diff-on-diff as dilf-on-dilf so thatβs how Monday is going
28.07.2025 14:13 β π 2 π 0 π¬ 0 π 0Let it be said that the insistence that universities must produce useful things is what is killing (or has killed) universities. The reluctance to value useful things to the same extent as other things is what allowed this to happen.
28.07.2025 06:59 β π 7 π 0 π¬ 0 π 0lol. Academias doesnβt value it at all. Useful is (generally) neither prestigious nor quantifiable. There are people in academia who value it and will help. There are even some rare places that will do it. But being useful (broadly or narrowly) really isnβt the goal of most academic systems.
28.07.2025 06:18 β π 11 π 0 π¬ 1 π 0New Yorker Instagram post of "Sniffies brings cruising into the digital age" by Emily Witt, with New Yorker style cartoons of naked men on their phones
hot take: cis women should not be writing about Sniffies for mainstream publications, and you should not be reading those pieces.
25.07.2025 16:31 β π 428 π 73 π¬ 16 π 22DO YEAR newyorkermag "On Grindr, the third thing you send is a picture of that your genitalia; on Sniffies, it's your profile picture." a Sniffies user THE NEW YORKER 2/5 :
newyorkermag "I can wake up on a Saturday and be, like, O.K., I don't have a lot to do today, I should clean my house, but instead I'm just going to go out and have sex with someone." a physician in L.A. .. 3/5 THE NEW YORKER
newyorkermag "I think it's understood that everyone there is probably there to fuck and not necessarily, like, 'Would you like to go to the theatre with me?"" a Sniffies user 4/5 THE NEW YORKER
In case anyone wants to try and pretend that @embits.bsky.social's New Yorker article on Sniffies is anything other than pure voyeurism, here are thre three pull quotes @newyorker.com is using to promote it.
They know what they're doing, and they're selling it hard.
While I was on vacation enjoying Finnish summer, Frank Weber finished some PRs and made a new major CRAN release of projpred mc-stan.org/projpred/ (for projection predictive variable selection). This release contains some major features add by me and Frank.
1/4
Needs another S
22.07.2025 19:28 β π 1 π 0 π¬ 0 π 0AIC comparing discrete to continuous (or mixed discrete/continuous to continous) models? Run away!
notstatschat.rbind.io/2025/07/22/a...
Same reason we shouldn't use LLMs to do arithmetic (although I've seen people with PhDs try). We already have calculators. They're reliable, fast, and widely available.
22.07.2025 04:02 β π 5 π 1 π¬ 1 π 0This will get taken up as a problem with the model (it gave a bad answer), when it really should be taken up as a problem with the use (people shouldn't turn to LLMs for this kind of question, though I'm sure some do).
22.07.2025 03:59 β π 6 π 6 π¬ 1 π 1Thatβs literally my strategy and I hate it.
22.07.2025 05:54 β π 2 π 1 π¬ 0 π 0I would really like them to add an accessibility setting thatβs basically just βI donβt want to fucking do thisβ
22.07.2025 01:44 β π 2 π 0 π¬ 0 π 0Elder scrolls oblivion is fun but the lock picking minigame can truly fuck right off. Atrocious.
22.07.2025 01:35 β π 3 π 0 π¬ 2 π 0You can have some of ours.
21.07.2025 20:29 β π 2 π 0 π¬ 1 π 0Is the appendix a grave?
21.07.2025 19:45 β π 2 π 1 π¬ 0 π 0In a stunning moment of carnal desire, my calculator said βBOOBIE5β
21.07.2025 08:44 β π 4953 π 1103 π¬ 16 π 6Image of spell check suggesting "reprehensibility" as a replacement for "reproducibility"
Spell check making some very strong claims this morning
21.07.2025 19:00 β π 17 π 1 π¬ 1 π 0More like Babylon Cis-ters, amirite
21.07.2025 18:32 β π 1 π 0 π¬ 0 π 0Like Iβm sorry if anyone here is impressed by leaderboards. Fucking grow up. To the extent that there are actually good metrics, I promise you they are very intentionally not public.
21.07.2025 18:12 β π 3 π 0 π¬ 0 π 0Like I donβt post about work on here because, like, that is not my journey anymore but this is definitely true. Both for regulatory oversight, but also like just for making the damn things work reliably in a production environment. Your metrics will decrease but they were bullshit anyway.
21.07.2025 18:10 β π 3 π 0 π¬ 1 π 0