Sign in

Thom Volker

@thomvolker.bsky.social
382 followers 343 following 759 posts

PhD Candidate in Statistics, Utrecht University Creates fake data for a living. thomvolker.github.io

PostsRepliesMedia
Reposted by Thom Volker
Hagen Blix @hagenblix.bsky.social · 29/09/2026
The problem that we need to solve isn't rogue AI, it's that companies increasingly think they can just price crime into the business model and get away with it
350787
Reposted by Thom Volker
Kat Abughazaleh @katmabu.bsky.social · 29/09/2026
A man replied to this saying in earnest that women need to get organized and fight for our rights and men will follow. We did that. We do that. We mobilized the largest single day protest in global history on every continent and yet I have fewer rights than my mom did at my age.
15284401506
Reposted by Thom Volker
Michael Mishak @mjmishak.bsky.social · 26/09/2026
Online sportsbooks say they do not want business from problem gamblers. So @propublica.org's Jake Pearson spent 10 weeks on DraftKings testing how the company would react to someone displaying all the hallmarks of a compulsive bettor. Spoiler: It made him a VIP. www.propublica.org/article/draf...
propublica.org
I Deliberately Bet Like a Problem Gambler. DraftKings Made Me a VIP.
Amid growing public health concerns, operators like DraftKings have touted “responsible gaming” tools and prompts. But as I went off the deep end, I found they were no match for the incentives and pro...
17510167
Reposted by Thom Volker
Darren Dahly @statsepi.bsky.social · 02/10/2025
Every online getting company has exactly the data they need to identify people with gambling problems, and the technology and know-how to maximally enable that problem. They also know exactly how to spot the rare skilled gamblers and ban them from playing. Zero economic value. It's basically theft.
312532
Thom Volker @thomvolker.bsky.social · 25/09/2026
So, your slides already suck and then you make them even worse by introducing typos? Why would you do that??
010
Reposted by Thom Volker
Will Lowe @conjugateprior.org · 24/09/2026
The boring but effective way to ensure 'AI alignment' has two steps: 1. make those who build and sell AI responsible for what it does. 2. apply our best existing alignment technology – the law – to them. Maybe the AI is sentient, malign or benign. Who cares? It doesn't matter for this to work.
43310
Reposted by Thom Volker
Jonathan Bartlett @jonathan-bartlett.bsky.social · 09/09/2026
I am seeking to support an applicant for a PhD scholarship to pursue a PhD at LSHTM, starting October 2027, on developing machine learning methods for handling missing data in statistical analyses. For more details, and to express interest, please complete the form docs.google.com/forms/d/e/1F...
docs.google.com
LSHTM PhD - machine learning for missing data
I am seeking to support an applicant for a PhD scholarship from the UBEL DTP to pursue at PhD at LSHTM, starting October 2027. The topic is about developing machine learning methods for handling missi...
064
Reposted by Thom Volker
Rafe Meager (they/them) @economeager.bsky.social · 08/09/2026
this story is crazy. also, looks like academia's not so obsolete after all.
I said that if OpenAI released its result in the way proposed I would go
public with what happened. The reply was, “Why would you ruin your career?”
I replied that I am an academic, and asked why he thought going public would
ruin my career. The reply was, “If you don’t want me to be nice, then I don’t
have to be nice.”
623146
Thom Volker @thomvolker.bsky.social · 04/09/2026
It’s only fraud if you signed off that you won’t commit fraud
110
Reposted by Thom Volker
Sam Gershman @gershbrain.bsky.social · 01/09/2026
If this makes you mad, submit your cognitive science research to @openmindjournal.bsky.social No APCs, no subscription fees. It's funded by university libraries, at low cost.
18934
Thom Volker @thomvolker.bsky.social · 31/08/2026
Must be nice if your academic genre is "fiction"
000
Reposted by Thom Volker
Dr. Samantha Hancox-Li @sjshancoxli.liberalcurrents.com · 24/08/2026
i do think a lot of mainstreamers have missed the ball on how musk has moved past "rich guy with funny opinions on race science" and into "literally just a nazi"
674227902
Reposted by Thom Volker
Darren Dahly @statsepi.bsky.social · 21/08/2026
Just as we don't extend academic freedom to flat-earthers, we shouldn't extend it to "race realists". They consistently ignore or misunderstand the vast quantity of research across multiple fields that rubbishes their "viewpoints", and they produce shockingly bad "research" to support their views.
3197
Thom Volker @thomvolker.bsky.social · 18/08/2026
Surely the free speech advocates will do everything in their power to revert this policy
000
Reposted by Thom Volker
Remco Breuker 🟥 @remcobreuker.bsky.social · 17/08/2026
In het kader van het nieuwe academische jaar zet ik de toon nu alvast even: www.mareonline.nl/opinie/door-...
mareonline.nl
Door de lat steeds lager te leggen beledigen we onze studenten
Ik constateer al langere tijd, en ik ben niet de enige, dat de onderwijskwaliteit achteruit blijft gaan. We geven hier normaliter geld- en dus tijdgebrek de schuld van, maar ook de doorgeschoten betut...
67221
Reposted by Thom Volker
Ran (abu_hobbit) @abuhobbit.bsky.social · 17/08/2026
There are more people telling how they were hit by bikes than by SUVs
4726105
Thom Volker @thomvolker.bsky.social · 04/08/2026
So? As long as it’s smaller than .05, right??
010
Thom Volker @thomvolker.bsky.social · 03/08/2026
Nou, alleen nog wachten tot de VVD het weer laat klappen, dan mogen we weer
100
Reposted by Thom Volker
Hassnae @hassnae.bsky.social · 01/08/2026
Grote bek voor een land dat niet eens fatsoenlijke opvang in Ter Apel kan organiseren.
1261859516
Thom Volker @thomvolker.bsky.social · 01/08/2026
The PI presenting the paper written entirely by the rest of their research team
0213
Thom Volker @thomvolker.bsky.social · 30/07/2026
“As a result of today’s discussion, no UEFA national teams will participate in any FIFA competition for so long as these proposals remain alive” UEFA is taking no prisoners www.uefa.com/news-media/n...
uefa.com
Statement on behalf of UEFA and its 55 national associations
UEFA and its national associations will not participate in FIFA competitions.
000
Thom Volker @thomvolker.bsky.social · 30/07/2026
If you thought your smoke alarm's false positives are annoying, ... HT @statsepi.bsky.social
150
Thom Volker @thomvolker.bsky.social · 30/07/2026
After mocking business schools yesterday, I have to face the uncomfortable fact that researchers in my university also produce junk science. If you compare your anti-hangover drug to a placebo and you restrict the analysis to subjects who report a hangover when on the placebo, that's not good.
2111
Reposted by Thom Volker
★ govert schilling @govertschilling.nl · 29/07/2026
lees dit. @sheilas.bsky.social overtreft zichzelf. helaas, gaat het niks uithalen, vrees ik 😢
23602221
Reposted by Thom Volker
Colin Camerer @cfcamerer.bsky.social · 29/07/2026
A terrific ad
021
Thom Volker @thomvolker.bsky.social · 29/07/2026
Every time I read these anti-AI posts that the point of education is not to produce essays, I get a little disappointed because how else am I supposed to heat the stove
000
Thom Volker @thomvolker.bsky.social · 29/07/2026
This is what I imagine when people talk about projection predictive inference
120
Thom Volker @thomvolker.bsky.social · 29/07/2026
Business schools please never change!
040
Thom Volker @thomvolker.bsky.social · 27/07/2026
Not sure whether this is fellowship-worthy content
100
Reposted by Thom Volker
Jason Koebler @jasonkoebler.bsky.social · 15/07/2026
Scoop: The AI music generator Suno was hacked. Hacker shared source code that shows how the tool was made and part of the music and podcasts that were scraped to create it. Decades worth of music, lyrics, and podcasts from YouTube, Deezer, Genius & more www.404media.co/hack-reveals...
# youtube_nusic: ~l.5h runtine # 113,879 hours of youtube_rusic 12,287 hours of youtube_rusic_lyrics 16,843 hours of youtube_rusic_lyrics_for genius_hg: -1.3h runting 17,615 hours of genius_hq 41, 544 hours of genius_hq lyrics 17,387 hours of genius ha_lyrics_ foreign freesound: -B.2h runting 410 hours of freesound # insip: -B.2h runtine 19,514 hours of insip # jamendo: ~B.1h runtine 3,726 hours of janendo * ponds_music: ~1.1h runtine # 62,117 hours of ponds_music # deezer: -8. 5h runtine 12,287 hours of deezer 10,175 hours of deezer_lyrics 6,699 hours of deezer_lyrics_foreign * ytn_tagged: ~2h runtine # 152,162 hours of ytn_ tagged # musescore: -B.1h runtine 183 hours of nusescore_lyrics
4232681529
Reposted by Thom Volker
Andreas Kirsch @blackhc.bsky.social · 14/07/2026
I'm a Senior Research Scientist at Google DeepMind, writing in my personal capacity. More than 600 colleagues and I signed an open letter asking Sundar Pichai not to put our AI models on classified networks. The deal was signed anyway. I learned about it from the press.
335139
Reposted by Thom Volker
Dirk Eddelbuettel @eddelbuettel.com · 11/07/2026
Norm Matloff (1948-2026) @matloff.bsky.social I was made aware of the post below made by his daughter. We lost another friend and valued long-time contributor the #Rstats community, gone far too soon, and sure to be missed. His memory will be a blessing.
Screenshot of post on X in the account of Norm Matloff in which his daughter announces his passing. Gone too soon, and sure to be missed.
313038
Reposted by Thom Volker
escapedexplorer.bsky.social @escapedexplorer.bsky.social · 08/07/2026
DAWG WE'RE DOING STATE-RUN GAS PRICES AT PRICE CONTROLLED RATES DESPITE INCREASING SHORTAGES TO PACIFY THE PEOPLE AND THEY'RE SCREAMING ABOUT HOW THEIR OPPONENTS WANT COMMUNISM I AM GOING TO LOSE MY ENTIRE GODDAMN MIND
313149944239
Reposted by Thom Volker
Robin Ryder @robinryder.bsky.social · 08/07/2026
I have just resigned from the board of "Statistics and Computing", along with 17 other Associate Editors. This was a difficult decision: this great journal has published many tremendous articles under the leadership of EiC Ajay Jasra, and before him David Hand, Gilles Celeux, and Mark Girolami.
713453
Reposted by Thom Volker
Noah Greifer @noahgreifer.bsky.social · 29/06/2026
This made me very happy to see :) www.nytimes.com/interactive/...
0742
Reposted by Thom Volker
Luca Migo @victinibcn.bsky.social · 28/06/2026
Jared Kushner wants Albania's protected wetlands. The government obliged. Protesters said no. 29 days later, they're still saying no — with flamingo cutouts and flags, in the streets, every single night. Europe is watching. Is anyone else?
767261617378
Thom Volker @thomvolker.bsky.social · 17/06/2026
Ivy League press releases be like
000
Thom Volker @thomvolker.bsky.social · 16/06/2026
There is one thing I’d rather see than Jake Paul getting his ass whooped by Anthony Joshua
000
Reposted by Thom Volker
ContempraInn 🌹 @contemprainn.bsky.social · 13/06/2026
This is so Artistic 😊
253108241768
Reposted by Thom Volker
Jeremy Labrecque @jeremylabrecque.bsky.social · 12/06/2026
Interested in doing a PhD with me in epidemiology on causal inference and triangulation? In beautiful Rotterdam no less. www.werkenbijerasmusmc.nl/en/vacancy/1... Pro tip: if you get to the interview stage, there WILL be a question about ice hockey
werkenbijerasmusmc.nl
Vacature: PhD student in epidemiology to develop and apply causal triangulation
Many key questions in epidemiology are fundamentally causal. Yet, we often have to rely on assumptions that cannot be directly verified to judge whether an observed association truly reflects causalit...
22724
Reposted by Thom Volker
Hayoung Jung @hayoungjung.bsky.social · 11/06/2026
First paper of my PhD with my amazing advisors! There’s been a ton of hype and media coverage on OpenEvidence as an “AI co-pilot for clinicians”… and our long-horizon benchmark puts them to the test!! Our results suggest they are far from reliable for downstream use.
1144
Reposted by Thom Volker
Manoel Horta Ribeiro @manoelhortaribeiro.bsky.social · 11/06/2026
New preprint! We introduce a new benchmark, SciConBench, with 9.11k scientific questions derived from Cochrane Systematic Reviews. We find evidence that frontier AI agents **cannot** synthesize scientific conclusions well. A thread 🧵 w/ @hayoungjung.bsky.social & others!
24917
Reposted by Thom Volker
ArXiv Paperboy (Stat.ME+Econ.EM) @paperposterbot.bsky.social · 09/06/2026
arXiv📈🤖 Rank Intervals for Leaderboards: A Hierarchical Framework for Model Evaluation By Neuhof, Benjamini
Pretrained models are often evaluated on multi-task leaderboards to measure their applicability in diverse contexts. However, current methods for aggregating performance across tasks into leaderboard-level rankings do not address the uncertainty and variability at the task level. While recent works have proposed interval-based model rankings, the principled aggregation of uncertainty from individual tasks to leaderboard-level rankings remains unaddressed, and variation in models' performance across tasks is frequently obscured. In this work, we introduce a hierarchical framework that constructs model rank intervals with statistical guarantees at both levels: task-level rank confidence intervals from pairwise comparisons, and leaderboard-level rank prediction intervals using a conformal approach. This enables reliable quantification of model rank for each observed task and for new potential tasks. Experiments on simulated data and the TabArena and PromptEval (MMLU) benchmarks show that our method yields statistically valid and informative intervals, enabling reliable, uncertainty-aware model ranking on leaderboards.
031
Reposted by Thom Volker
Jake Grumbach @jakemgrumbach.bsky.social · 06/06/2026
Gelman absolutely on fire on the blog statmodeling.substack.com/p/13-aspects...
69521
Reposted by Thom Volker
Thom Volker @thomvolker.bsky.social · 05/06/2026
I knew that quarto and revealjs slides allow you to incorporate .html webpages directly in your slides, but I did not realize that this allows you to embed interactive games directly in your slides!
101
Thom Volker @thomvolker.bsky.social · 05/06/2026
I knew that quarto and revealjs slides allow you to incorporate .html webpages directly in your slides, but I did not realize that this allows you to embed interactive games directly in your slides!
101
Reposted by Thom Volker
Elena Candellone @elenacandellone.bsky.social · 03/06/2026
New preprint out! When people debate controversial topics online, two forces compete: personal convictions and peer influence. But how much does each one actually matter, and does it vary by topic?
1113
Reposted by Thom Volker
Thom Volker @thomvolker.bsky.social · 28/05/2026
Only less than two months before our Missing Data Summer School starts! If you want to learn how to appropriately handle missing data in your statistical analysis, this course might be for you. We cover cross-sectional, longitudinal and experimental data and both inferential and predictive models.
152
Thom Volker @thomvolker.bsky.social · 28/05/2026
Only less than two months before our Missing Data Summer School starts! If you want to learn how to appropriately handle missing data in your statistical analysis, this course might be for you. We cover cross-sectional, longitudinal and experimental data and both inferential and predictive models.
152
Reposted by Thom Volker
ArXiv Paperboy (Stat.ME+Econ.EM) @paperposterbot.bsky.social · 21/05/2026
arXiv📈🤖 Laplace Approximations for Mixed-Effects and Gaussian Process Quantile Regression By Nava, Sigrist
Laplace approximations are a standard tool for computationally efficient inference in latent Gaussian models, but they fail for quantile regression with the asymmetric Laplace likelihood because the observed Hessian vanishes almost everywhere. We show that this obstacle can be overcome without smoothing the likelihood: the relevant local curvature is given not by the observed Hessian, but by the Fisher information when the model is correctly specified and by the population curvature of the expected loss under misspecification. On this basis, we develop a Laplace approximation framework for quantile regression with mixed-effects and Gaussian process models. We propose practical curvature estimators, including the triangular kernel curvature (TKC) estimator, that yield approximations for posterior distributions and marginal likelihoods, and we establish their asymptotic validity. Empirically, the proposed methods are scalable and numerically stable, and for latent Gaussian models, they achieve accuracy comparable to or better than MCMC and variational competitors at substantially lower computational costs. More broadly, the framework clarifies how Laplace approximations can be justified for non-smooth generalized posteriors through local quadratic behavior of the expected loss.
021