Sign in

Jared Huling

@jaredhuling.bsky.social
507 followers 167 following 94 posts

Associate Professor of Biostatistics, University of Minnesota jaredhuling.org

PostsRepliesMedia
Jared Huling @jaredhuling.bsky.social · 27/09/2026
I hope I'm not the only one who thinks the AP stuff has gone way too far. No one actually believes most of this stuff is genuinely at a college level. I don't see how it actually helps students except pad credit hours
210
Jared Huling @jaredhuling.bsky.social · 23/09/2026
gooooood question
110
Reposted by Jared Huling
anil oza @aniloza.bsky.social · 21/09/2026
Congress was not a fan of the reports circulating this weekend that Vought is pushing for a commission to review all NIH grants www.statnews.com/2026/09/21/t...
statnews.com
Members of Congress push back on plan to create a commission of political appointees to review NIH grants
Several members of Congress voiced strong opposition about a reported plan to create a panel of political appointees to sign off on NIH grants
29046
Reposted by Jared Huling
Garrick Aden-Buie @grrrck.xyz · 21/09/2026
At conf I saw so many humans sharing ideas, showing each other how to write and use new tools, new packages, new #RStats and #Python code. We may not be writing or reading code in the same way, but we're still out there building tools for each other, trying to solve big and important problems.
2151
Jared Huling @jaredhuling.bsky.social · 18/09/2026
ooooh this is something I should implement in fastglm. I've already implemented it for general GLMs and it's much faster than brglm2 and more stable: cran.r-project.org/web/packages...
cran.r-project.org
Firth Bias-Reduced GLMs with ‘fastglm’
010
Jared Huling @jaredhuling.bsky.social · 17/09/2026
We simply know that the true probability of a given level isn't zero in most scenarios, so doing something like Firth (sort of like adding a particular prior) makes more sense than trying to plod along as if you didn't know a given level doesn't have 0 probability
010
Jared Huling @jaredhuling.bsky.social · 17/09/2026
Firth bias correction would be my first thought. It's a sort of analog to the Haldane-Anscombe correction (i.e. adding a half) to contingency tables to deal with zero cells and it *might* make sense here
110
Jared Huling @jaredhuling.bsky.social · 17/09/2026
"The University is not here to teach you how to make a shiny, well-polished report. It’s to teach you how to think, how to learn, how to communicate...if you start relying on GenAI to replace the work of learning, or use it as a crutch, ... you’ll stop doing the hard work of learning."
040
Jared Huling @jaredhuling.bsky.social · 11/09/2026
"...turn the tool against the primary goal. Indeed, the mass production at faster and faster pace of 'true/false' statements could destroy fertile ground instead of breathing life into new ideas."
041
Jared Huling @jaredhuling.bsky.social · 11/09/2026
"In recent months, the success of AI in solving major mathematical problems has made headlines even outside mathematical circles. But solving problems is only a tool and proxy for achieving the primary goal of conceptual understanding and insight. Forgetting this in the world of AI may..."
2105
Reposted by Jared Huling
Terence Tao @teorth.bsky.social · 11/09/2026
A group of 25 Fields Medalists, including myself, have made a joint declaration on Math and AI: mathandai.org . We welcome additional signatories. See also this article in the Economist announcing the declaration: www.economist.com/science-and-...
mathandai.org
Declaration — Math and AI
Read the declaration and add your name.
422052929
Reposted by Jared Huling
Molly Offer-Westort @mollyow.bsky.social · 26/08/2026
I participated in a reading group of Asymptotic Statistics this summer (thanks @bobbygulotty.bsky.social!). The text is very dense and there is not a lot of hand-holding, so I used an AI assistant to write up notes informed by our discussions mollyow.github.io/bloomsday/
mollyow.github.io
Reading Companion for Asymptotic Statistics
173
Jared Huling @jaredhuling.bsky.social · 08/09/2026
okay this is very good! please read this
000
Jared Huling @jaredhuling.bsky.social · 08/09/2026
I also put together (with the help of Claude) a 3D live replay of each of the key climbs that were featured in the 2025 and 2026 races (the routes differed so some climbs aren't compared): jaredhuling.org/power-files/...
jaredhuling.org
Gray Duck Grit climb replay
000
Jared Huling @jaredhuling.bsky.social · 08/09/2026
In short: the climbs. This year we raced the climbs at a much more aggressive pace but went easier in between where the riders knew they couldn't get rid of each other. This report goes in depth comparing data from the races: jaredhuling.org/power-files/...
100
Jared Huling @jaredhuling.bsky.social · 08/09/2026
I like bike racing and I like data. Each of the last 3 years I've raced the Gray Duck Grit gravel race, a favorite of mine. This year felt much harder than last year's edition, but the typical aggregate measures (normalized power/HR) were almost identical. What made this year's so much harder?
130
Jared Huling @jaredhuling.bsky.social · 07/09/2026
I'm very big on leaning into whatever I have the most energy for in the moment (within reason; some things just need to get done). I find overall I do much better and efficient work this way
131
Reposted by Jared Huling
John Mullahy @johnmullahy.bsky.social · 05/09/2026
Frog tied some string around the box. “There,” he said. “Now our manuscript is completed.” “But we can do lots more robustness checks,” said Toad. “That is true,” said Frog.
072
Jared Huling @jaredhuling.bsky.social · 26/08/2026
Can't imagine why you'd do it before then
130
Jared Huling @jaredhuling.bsky.social · 26/08/2026
I hate how I know exactly what this is and that it's accentuating my own procrastination for the exact same thing
140
Jared Huling @jaredhuling.bsky.social · 21/08/2026
maybe, but uploading raw AI generated output and passing it off as your own review so when that is explicitly against the code of conduct for a journal is bad and, in my opinion, dishonest at best. If journals want to use AI generated reviews they should be explicit about it
100
Jared Huling @jaredhuling.bsky.social · 20/08/2026
I'm now at the point where I'm not even sure peer review is doing anything at all. I don't think people in our field understand how prevalent copy-pasting AI output into peer reviews is. It genuinely makes me question what we're all even trying to do anymore. Our field needs to decide what it values
0111
Jared Huling @jaredhuling.bsky.social · 19/08/2026
entirely unsurprising, but still interesting to see quantitatively
020
Jared Huling @jaredhuling.bsky.social · 19/08/2026
yes, plainly
010
Jared Huling @jaredhuling.bsky.social · 17/08/2026
I will say, there are far more stupidly dangerously designed highway interchanges in the twin cities than in, say, Columbus, OH where I grew up. Why we have cloverleaf interchanges (and worse) in the year 2026 is beyond me
110
Jared Huling @jaredhuling.bsky.social · 17/08/2026
How nice that would be. I love Minneapolis but it has some of the absolute worst drivers in the country because there is essentially zero traffic law enforcement anywhere. I've seen people run through red lights in the city while 15mph over the limit in front of police and they do nothing
100
Jared Huling @jaredhuling.bsky.social · 14/08/2026
instead we could be talking about how great transit is getting in Minneapolis www.youtube.com/watch?v=J1qN...
youtube.com
Is Minneapolis Too Nice for Its Own Good?
YouTube video by Ray Delahanty | CityNerd
110
Reposted by Jared Huling
Alex Schieferdecker @alexschief.bsky.social · 13/08/2026
I appreciate this video, not just because it's complimentary towards the work that we're doing in Minneapolis, but because it reflects the same sense that I have about our *momentum* towards building a better city. I think we've got something special here and I'm glad people are noticing.
19118
Reposted by Jared Huling
Jessica Hullman @jessicahullman.bsky.social · 11/08/2026
Nice post from @eytan.adar.prof on the cargo cult science of "AI-native" PhD students, who come in super productive because they perform all major parts of research with genAI. No one needs all these quick papers, and they can really hurt students down the road. eytanadar.medium.com/ai-native-ph...
eytanadar.medium.com
AI-Native PhD Students
If you’re a student and just want the tl;dr, feel free to jump to the end.
24312
Jared Huling @jaredhuling.bsky.social · 10/08/2026
Bleak!
000
Jared Huling @jaredhuling.bsky.social · 06/08/2026
cycling gives you more freedom to control your own schedule. it's also just so much more enjoyable than any alternative mode of transportation. and the more, the merrier!
000
Reposted by Jared Huling
Erin J. Wamsley 🧠📈😴 @wamsleylab.bsky.social · 06/08/2026
🚨🚨🚨 ALERT: Vanderbilt's board just launched a horrifying assault on academic freedom that threatens the legitimacy of their university. 1/x #AAUP #academicsky #edusky
insidehighered.com
Vanderbilt’s Board Declares War Against Academic Freedom
A new policy punishes faculty for speech the trustees deem strays from the core purpose of the university.
13199163
Jared Huling @jaredhuling.bsky.social · 02/08/2026
I'm not challenging how impressive it is! Just that we aren't yet 100% sure what these 10 proofs contain.
110
Jared Huling @jaredhuling.bsky.social · 02/08/2026
Definitely impressive, but worth noting that at least one of these 10 proofs is incorrect: philpapers.org/rec/NIEWTC
philpapers.org
100
Jared Huling @jaredhuling.bsky.social · 29/07/2026
it is definitely over-engineered and isn't a natural idea for that intersection. That people don't understand how to use it 'correctly' isn't user error imo
141
Jared Huling @jaredhuling.bsky.social · 27/07/2026
I started about 3 years ago. last two years I biked 10k+ miles and have really gotten into gravel racing mostly
010
Jared Huling @jaredhuling.bsky.social · 27/07/2026
'What?' indeed!
100
Jared Huling @jaredhuling.bsky.social · 26/07/2026
thanks!!
010
Jared Huling @jaredhuling.bsky.social · 26/07/2026
Today I won the Minnesota Cat 4 State Cycling Road Race championship. A hot and hard day -- over 105F at points according to my Garmin device
2132
Jared Huling @jaredhuling.bsky.social · 10/07/2026
I'll never forget coming across this part of his book, it really left an impression on me. Teaching is not only such an important part of what we do as academics, but also truly does have benefits for our own work!
030
Jared Huling @jaredhuling.bsky.social · 02/07/2026
I know this is a joke, but "energy balancing" comes from the energy distance (en.wikipedia.org/wiki/Energy_...) which measures the distance between two joint distributions
000
Reposted by Jared Huling
Alan Grossfield @agrossfield.bsky.social · 30/06/2026
Another important thread from @jeremymberg.bsky.social , this time on data showing nih programs to improve minority access to research careers worked.
0146
Reposted by Jared Huling
arxiv.stat.ME @arxiv-stat-me.bsky.social · 29/06/2026
Martha Barnard, Jared D. Huling, Julian Wolfson Causal Inference for Functional Treatments with Stochastic Policies arxiv.org/abs/2606.27518
021
Reposted by Jared Huling
Noah Greifer @noahgreifer.bsky.social · 29/06/2026
@natecohn.bsky.social (the author of this post) reached out to me for some help with it. Energy balancing is a weighting method (similar to entropy balancing) that balances the entire covariate distribution, not just the means. It's described here by @jaredhuling.bsky.social:
degruyter.com
141
Jared Huling @jaredhuling.bsky.social · 29/06/2026
I find this fascinating because my reaction to Dutch suburbs was how wonderful, sane, and livable they are in comparison to suburbs in the US
010
Reposted by Jared Huling
Eric Topol @erictopol.bsky.social · 17/06/2026
Academic freedom assessment across the world. The US is on a steep decline, below >120 countries www.scientificamerican.com/article/see-...
11400218
Jared Huling @jaredhuling.bsky.social · 08/06/2026
With the latest update on CRAN, I've made additional changes to fastglm, including Firth bias correction for all standard GLMs. If you need to fit negative binomial models, try fastglm, as it's much more likely to converge than MASS::glm.nb(), see cran.r-project.org/web/packages...
cran.r-project.org
Negative Binomial Stability Benchmark: fastglm vs MASS
022
Reposted by Jared Huling
Nicholas A. Christakis @nachristakis.bsky.social · 06/06/2026
“You were respectfully informed that distribution of materials [critical of the Trump administration] was not permitted and given the opportunity to remain in the meeting if you stopped handing out the materials. When you continued the behavior, we had no choice but to remove you from the meeting.”
nytimes.com
Police Remove Diabetes Experts From Conference for Distributing Critique of Trump Administration
2236
Jared Huling @jaredhuling.bsky.social · 27/05/2026
This is huge. You could have perfectly good cholesterol, do all the right things in terms of diet, exercise and sleep and yet still be at major risk of CVD due to high LP(a). Lots of promising therapies and we'll get to hear whether one of them works this year
020