Lightnews — Scholar-powered news

Jessica Hullman @jessicahullman.bsky.social · 11h

Thanks, that makes my day!

Reposted by Jessica Hullman

Gretchen McCulloch @gretchenmcc.bsky.social · 14h

Anyway I wrote an article back in 2013 analyzing permutations on Benedict Cumberbatch which you may read and sub in Timothée Chalamet should you so desire
web.archive.org/web/20250322...

A Linguist Explains the Rules of Summoning Benedict Cumberbatch -The Toast

A willing foe, and sea room.

web.archive.org

5 17 77

Jessica Hullman @jessicahullman.bsky.social · 20h

Some related papers I plan to mention:
arxiv.org/pdf/2502.02561
arxiv.org/abs/2306.00074
arxiv.org/abs/2401.08876
arxiv.org/abs/2503.11709

5

Jessica Hullman @jessicahullman.bsky.social · 20h

I'm giving keynote at a workshop on predictive uncertainty for decisions tomorrow & realizing I may have more questions than answers

What do ppl really do w/conformal prediction sets?
What should they do?
What to do when calibration isn't enough for trust?
What's shape of uncertainty in high dim?

1 7

Reposted by Jessica Hullman

Jessica Hullman @jessicahullman.bsky.social · 1d

For XAI it’s often thought explanations help (boundedly rational) user “unlock” info in features for some decision. But no one says this, they say vaguer things like “supporting trust”. We lay out some implicit assumptions that become clearer when you take a formal view here arxiv.org/abs/2506.22740

Explanations are a means to an end

Modern methods for explainable machine learning are designed to describe how models map inputs to outputs--without deep consideration of how these explanations will be used in practice. This paper arg...

arxiv.org

2 3 22

Jessica Hullman @jessicahullman.bsky.social · 1d

For XAI it’s often thought explanations help (boundedly rational) user “unlock” info in features for some decision. But no one says this, they say vaguer things like “supporting trust”. We lay out some implicit assumptions that become clearer when you take a formal view here arxiv.org/abs/2506.22740

Explanations are a means to an end

Modern methods for explainable machine learning are designed to describe how models map inputs to outputs--without deep consideration of how these explanations will be used in practice. This paper arg...

arxiv.org

2 3 22

Reposted by Jessica Hullman

Dan Goldstein @dggoldst.bsky.social · 2d

🚨 Postdoctoral Research Economist positions at Microsoft Research New England or New York City 🚨

www.aeaweb.org/joe/listing....

Open to PhDs in econ-related fields (e.g. quant marketing, econ-CS). Seeking those with high collaboration potential with the team: www.microsoft.com/en-us/resear...

American Economic Association: JOE Listings - August 1, 2025 - January 31, 2026

www.aeaweb.org

5 10

Jessica Hullman @jessicahullman.bsky.social · 2d

Looking forward to talking about our work on the value of explanation for decision-making at this workshop

Gunnar König @gunnark.bsky.social · 2d

Interested in provable guarantees and fundamental limitations of XAI? Join us at the "Theory of Explainable AI" workshop Dec 2 in Copenhagen! @ellis.eu @euripsconf.bsky.social

Speakers: @jessicahullman.bsky.social @doloresromerom.bsky.social @tpimentel.bsky.social

Call for Contributions: Oct 15

Theory of XAI Workshop

Explainable AI (XAI) is now deployed across a wide range of settings, including high-stakes domains in which misleading explanations can cause real harm. For example, explanations are required by law ...

sites.google.com

3 11

Reposted by Jessica Hullman

Heartland Signal @heartlandsignal.bsky.social · 3d

Chicago Mayor Brandon Johnson:

"Private establishments can partner with us to make it clear that private property is private property and ICE cannot enter establishments without a valid warrant ... If Congress will not check this administration, then Chicago will."

330 5.4K 19K

Jessica Hullman @jessicahullman.bsky.social · 3d

This is true even with slide reuse. Each audience is different, talk goals might be subtly different, and your own ability to explain the concepts may be different, even if you gave the talk recently.

I have a special kind of disdain for the attitude that one doesn't have to do their best on a talk

Nathan Lambert @natolambert.bsky.social · 3d

You should be spending 10+minutes on slides per minute of your talk.
Doing too many talks makes it so you don't have time for top quality.

11

Jessica Hullman @jessicahullman.bsky.social · 6d

one might hope this phenomena skips the academics but no such luck

3

Reposted by Jessica Hullman

Andrew Marder @andrewmarder.net · 6d

This spoke to me today: bsky.app/profile/shaw...

Shawn @shawnhils.bsky.social · 6d

1 1 5

Jessica Hullman @jessicahullman.bsky.social · 6d

The answer is more reasonable expectations, not ignorance.

2 5

Jessica Hullman @jessicahullman.bsky.social · 6d

No polls
No fundamentals forecasting
No RCTs
No inference

What’s up with anti-data extremism in the op-ed/blogosphere these days? It’s one thing to critique horse race polling or stat testing in place of thinking, but very different to conclude we’re better off not trying to model at all. 1/2

1 11

Reposted by Jessica Hullman

Peter Tennant @pwgtennant.bsky.social · 6d

"Uncooperative statistician": the term used (typically by a senior clinician) to describe a well-trained and knowledgeable statistician who refuses to conduct flawed or fraudulent research.

3 14 52

Jessica Hullman @jessicahullman.bsky.social · 7d

Also on substack

jessicahullman.substack.com/p/howl-after...

Howl, after Allen Ginsberg

for the AI-headed hipsters

jessicahullman.substack.com

1 1

Jessica Hullman @jessicahullman.bsky.social · 7d

statmodeling.stat.columbia.edu/2025/09/10/h...

Howl, after Allen Ginsberg (for the AI-headed hipsters) | Statistical Modeling, Causal Inference, and Social Science

statmodeling.stat.columbia.edu

1 1

Jessica Hullman @jessicahullman.bsky.social · 7d

My full rewrite of the (very long) first part of Howl is now live at the link below👇

I hope someone somewhere enjoys reading this as much as I enjoyed writing it!

Jessica Hullman @jessicahullman.bsky.social · 29d

I rewrote Allen Ginsberg's Howl, but for the AI-headed hipsters👇

"I saw the best minds of my generation distracted by machines, shining-eyed delusional demo gods, rubbering through polished cities at dusk Capitalizing The Next Big Push to the end of Introspection,

1/N

1 7

Jessica Hullman @jessicahullman.bsky.social · 12d

This is true, but is the devil ever not in the details? Even in non-methods empirical work, it arises in interpreting results unless we treat it all as exploratory existence proof

Only time its absent is when communities normalize ignoring it. Convention dictates so much of the knowledge we produce

John Sakaluk @johnsakaluk.bsky.social · 14d

This 🧵 has solidified for me a belief that there just are some scientific arguments that are harder to win than others, because the devil is in the belaboured boring-ass details, and those kinds of arguments are vulnerable to alternative positions that are easier (and more entertaining) to convey.

4

Reposted by Jessica Hullman

John Sakaluk @johnsakaluk.bsky.social · 14d

This 🧵 has solidified for me a belief that there just are some scientific arguments that are harder to win than others, because the devil is in the belaboured boring-ass details, and those kinds of arguments are vulnerable to alternative positions that are easier (and more entertaining) to convey.

1 4 13

Jessica Hullman @jessicahullman.bsky.social · 18d

Perhaps, but not rigorously. They wanted people to just take their word for it.

Jessica Hullman @jessicahullman.bsky.social · 21d

statmodeling.stat.columbia.edu/2025/09/18/m...

More Howl, after Allen Ginsberg for the AI-headed hipsters | Statistical Modeling, Causal Inference, and Social Science

statmodeling.stat.columbia.edu

3

Jessica Hullman @jessicahullman.bsky.social · 21d

who fed it words, who watched it excrete words, who fed it excreted words again & again waiting, who ran the primitive nervous system with sloshing words & prayed & made words of excrement & finally brought it out to check the Time, & now mistake the words for heroes

(more from my Howl rewrite)
1/2

1 1 6

Jessica Hullman @jessicahullman.bsky.social · 22d

The best time to work on something obscure is when all the men around you are caught up in posturing and arguing loudly.

7