Jason Hartline
@jasonhartline
Professor at Northwestern CS. Economics, by courtesy. Study mechanism design, economics of algorithms, regulation of algorithms, AI and society. My opinions. virtualchair.net, socialcal.org, pubchat.org,
If you're coming to Paul Dütting's plenary lecture, I'm sitting at a four-top on the left. Come join me.
This is what it looks like at one of the posters, a great paper by Cruz, Kleinberg, and Abebe.
The MIW program has 5 minute lightning talks followed by a poster session that is exceptionally good. Check it out. 👇
Great to be attending a virtual conference again, from the comfort of my office. If you do market design and you aren't at the Marketplace Innovations Workshop this week, you're not doing it right. marketplaceinnovation.net
In preparation for a talk on algorithmic collusion, I prepared the following diagram which explores the cross-product between independent v coordinated behavior with competitive v supra-competitive prices. Importantly, collusion is only the bottom right. The top right is also bad but not collusion.
I joined the #nokings protest in Chicago. With the war between the USA, Iran, and Israel and grave human rights violations by the regimes of all three, the good people of these countries must not allow the acts of their leaders to be normalized. March 28, 2026.
Hello @bsky.app friends, I would be happy to have some more people trying out #socialcal for turning Bluesky posts into calendar events. Details. 👇
Great presentation by UIC Prof. Michael Curry on "Language Models and Text Elicitation Mechanisms" We have an upcoming #IDEAL workshop on this topic on May 1. Pictured, Prof. Curry is overviewing our recent joint paper "Aligned Textual Scoring Rules". arxiv.org/abs/2507.06221
Great presentation by Northwestern postdoc Pooja Kulkarni on "Emerging Markets: Data, Models, and Tokens" What's different about data: non-rival, freely replicable, costly to share, and value a priori unknown. #IDEAL
I don't agree with "a computer can never be held accountable". But also, apparently the logic of this meme doesn't work, at least not according to the U.S. Supreme Court. Cf. Trump v. United States. en.wikipedia.org/wiki/Trump_v...
This is what Google thinks I'm looking for when I search for "Chernoff"?
Hello @chicagocta.bsky.social, for several days at least, the CTA train tracker is not showing the northbound purple line at Wilson. (These are usually above the red line in the attached screenshot.) www.transitchicago.com/traintracker...
Thanks to Kangning Wang for working with me on this "A Geometric Analysis of Gains from Trade". There is a very simple geometric proof that the "random proposer" mechanism gives a 4-approximation to first-best gains from trade. And some calculus gives the state-of-the-art 3.15-approximation.
Congratulations Dr. Paula Kayongo for successfully defending her Ph.D. dissertation! "Behavioral Information Design for Forecasting Dashboards: Equilibria, Mechanisms, and Calibration" Thanks to her stellar committee: @jessicahullman.bsky.social, @mjskay.com, and Annie Liang!
It was wonderful, but not easy, to try to keep up with the many projects that have gone into your formulation of Trustworthy AI, @yifanwu.bsky.social. Thanks for being patient and teaching me so much!
Congratulations to Dr. @yifanwu.bsky.social for successfully defending her Ph.D. dissertation! "Trustworthy AI: Foundations from Proper Scoring Rules" Thanks to her committee: @jessicahullman.bsky.social, Bobby Kleinberg, and Annie Liang.
Dashboards are visualizations of uncertainty. We compared: 1. allocation rules (probability of winning as function of bid). 2. (hypothetical) utility curves (for selected values, as function of bid). 3. heatmap (showing utility on a grid of bid,value pairs). Each are depicted in this figure.
The Serpentine at Chefs Special Cocktail Bar in Chicago is one of the best cocktails in the city. Got my grades in. Hello summer!
Once upon a time, five years ago, before BlueSky, I wrote the LaTeX code to to make a one-line Beamer block. Reposting. www.overleaf.com/read/whygfgw...
Here is my receipt for a Divvy scooter ride where I parked right next to a full station. Why is it costing me 50% more when a Divvy station is full? That's the wrong incentive for me and for Divvy (to keep the stations free).
Here is an essay titled "Frustrated with Oranges" by CharGPT. For the best experience, click the link and have ChatGPT read it out loud. chatgpt.com/share/67a8ef...
A bonus from visiting Morocco in January is surfing! (When the technical program allows.)
Had a great visit to @mcgt-um6p.bsky.social this January. Co-organizing a winter school on "Data Economics". Teaching half of my PhD "Mechanism Design" class to their first cohort of PhD students. Closed out the visit with this seminar on "Regulation of Algorithmic Collusion" (video below).
Last Friday at @mcgt-um6p.bsky.social's Winter School on Data Economics!!
Thursday at @mcgt-um6p.bsky.social's winter school on Data Economics!