Aligning an AI with human preferences might be hard. But there is more than one AI out there, and users can choose which to use. Can we get the benefits of a fully aligned AI without solving the alignment problem? In a new paper we study a setting in which the answer is yes.

Thing with Grok is that better versions will be less overt and more convincing. An expert on arguing about the deficit, immigration, public health, etc. with a particular political slant

As AI systems continue to become more complex and hard to reason about, one increasingly powerful lens through which to understand them is through the incentives of AI creators.

Ecstatic and deeply honored by this award. I've had great fun thinking about algorithms as strategies for repeated games over the past few years and hope that this highlight will push more researchers to come up with exciting directions in this field! Come to our talk on Monday to learn more!

This best paper news is a good opportunity to highlight that a month or so ago I started maintaining CV of failures on my website. It will almost certainly continue to grow linearly in the number of things I attempt to do, and that’s a good thing! www.seas.upenn.edu/~ncollina/Fa...

Delighted (and honestly a little bit stunned) that our paper “Swap Regret and Correlated Equilibria Beyond Normal-Form Games” was just awarded both the “Best Paper” and “Best Student Paper” at EC! arxiv.org/abs/2502.20229

Harriet Tubman #quilt by my Mom, Vera P. Hall, who makes quilts celebrating Black people who fought for their own freedom. This seems to be the crowd favorite of the “We Didn’t Wait for Freedom” series. Happy #Juneteenth #quilting

NEW: Democracy is alive in Philly! Amazing scenes in Philadelphia as thousands hit the streets for the anti-Trump protest. #NoKings #50501Movement

I hugely recommend Harry as a researcher, mentor and person. If you are interested in the stuff he does, apply to do a PhD with him!

Check out the terrific set of EC 2025 accepted papers! ec25.sigecom.org/program/acce...

The idea that you can just “teach computer science” and be apolitical is a beautiful dream that expired in the 2000s, at the latest. Computer science has re-organized every facet of our society: it is inherently political. Instead of taking this idea seriously, we ran from it. Now we live in hell.

PhD Timeline xkcd.com/3081

I'm an AI researcher. The proposed devastating cuts to US science are supposed to spare AI. That's like a football team saying we're going to slash spending on everything but the quarterback. That quarterback will not be successful (or will go play for another team). www.nature.com/articles/d41...

So proud of my sibling, who is doing amazing work in the Philly community. A reminder to all of us who are overwhelmed with the state of the world right now—sometimes positive impact starts with the community directly around you. www.linkedin.com/posts/temple...

🚨 WaPo exclusive — The Trump admin has set its sights on deporting **1 MILLION** immigrants this year. How it went from a quiet ambition to a White House strategy. Scooping with @mariasacchetti.bsky.social.

I am a professor at Columbia University. All of the student NIH training grants have been canceled and now there are reports that ALL funding will be frozen. Why are Universities not banding together and speaking out publicly and forcibly about governmental attacks on biomedical research?

Kseniia Petrova, a Harvard scientist who fled Russia after protesting its invasion of Ukraine, has been captive to the hard-line immigration policies of the Trump administration for nearly eight weeks. She fears arrest if she is deported.

Check out our new paper on collaborative prediction!

I’m in Munich this week for CSLAW, presenting our work on the difficulty of defining and regulating algorithmic collusion! arxiv.org/abs/2409.03956 If you’re in the area and want to chat about coordination between algorithms (or anything else related to the game theory/ML intersection) let me know!

Whoa, was half-listening to a @thedailyshow.com segment with @linamkhan.bsky.social and suddenly heard talk of Algorithmic Collusion! (11 min in) Way to go, people researching this! @jasonhartline.bsky.social @aaroth.bsky.social @jubaz.bsky.social @ncollina.bsky.social @yannaigonch.bsky.social

Excited about our new work, asking what should count as “swap regret” in Bayesian games and beyond! We think we’ve pinned down a really nice answer, as @epsilonrational.bsky.social describes in this thread:

Good rule of thumb in life: David Blackwell already figured this out

I wish the faculty/students were given a chance to decide on this. Do we want to stand up for our students and first amendment rights or care more about money at the expense of what someone phrased as "only a minority concern".

And speaking of good news, this paper was accepted to STOC’25! It’s my first STOC/FOCS paper, which, as someone who likes to think of herself as a theorist, is an exciting milestone!

Shout out to @aaroth.bsky.social for helping make this possible! He has been a fantastic advisor, giving me the freedom to do whatever research I want while also providing guidance towards very cool lines of thought.

Go, and I say this with the utmost sincerity, Birds

seen in Philadelphia: FUCK ICE FUCK TRUMP + MUSK GO BIRDS

NASA did a feature on me last year. It was up this weekend but now when you click on the link you get a 404 message.

NEW: 2024 has just been confirmed as the warmest year on record, and the first to breach the 1.5C threshold. We used a ridgeline (Joy Division inspired) chart to visualise daily temperature anomalies since 1940. 2024 clearly stands out with 100% of its days above 1.3C and 75% above 1.5C.

Today I got some good news—I’ve been awarded the 2024-2025 IBM PhD Fellowship! A really wonderful way to close out 2024, a year in which I poured a lot of time, energy and excitement into my research.

William is doing a donation drive for these charities: between him, his employer, and me, we’re matching donations up to $2000 four-fold to $10000. Send him receipts!

New paper with @epsilonrational.bsky.social and Jon Schneider! Say you’re playing a repeated game against an opponent who will best-respond to your algorithm, but you only have a prior over their utility. What algorithm should you deploy to maximize your expected utility? arxiv.org/abs/2412.182...

I’m a monthly donor 💞 Wikipedia is the one nice thing we have

Family holiday “ice hockey” on a frozen river: no major bodily injuries!

Stay kiki

Giving a talk on this at the Junior Theorists Workshop tomorrow (12/6) at TTIC! If you’re in the area come through, I have nice pictures and promise to make at least one (1) joke theory.cs.northwestern.edu/junior-theor...

A starter pack of researchers in AI and ML at Penn, currently woefully incomplete. Please PM to be added or suggest names. go.bsky.app/5SxWsoZ

Check out our new paper, where we use calibration to substantially generalize results in the Agreement literature, including extending to the setting of human-AI interactions! I think it’s really cool, and if you don’t agree, we can engage in a short protocol to fix that with high probability :)

you don't know me. I don't have a driver's licence. and maybe I want a goat as a pet. I'm not switching my door, monty

“Ten Dual Commandments,” a Hamilton parody that helps you remember how to take the dual of an LP

Simplicity is underrated! Check out our 3-page paper in SODA 2025, solving an open problem from COLT 2024.

Everyone follow Eshwar @epsilonrational.bsky.social, a fantastic researcher and person! We’ve worked together on characterizing optimal learning algorithms for repeated games (arxiv.org/abs/2402.09549, arxiv.org/abs/2207.04192) and implications for algorithmic collusion (arxiv.org/abs/2409.03956).

The starter pack of all starter packs that don’t contain themselves

bird sounds

This year’s SIGecom job market candidate profiles are now available. Please share the following link with anyone you think may be interested in hiring from our community in either postdoctoral or faculty positions: www.cs.drexel.edu/~vg399/SIGec...

I made a theory starter pack. It includes some CS/Econ, pol-econ, and behavioral types as well... Lmk if you'd like to be added (or removed!) and I'll do my best - the edit functionality is a bit wonky atm go.bsky.app/KAGmZH4

Excited about this recent work on algorithmic collusion, which will be at ITCS’25!