I poured my soul into building this course last fall: 📚 Graph Algorithms via Graph Decomposition 📚 Graph decomposition has been a powerful framework in graph algorithms for over 20 years, but the literature is scattered and technical. Thus, I tried to organize part of it into one coherent story.

I really like this paper, one of my favorites in recent memory. Quick explainer of what's going on:

I am thinking of starting a cult that believes that fast max flow algorithms might someday be used to produce paperclips in a manner that will destroy humanity. It's unfair that the AI subfield has monopolized this.

Academia hack: when writing scathing rejections, also demand that the authors add citations to your enemies

Motion to bring medieval kerning back in our papers

stoc deadline tomorrow

Kevin Pratt, Yahel Uffenheimer, Omri Weinstein: (Approximate) Matrix Multiplication via Convolutions https://arxiv.org/abs/2510.22193 https://arxiv.org/pdf/2510.22193 https://arxiv.org/html/2510.22193

this is a pretty accurate summary of my research area

I always thought it seemed weird how fancy waiters in movies would be like "Excellent choice, sir" after someone orders off a fixed menu. Now ChatGPT does this, and I can confirm: it is indeed weird

Some UMich TCS lore: a (now-retired) professor once photoshopped these shocked Hilberts to express his surprise at increasingly complicated Hilbert's Hotel situations. They are now in use as all-purpose math reactions. Please enjoy

Some questions on spanners in my talk at the Simons Institute. Since the talk, progress has been made on a few questions, but most are open. minorfree.github.io/SpannerQues/

adding a "Do you like this personality 👍👎" box to my email signature

Thinking about how the security people give their results metal names like "spectre" and "meltdown" and design cool logos for them. We should give that treatment to our theorems

Incredibly well deserved!!

Hot CS take: Big-O notation should have been defined to hide constant factor changes in the input variable, not the output function value

Lean with it Coq with it

Huge congratulations to my amazing student Yeyuan Chen (+co-author Zihan Zhang of OSU advised by Zeyu Guo) for being awarded the STOC 2025 Best Student Paper Award! Their monumental result proves that explicit Reed-Solomon codes can correct more errors than previously known: arxiv.org/abs/2408.15925

I find that this works, mostly, but requires the right amount of skepticism. Too much leads me to the “dual” of “fully trustful” reading, bogged down in doubting every little piece of ink, and not seeing the shapes or overall picture.

when my family asks me about the impact of my research

frantically changing "graph theory" to "autonomous killer graph theory" in all my papers

This is terrible news, and part of the ongoing assault on science in the U.S. by the new administration. To honor Tracy Kimbrel and his service to the NSF's AF division, here's a short thread about a beautiful algorithm of his, joint with Rakesh Sinha (www.sciencedirect.com/science/arti...). 1/

A proof is a logical argument written to convince a skeptical audience. A corollary is that the best way to read a proof is to roleplay as a skeptical audience.

Same

Guy who updates his beliefs towards frequentism after witnessing examples of it working

One of my favourite things anyone has ever said about art. You'll be missed, David. 🌹

With @adamsmith.xyz and @thejonullman.bsky.social, we have compiled a set of profiles of 29 people in the "foundations of responsible computing" community ("mathematical research in computation and society writ large") who are on the faculty job market. Link: drive.google.com/file/d/1Hyvg... 1/3

In Vietnam they call these "dragon powder." That is so much cooler than "fire extinguisher" what are we doing

I might be the last horse to cross the finish line here, but I learned today that the length "1em" in tex doesn't stand for anything, rather, it's called that because it's exactly the length of one uppercase M 🤯

I was reminded this week of an open problem in graph theory that has absolutely no business being open. 🧵

Pepper illegally attempted to stop us from leaving this morning

A reminder about NY Theory Day in a week! Fri Dec 6th! Talks by Amir Abboud, Sanjeev Khanna, Rotem Oshman, and Ron Rothblum! At NYU Tandon! sites.google.com/view/nyctheo... Registration is free, but please register for building access. See you all there!

New post on additive spanners. Oustanding open problem: construct a +4 spanner with O(n^{4/3}) edges.

♪ when the function behaves like a sum of sine waves, that's a fourier ♪

New platform so I get to repost things. First of all, here is the greatest proof of all time