A young aspiring mathematicians asks about the future of math with all the AI advances. My advice: Embrace research but be ready to pivot as needed. We many be seeing a new age in mathematics, how could you not be part of it.

Mathematicians find largest prime number to date www.der-postillon.co... HT: Marco Lübbecke

Bill on the future historians access to our times. I'm more interested in how they will judge us, assuming there is a future with historians.

The new SIGACT Luca Trevisan Award for Expository Work promotes and recognizes high-impact work expositing ideas and results from the theory of computation. Nomination deadline is April 10th. sigact.org/prizes/tr...

Google put out an article on accelerating research with Gemini. I'm one of the 34 authors for my adventures in vibe-coding a research paper.

One of the humanities fellows at Oxford called a "computer science library" an oxymoron. Yet, the university has one and a visit helped bring new life to an old paper.

Thomas Watson has a new computational complexity textbook about to be published by Cambridge University Press. There's a free version online for personal use. complexityincs.com

Bill talks about how technology leads to fears of less learning throughout history.

Large-language models have gotten very good at academic literature search, but it often requires two prompts where the second prompt is "You made up those references. Try again."

Revisiting an old friend by visiting it for the first time. blog.computationalco...

I'm perfectly happy having AI models trained on all my research papers and everything else I've publicly written, and I hope others feel the same. Best to have the models know as much about computational complexity as possible if we want to best further the field.

England weather has been as advertised, cold, dark and damp. I haven't seen dry ground since I got here two weeks ago. Still we're 50 degrees warmer than Chicago today.

The 2026 Michael and Sheila Held Prize goes to Irit Dinur, Subhash Khot, Guy Kindler, Dor Minzer and Muli Safra for their work on the 2-to-2 Games Theorem.

Congrats to Irit Dveer Dinur of the Institute for Advanced Study & Weizmann Insitute, #NASmember Subhash Khot of New York University, Guy Kindler of @hebrewuniversity.bsky.social, Dor Minzer of @mit.edu, and Muli Safra of Tel Aviv University, winners of the 2026 Michael and Sheila Held Prize! (1/2)

In academia we have two main communities, our department and our academic field. At Oxford, community seems to lie mainly in the colleges.

New ACM Fellows include Swarat Chaudhuri, Javier Esparza, Paolo Ferragina, Ken-Ichi Kawarabayashi, Aggelos Kiayias, Alistair Moffat, Noam Nisan, Ariel Procaccia, Oded Regev, Natarajan Shankar, Stephanie Weirich, Adam Wierman and Ke Yi. awards.acm.org/fello...

Bill asks how can we handle students using AI for homework and I give my thoughts. Short Answer: Not much.

This is scientists’s version of tearing down parts of the White House.

A three-decade-old open problem of mine just got solved! Matt Kovacs-Deak, Daochen Wang and Rain Zimin Yang showed that if a Boolean function f is exactly computed by a low-degree rational function, the quotient of two polynomials, then f has low decision-tree complexity. 1/2

RIP Scott Adams www.nytimes.com/2026... Whatever you think of the man, his Dilbert creations captured the tech world. Just call me Dan. 1/2

In the LLM era, I'm finding people are not very tolerent of typos and grammatical issues.

Bill's post smells like an egg frying on the sidewalk on a winter's day.

You rarely see mathematicians claiming that AI systems will solve any of the Clay Math Millennium problems in the near future. It's not that AI is improving its math capabilities quickly. More so that the Millennium math problems are far more difficult that any of us can imagine.

I made it to Oxford for my visiting fellowship at Magdalen College. Time to return to Computational Depth, measuring information with randomness removed.

Good budget news for the National Science Foundation. Well, good compared to what might have been.

From Slop to Verification to Collaboration to Generation, AI is changing how we write research papers and its getting better quickly.

And God Created Woman, then took her away at the end of 2025, making Brigitte Bardot a recipient for Bill's Betty White Award.

A really important piece. The times when we have opened our country to talented people from around the world are among the moments when America has been truly great. Trump is destroying that legacy and fomenting unconscionable racism. www.nytimes.com/2025/12/29/o...

ACM digital library is now open access. https://dl.acm.org/ They have moved to an author/institution pay model, and features like stats and advanced search now require a "premium" subscription. I do my searching in Gemini now, and Google Scholar has good stats.

TIL that now, in Australia's summer, Brisbane is a half hour behind Adelaide in time despite being nearly 900 miles further east.

Sometimes I feed my favorite open complexity problems into LLMs to see what they say. Too often they say something like "I can't solve that because Fortnow's blog says its an open question."

Time for the Computational Complexity Year-in-Review and this year's theorem of the year blog.computationalco... New this year, The Complexity Blog Wrapped by Claude lance.fortnow.com/bl...

A different way to fund science? www.wsj.com/opinion/...

In the holiday spirit, my trip to a magical place where people create things without new fangled technology. And I didn't even have to leave Chicago.

Google reports out on their experiment giving optional feedback generated on Gemini to pre-submission STOC 2026 papers. Upshot: Pretty useful. research.google/blog... Honestly, everyone should run their paper through a reasoning LLM before submitting to an archive, conference or journal.

I'll be in Oxford in England from January 8 to March 26 as a visiting fellow at Magdalen College. Feel free to reach out if you want to connect.

Weird AI is yanking Bill's chain

Mathematicians work in mysterious ways. Perhaps that's why AI find proving theorems so challenging.

Bill Gasarch is running REU-CAAR (Research Experience for undergraduates- Combinatorics, algorithms, and AI for Real Problems) at the University of Maryland next summer. A great opportunity for undergrads. Comes with $7K+room and board. www.cs.umd.edu/proje...

The pre-recorded videos (for the talks whose authors opted in to submit one) are now available on our YouTube channel: m.youtube.com/@FOCS2025 They're also linked from the main website's schedule!

The architect Frank Gehry passed away last Friday at 96. Among many other projects, he designed the Stata Center, the CS building at MIT.

Bill remembers Tom Stoppard

University of Wisconsin is establishing a College of Computing and Artificial Intelligence, one of the first major research university with a college with AI in its name.

My thoughts on the late 80'-early 90's transition for research paper distribution from mail to email to the web.

My little theorem paper is now on arXiv: arxiv.org/abs/2512.0... How I had AI write the paper is more interesting than the paper itself. It felt like working with a grad student. docs.google.com/docu...

It seems the programming world is divided among the people who say AI is useless to them, and others who just use it effectively. Bill hosts a debate.

Happy AI day! November 30, 2020: AlphaFold announced. November 30, 2022: ChatGPT released. We're only three years into the generative AI era.

Following my blog post yesterday, I "vibe coded" a short research paper on my little theorem on S2P search, maybe more of an observation. lance.fortnow.com/pa... Blog post: blog.computationalco... Now what should I do with it?

Let's hear it for the little theorems!

Well, it works in theory...