Erin Carmody

✨ A special delivery from THE SET THEORY GUIDE FOR ARTISTS ✨ open.substack.com/pub/erincarmod…

These essays are so delightful (and challenging) with amazing images. Can't wait for the physical book!

@ShakeelHashim I've realized that since they spent so much money for their AI machines to work it's just ok to lie and make things up.

New terrible definition of superintelligence just dropped:

Let's talk about talking about strongly compact cardinals. 🏋️‍♀️🎖️ erincarmody.substack.com/p/math-in-the-…

"Verification is the real bottleneck of AI" good old truth always getting in the way..

STATUS UPDATE: REVISION NEEDED Currently there's an error flagged by @littmath in the proof of Proposition 12.55 that needs correction to have the full argument. There are also some issues raised by @Tomodovodoo here: github.com/Tomodovodoo/de… (haven't checked them yet) If you found other issues let me know, happy to add them to the working queue. If you want to collaborate on making it into a publishable result, also DM!

@prz_chojecki The language "this settles the problem to the negative" does not match "I trust the process [of the agents having worked out the proof correctly]" at all. I really think you should be much less definitive in your wording.

@prz_chojecki @littmath You "trust the process"? Why AI makes many errors constantly, doesn't even know what a prime number is...

I trust the process here. Models didn't write more than 5-10 pages at a time, there was always a verifier/reviewer model looking over individual additions for correctness. Finally the orchestrator mapped out the arguments and check them too. I didn’t check all the computations myself, but I've read through the strategy and some individual pieces in more detail. That said, I expect there will be more than one correction to be made in the end - as is the case with any human written paper too of that size.

My multi-agent harness powered by GPT-5.4 settled a FrontierMath Open Problem. The result of 2 weeks of 5-10 agents working 24/7: there are no char 3 rank 1 del Pezzo surfaces with more than 7 singularities. This settles the problem to the negative. Details below.

@prz_chojecki Sorry, I'm not going to go back and forth with your LLM. Please check the paper yourself!

Had a great time at the conference! ❤️ erincarmody.substack.com/p/math-in-the-…

I finally published my work on the modal theory of the category of sets on arXiv. This work introduces a new categorical perspective on potentialist systems—the backbone of a series of papers coming out of stealth soon! 🙂

@JDHamkins Seen on Facebook:

@JDHamkins Congrats Joel! It looks great! The essays and images are so good. Such a variety of topics. I feel like this could even complement a calculus class.

I am very pleased to announce that The Book of Infinity is available for pre-order! All my favorite paradoxes and conundrums. #BookOfInfinity #InfinitelyMore

I gave a talk yesterday at the CUNY grad center and had a lot of positive feedback and comments. A collaborator of some of the students sent me a link to a coding challenge that has a similar vibe to killing primes check it out: codegolf.stackexchange.com/questions/1489…

I saw Astoria today in battery park! And it had a human friend nearby watching out 😊

🔥🚨JUST IN: The Unitree humanoid robot was spotted running and playing with children on the streets of New York City.

Throwing darts at the real line. My latest Short Stories column in the AMS Notices. ams.org/journals/notic…

Arthur W. Apter, Eyal Kaplan, Alejandro Poveda: The number of measures on very large measurable cardinals arxiv.org/abs/2603.11710 arxiv.org/pdf/2603.11710 arxiv.org/html/2603.11710

@wtgowers I tried a game and got something right accidentally. Then tried again today and didn't guess correctly but it looks like something is working. I'm going to read it carefully and try to actually play it. Looks interesting!

I think now both games are working properly, with all puzzles showing up. (I've checked this by opening them in incognito mode.)

I've created a couple of mathematical games, both based on word problems in groups or semigroups. One of them could lead to a Polymath project if enough people are interested in it, as it is connected with an open problem. More details in a blog post linked below. 1/2

@WAWoloszyn @unusual_whales I see so it's interested possible buyers. I was trying to apply the pigeonhole principle :)

@uncountableart @unusual_whales But then the same logic would apply—every trade has two sides :) I just checked and by sellers they mean number of homes listed for sale and by buyers they mean an estimated number of people actively trying to buy :)

Home sellers now exceed buyers by over 600,000, marking the widest gap on record, per Redfin.

@WAWoloszyn @unusual_whales Lots of buyers are buying more than one house?

@unusual_whales What is this measuring exactly? For every sale, there needs to be a buyer on the other side of the trade, hence my confusion.