Manifold markets: Will Gary Marcus still say that deep learning models can't reason after they prove the Riemann hypothesis?
16.09.2025 22:49 โ ๐ 4 ๐ 0 ๐ฌ 0 ๐ 0@vatter.bsky.social
Mathematician at the University of Florida. Here for the hot takes. Also @VinceVatter@mathstodon.xyz
Manifold markets: Will Gary Marcus still say that deep learning models can't reason after they prove the Riemann hypothesis?
16.09.2025 22:49 โ ๐ 4 ๐ 0 ๐ฌ 0 ๐ 0Thanks to co-PIs Sarah Sword, @jaypantone.com, and Ryota Matsuura, and to our advisory board, for countless hours shaping what good proof feedback should look like.
14.08.2025 21:08 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0Teaching intro to proofs this semester or next? We have a couple of spots left for college/university faculty who would be interested in participating in a field test. Details on the site, or reply or DM.
14.08.2025 21:08 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0Try it: hallmos.com
Free, browser-based, no account. Students, professors, and the proof-curious are welcome. Choose from common intro-to-proofs exercises, or use the sandbox with your own exercise.
I'm excited to share: we built a free AI coach for intro-level proof writing. It critiques drafts, points out gaps, and helps you or your students iterate without giving away the answer. Also for lapsed mathematicians to see if they've still got it. Demo below, link in reply.
14.08.2025 21:07 โ ๐ 6 ๐ 0 ๐ฌ 1 ๐ 0Federal Funding for Mathematics ResearchโWhatโs Changed in the Last Six Years? This article was solicited in early 2024 and completed in Fall 2024. It therefore does not reflect the changes that have occurred, and which are still going on, at federal science-funding agencies in 2025. For example, website links given in the article may not be active at the time of publication. - Ed.
Published today in the Notices of the AMS. Lol.
16.06.2025 16:03 โ ๐ 3 ๐ 0 ๐ฌ 0 ๐ 0The number of American kings over time: a harrowing saga of job insecurity, with a notable dip after 1776 and a suspicious lack of improvement since.
15.06.2025 19:40 โ ๐ 3 ๐ 0 ๐ฌ 0 ๐ 0meanwhile, wikipedia is in existential crisis over whether โ4 deserves its own page
30.05.2025 15:34 โ ๐ 25 ๐ 2 ๐ฌ 1 ๐ 0Line graph showing the frequency of years ending in โ00,โ from 1800 to 2020, in English books.
people like years more than ever though. reliably, year X mentions peak during year X+2.
29.05.2025 20:15 โ ๐ 8 ๐ 1 ๐ฌ 0 ๐ 0Google Ngram Viewer line graph showing the frequency of digits 0 through 9 in English books from 1800 to 2019, with usage peaking around 1980 and declining sharply afterward.
interest in the natural numbers is at an all-time low
29.05.2025 19:09 โ ๐ 16 ๐ 5 ๐ฌ 5 ๐ 5Just to say how much I like the digit party game digit.party created by Robert Brignall and @vatter.bsky.social.
28.05.2025 21:04 โ ๐ 1 ๐ 1 ๐ฌ 0 ๐ 0Of course they would choose a math major
08.05.2025 23:00 โ ๐ 7 ๐ 0 ๐ฌ 0 ๐ 0FOX NEWS: Cardinals assess Pope Francis' legacy before vote. Delegates include $\aleph_0$, $\aleph_1$, $\beth_2$, and others, as the continuum hypothesis looms over the ceremonies.
06.05.2025 16:38 โ ๐ 23 ๐ 3 ๐ฌ 5 ๐ 5Itโs final exam season again, and I like to display a clock while proctoring.
But all the web clocks google suggests suck.
So I made one (okay, tbh, chatgpt made it under my supervision):
digit.party/clock
For a fuller glimpse of our dystopian future:
spinup-000d1a-wp-offload-media.s3.amazonaws.com/faculty/wp-c...
AI: Hello, is this Ms. Zhang San? User: This is Ms. Zhang Sanโs robot phone assistant, may I ask what you want to discuss with her? AI: Sorry for disturbing you, goodbye!
AI debt collector exhibiting speciesism
14.04.2025 11:48 โ ๐ 4 ๐ 0 ๐ฌ 2 ๐ 0A001339 a(n) = Sum_{k=0..n} (k+1)! binomial(n,k). 1, 3, 11, 49, 261, 1631, 11743, 95901, 876809, 8877691, 98641011, ...
Needless to say, A001339 is not the same sequence, does not count preferential arrangements, and neither sequence has anything to do with the egf 1โ(2โexp(x)). All of this was completely made up.
13.04.2025 14:47 โ ๐ 12 ๐ 0 ๐ฌ 1 ๐ 0'bc I'm tired'
Excellent example of how LLMs just reproduce what they've seen in their training data.
ChatGPT-4o: "go to OEIS bc iโm tired: [1, 1, 1, 2, 5, 17, 73, โฆ] โ OEIS A001339"
you're what now chatgpt?
12.04.2025 19:30 โ ๐ 28 ๐ 3 ๐ฌ 3 ๐ 2I mean, itโs still better to have money than not have money right?
04.04.2025 17:08 โ ๐ 2 ๐ 0 ๐ฌ 0 ๐ 0Instructions for assignment: Color each square blue. Color each rectangle yellow. Color each circle red.
you'll never guess what color my daughter's squares ended up
03.04.2025 23:31 โ ๐ 11 ๐ 1 ๐ฌ 3 ๐ 0Everyone hates springing forward but loves falling back. Proposal: Abolish leap days, and redistribute those hours in the form of 6 annual "fall back" holidays. More sleep, less leap. A clear pareto improvement to our calendar.
09.03.2025 15:45 โ ๐ 3 ๐ 1 ๐ฌ 0 ๐ 0Andrรฉ's Reflection Method ๐
07.03.2025 18:57 โ ๐ 2 ๐ 0 ๐ฌ 1 ๐ 0Logical proof that "months" are fake.
Most people born in April are Aries.
All Aries are born in March or April.
Most people born in March or April were born in March.
Therefore, most people born in April were born in March.
I asked GPT-4o to flip a coin 10,000 times. Same prompt, same parameters, but the *odds of getting Heads* took on 42 different probabilities (none below 85% btw). Obv don't use gpt-4o to flip your coins, but what the hell is going on? Here are some thoughts.
towardsdatascience.com/avoidable-an...
Actually it's ok to divide by zero as long as you use the proper facilities. ๐งฎ
04.03.2025 13:22 โ ๐ 31 ๐ 5 ๐ฌ 3 ๐ 0A screenshot from Containing All Permutations featuring Theorem 5 (Anonymous 4chan poster): For all $n\ge 1$, every $n$-superpermutation has length at least $n!+(n-1)!+(n-2)!n-3$.
Want more superpermutation fun? My free article Containing All Permutations in the Mathematical Monthly covers this and related problems, and features, of course, Theorem 5 (Anonymous 4chan poster).
www.tandfonline.com/doi/full/10....
There's an article in Scientific American today about superpermutations mentioning @robin.boskerenterprises.com, @gregegansf.bsky.social, @njohnston.ca, @jaypantone.com, me, and of course, โAnonymous 4chan Poster.โ
www.scientificamerican.com/article/the-...
Reference list with a citation for The Terminator (1984) by James Cameron, published by Orion Pictures, nestled alongside academic papers on AI and security.
26.02.2025 16:14 โ ๐ 9 ๐ 0 ๐ฌ 0 ๐ 0'Please help us improve Google' popup with buttons 'NO THANKS' and 'SHOW QUESTION'.
the first step is admitting you have a problem. the second step is asking random users to fix it for free.
25.02.2025 15:22 โ ๐ 3 ๐ 0 ๐ฌ 0 ๐ 0