{{}, {{}}, … }
Joined July 2012
- Tweets 775
- Following 243
- Followers 1,582
- Likes 23,670
148 Photos and videos
May 30
so many advisories for zcash/zebra 🦓
makes it an interesting repo to study
github.com/ZcashFoundation/z…
7
209
May 21
I remember proving f(2) = 1 long time ago. true about showing f(k²) = k is easy.
This belongs to a pretty interesting set of combinatorial geometry problems by Erdös and friends ... ⊞
3
257
May 20
It’s cool that they have released two companion papers and also a 125 page chain of thoughts pdf.
should be interesting to read these papers. 🌐
AI has now solved a major open problem -- one of the best known Erdos problems called the unit distance problem, one of Erdos's favourite questions and one that many mathematicians had tried.
openai.com/index/model-dispr…
1
3
361
May 14
The constructions 🏗️ of the DGAs based on undecidability of problems in group (monoid, algebra, ...) theories is pretty cool. It's nice to see concrete examples. The arxiv paper has only sampled some of the more elementary examples. But GDM has more:
github.com/google-deepmind/s…
May 13
Recently, Aletheia (courtesy of Tony Feng) helped Stanford mathematician Ciprian Manolescu tackle Problem 5.16 from the K3 list (3rd version of Kirby’s List) in low-dimensional topology, autonomously generating proofs for the new paper “Undecidability problems for semifree DG algebras.”
Per Ciprian: “It’s a problem in pure algebra but one of interest to topologists. By the work of Chekanov and Eliashberg, semifree differential graded algebras (DGAs) appear as invariants of Legendrian knots. The problem asked whether there is an algorithm to tell these algebras apart from one another, and the answer turned out to be no.”
Ciprian also would like to “propose K3 as a challenging, long-term benchmark for the progress of AI on math research problems. I expect that even if we reach AGI, most of these problems won’t be immediately solved.”
Paper link: arxiv.org/abs/2605.08122
1
2
387
May 14
more fun to read are the refs actually regarding DGA invariants for 🪢:
- arxiv.org/pdf/math/9709233
- arxiv.org/pdf/math/0101145
94
May 13
I was excited to read this. The proof is very elementary.
Lemma 4. & Lemma 5. are very simple to prove (can be even left as an ex. for the reader).
Main ingredients are:
- the Neumann transformation of the group representation.
- ✨ Serre's work on actions of groups on trees
May 12
"Every finite group admits a just finite presentation"
New preprint by Marc Lackenby, with significant assistance from the AI co-mathematician
arxiv.org/abs/2605.10402
1
361
May 13
138
May 6
Fun night read. Timeline of public security releases are interesting 🐢
May 6
For a Fistful of Dollars: Less than $100 of Compute Surfaces Pre-auth RCE in Apache httpd
Write-up: striga.ai/research/apache-ht…
2
2
530
May 6
Some thoughts from one of the HTTPD devs
eissing.org/icing/posts/resp…
Looks like they are also flooded by bug reports …
1
136
Apr 29
Created some exercises to prove Erdös problem 1196 on my own (for those of us who don’t want to spoil the proof).
At the heart of it are two tail estimations and the basic identity regarding von Mangoldt function plays the key role:
log = Λ ∗ 𝟙
🪡
1
236
Apr 29
🤖 also came up with an adjacent idea of instead of phrasing the solution using markov chains, one can instead look at divergence of flows on the related poset graph. It does again require tail estimates of some sums but the idea is pretty intuitive …
114
Apr 28
Pretty cool through unsanitised X-Stat header key=value injection …
Huge kudos to GitHub for addressing the issue 👏
Full technical breakdown here → wiz.io/blog/github-rce-vulne…
1
407
Apr 24
⟁ can you cover an equilateral triangle of side length n ɛ with:
- n² 2 unit equilateral triangles ?
- n² 1 unit equilateral triangles?
fun puzzle to think about. I came up with Alexander Soifer's construction.
2
1
3
340
Apr 24
some refs:
- wfnmc.org/mc20101.pdf
- seewoo5.github.io/jekyll/upd… (many related refs ...)
funny coming across @antimath3 again after the @axiommathai 's (@KenOno691) announcement from a few months ago ...
1
1
217
Apr 21
pretty cool to explore:
mathnet.csail.mit.edu/
🧮
Apr 21
Today, MIT & the IMO released MathNet, the world’s largest dataset of International Math Olympiad problems & solutions 🌍
MathNet is 5x larger than previous datasets & is sourced from over 40 countries across 4 decades: bit.ly/4u1bhBC
3
289
Mar 31
A simple writeup would be more educational than a talk. Main point is clear regarding the AI trend. It's more fun to read about the bugs discovered:
- github.com/torvalds/linux/co…
- github.com/TryGhost/Ghost/se…
- gist.github.com/Saw-mon-and-…
📚 arxiv.org/search/?searchtype…
Mar 30
I think we just got a demo of Mythos and I’m surprised nobodies talking about it.. 💔
In what might be the first instance the general public has seen of Claude Mythos
Mythos (TBD) just uncovered a critical zero-day vulnerability in Ghost, an open-source platform with over 50,000 stars on GitHub that has never had a critical security flaw in its entire history.
It identified a highly complex "blind SQL injection" a flaw so subtle you can't even see the output, only how the server delays its response. When asked to prove the severity of the bug, the model autonomously wrote a custom Python exploit script that successfully navigated the blind injection to extract the admin API key, secret, and password hashes from the database completely unauthenticated…
This is genuinely game changing because it proves frontier models can now actively discover, reason through, and successfully build exploits for invisible vulnerabilities in enterprise-grade architecture that human developers missed for years.
Cyber security companies are cooked.
5
562