Theo Conjecture solves 35-year-old math problem, finds a term no one predicted
otalp
33 points
9 comments
July 29, 2026
Related Discussions
Found 5 related stories in 360.2ms across 15,380 title embeddings via pgvector HNSW
- AI solves 20 year old conjecture in graph theory never_giveup · 15 pts · July 20, 2026 · 60% similar
- Amateur armed with ChatGPT solves an Erdős problem pr337h4m · 238 pts · April 25, 2026 · 60% similar
- Epoch confirms GPT5.4 Pro solved a frontier math open problem in-silico · 218 pts · March 24, 2026 · 56% similar
- The AI revolution in math has arrived sonabinu · 74 pts · April 13, 2026 · 56% similar
- An OpenAI model has disproved a central conjecture in discrete geometry tedsanders · 962 pts · May 20, 2026 · 55% similar
Discussion Highlights (8 comments)
Chinjut
This article appears to be AI-written slop.
KaiserPro
I thought this turned out to be a bug in the solver?
margalabargala
The AI-generated writing makes this nearly impenetrable. The AI used does not have the same intuition as a human about what math concepts will be easy vs hard for its audience to grasp, and as a result this is simultaneously rambling and opaque. I'm sure that if I had a back-and-forth with AI about the topic, it would answer all my questions and explain it well. This article isn't that. At least the site had the grace not to list an author's name.
pretzellogician
The article isn't great. But it's nice to know about TxGraffiti! I always wondered what happened with Graffiti.
BigTTYGothGF
One program makes up questions, another one solves them.
firasd
If I'm understanding this correctly, here 'solve' is being used in a loose or miniature sense. ie. it's not proving or disproving the conjecture but is a solution for how one part of the expression involved in analyzing it is evaluated
Mathnerd314
Paper is easier to read https://www.researchgate.net/publication/410906746_The_Havel...
eulgro
The website is unusable, scrolling is slow as hell.