/AI7h ago

OpenAI's Sebastien Bubeck says GPT-5.5 disproved the sum-product conjecture using Boris Alexeev's prompting workflow

The prompt included a unit distance problem solution

--0--
Quote posts
Jimmy Apples 🍎/acc#837
Comments
Luca Ambrogioni
Csaba Szepesvari
Sebastien Bubeck#51
Original post
Sebastien Bubeck@SebastienBubeck#51inAI

Expert prompting was done by Boris Alexeev btw 😅

Sebastien Bubeck@SebastienBubeck

GPT-5.5 (the one available right now to everyone) can also disprove the sum-product conjecture: https://chatgpt.com/share/6a187d12-7d58-83e8-87bd-0ee312378028 . I didn't reveal it before because I think it is good to give some space to the community to absorb these new capabilities. In particular the humans involved in the discovery should get all the credit for this amazing breakthrough. We all have some work to do to align on cultural norms in this new world.

6:01 PM · Jun 3, 2026 · 6K Views
Sentiment
Sentiment building, check back later.
Cluster Engagement
-
Views
-
Comments
-
Reposts
-
Bookmarks
Expand data
Posts from X
Most Activity
Most ActivityTimeline
VIEWS7.1KLIKES95RETWEETS9
Jimmy Apples 🍎/acc@apples_jimmy

the public doesn’t see that current models can do so much, you don’t need the internal locked behind lab doors.

Also not sure I agree with Sebastien on letting the math community slowly work out these problems themselves, they’ve been notorious gatekeepers.

Sebastien Bubeck@SebastienBubeck

GPT-5.5 (the one available right now to everyone) can also disprove the sum-product conjecture: https://chatgpt.com/share/6a187d12-7d58-83e8-87bd-0ee312378028 . I didn't reveal it before because I think it is good to give some space to the community to absorb these new capabilities. In particular the humans involved in the discovery should get all the credit for this amazing breakthrough. We all have some work to do to align on cultural norms in this new world.

6hViews 7.1KLikes 95Bookmarks 5
BOOKMARKS10REPLIES3
Sebastien Bubeck@SebastienBubeck

@littmath totally. I am however a big fan of moving into the direction of "Kolmogorov Appendix" that I mentioned a while ago, where you show the "minimal prompt" that allows the model to recover your result. Details TBD (and there are many!)

Daniel Litt@littmath

@SebastienBubeck Prompting with solution to unit distance and indicating the same approach can be used for sum-product is a pretty big hint!

7hViews 5KLikes 64Bookmarks 10