Sign in

Sebastien Bubeck

@sbubeck.bsky.social
4.1K followers 171 following 29 posts

I work on AI at OpenAI. Former VP AI and Distinguished Scientist at Microsoft.

PostsRepliesMedia
Sebastien Bubeck @sbubeck.bsky.social · 07/08/2025
www.youtube.com/live/0Uu_VJe... 10am PT today!
youtube.com
Introducing GPT-5
YouTube video by OpenAI
270
Sebastien Bubeck @sbubeck.bsky.social · 12/06/2025
This was a really fun conversation with @billgates.bsky.social and Peter Lee, and I hope you enjoy it too! (roughly we talk about our respective first contacts with genAI/LLM/sparks and what we see as missing steps/future steps for it to truly transform healthcare) youtu.be/N0w9uO7kH80?...
youtu.be
How AI is reshaping the future of healthcare and medical research
YouTube video by Microsoft Research
170
Reposted by Sebastien Bubeck
The Onion @theonion.com · 02/05/2025
Impressive ‘GTA VI’ Trailer Features Characters Claiming They’re Sentient, Begging For Release From Digital Prison theonion.com/impressive-g...
Impressive ‘GTA VI’ Trailer Features Characters Claiming They’re Sentient, Begging For Release From Digital Prison
393256309
Sebastien Bubeck @sbubeck.bsky.social · 25/03/2025
0100
Sebastien Bubeck @sbubeck.bsky.social · 02/02/2025
Correct!
030
Sebastien Bubeck @sbubeck.bsky.social · 31/01/2025
Overall I think o3-mini will be a very useful model for the academic community. Learn more here: openai.com/index/openai...
openai.com
OpenAI o3-mini
Pushing the frontier of cost-effective reasoning.
050
Sebastien Bubeck @sbubeck.bsky.social · 31/01/2025
Interestingly though the reference it gives is not quite the correct one, but it is very closely related! In general I have found that references are "fuzzily correct", giving some mixed up version of authors/journals/titles, but surprisingly still useful!
160
Sebastien Bubeck @sbubeck.bsky.social · 31/01/2025
o3-mini is a remarkable model. Somehow it has *grokked arxiv* in a way that no other model on the planet has, turning it into a valuable research partner! Below is a deceitfully simple question that confuses *all* other models but where o3-mini gives an extremely useful answer!
2140
Sebastien Bubeck @sbubeck.bsky.social · 08/01/2025
With MIT license. Enjoy! huggingface.co/microsoft/ph...
huggingface.co
microsoft/phi-4 · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
3276
Sebastien Bubeck @sbubeck.bsky.social · 03/01/2025
Nice to see SLMs making it to the 10 Breakthrough Technologies of last year!!! www.technologyreview.com/2025/01/03/1...
technologyreview.com
10 Breakthrough Technologies 2025
What will really matter in the long run? That’s the question we tackle each year as we compile this annual list.
2113
Sebastien Bubeck @sbubeck.bsky.social · 13/12/2024
and if you're the type to be worried about contamination (American Mathematics Competition from last month, after phi-4 was trained):
0112
Sebastien Bubeck @sbubeck.bsky.social · 13/12/2024
151
Sebastien Bubeck @sbubeck.bsky.social · 13/12/2024
arxiv.org/abs/2412.08905 Hope you like it!
arxiv.org
Phi-4 Technical Report
We present phi-4, a 14-billion parameter language model developed with a training recipe that is centrally focused on data quality. Unlike most language models, where pre-training is based primarily o...
3385
Sebastien Bubeck @sbubeck.bsky.social · 09/12/2024
Pretty crazy ...
081
Sebastien Bubeck @sbubeck.bsky.social · 06/12/2024
It's getting there ...
0152
Sebastien Bubeck @sbubeck.bsky.social · 04/12/2024
They made a pretty cool poster for the event too 😄
2152
Sebastien Bubeck @sbubeck.bsky.social · 04/12/2024
Tomorrow morning at the Simons Institute : Sparks versus embers, can LLMs solve major open mathematical conjectures? (FWIW I agree with everything in the Embers paper, so I guess the debate will be about the conclusions to draw from current evidence!) simons.berkeley.edu/talks/sebast...
simons.berkeley.edu
Debate: Sparks versus embers
Debaters: Sebastien Bubeck (OpenAI), Tom McCoy (Yale) Discussants: Pavel Izmailov (Anthropic), Ankur Moitra (MIT) Moderator: Anil Ananthaswamy
0196
Sebastien Bubeck @sbubeck.bsky.social · 03/12/2024
Sure
010
Sebastien Bubeck @sbubeck.bsky.social · 02/12/2024
Yes, I was a bit hyperbolic:-). One of the cleanest open problem that is still open and doable I think is to prove a n^{3/2} sqrt(T) lower bound for BCO. I'd love to see that.
010
Sebastien Bubeck @sbubeck.bsky.social · 02/12/2024
This should be interesting! forum.openai.com/public/event...
forum.openai.com
Virtual Event: The Future of Math with o1 Reasoning - Event | OpenAI Forum
About the Talk:Fields Medal-winning mathematician Terence Tao makes his second appearance in the OpenAI Forum alongside OpenAI’s SVP of Research, Mark Chen to explore a future where mathematics and ar...
24811
Sebastien Bubeck @sbubeck.bsky.social · 01/12/2024
I guess before various versions of what's described here www.scientificamerican.com/article/ai-w... take place
scientificamerican.com
AI Will Become Mathematicians’ ‘Co-Pilot’
Fields Medalist Terence Tao explains how proof checkers and AI programs are dramatically changing mathematics
040
Sebastien Bubeck @sbubeck.bsky.social · 01/12/2024
Some more of the resolutions that happened in the last few years: arxiv.org/abs/2406.18672 arxiv.org/abs/2406.06506 arxiv.org/abs/2303.04752 arxiv.org/abs/2211.05753 arxiv.org/abs/2111.11316
arxiv.org
Eve, Adam and the Preferential Attachment Tree
We consider the problem of finding the initial vertex (Adam) in a Barabási--Albert tree process $(\mathcal{T}(n) : n \geq 1)$ at large times. More precisely, given $ \varepsilon>0$, one wants to outpu...
063
Sebastien Bubeck @sbubeck.bsky.social · 01/12/2024
It's looking like all the open problems I have thought about in the last 10 years are now solved (or in some cases on the verge of being solved)? Latest case in point this beautiful new paper: arxiv.org/abs/2411.18614 . I'm glad we (humans) got all this results just in time!
arxiv.org
Optimal root recovery for uniform attachment trees and $d$-regular growing trees
We consider root-finding algorithms for random rooted trees grown by uniform attachment. Given an unlabeled copy of the tree and a target accuracy $\varepsilon > 0$, such an algorithm outputs a set of...
3305
Sebastien Bubeck @sbubeck.bsky.social · 27/11/2024
Was there a paper hinting at that 18 months ago? Hmmm 🤔🤣🤣
071
Reposted by Sebastien Bubeck
Thomas Wolf @thomwolf.bsky.social · 22/11/2024
Great AI people starter pack from @chris.bsky.social! go.bsky.app/KRsy8pF
go.bsky.app
AI Bluesky
Join the conversation
77212
Sebastien Bubeck @sbubeck.bsky.social · 24/11/2024
This was a really special time, including for the fact that it was pre-"academic twitter" ... I miss those days.
130
Sebastien Bubeck @sbubeck.bsky.social · 24/11/2024
No, but it's hard to attain the same clarity with the current methods ..
020
Sebastien Bubeck @sbubeck.bsky.social · 23/11/2024
Blast from the past, wrote these equations in this form 6 years ago, I remember it felt quite clarifying at the time ... A bygone era unfortunately
2485
Reposted by Sebastien Bubeck
Ananya Kumar @ananyak.bsky.social · 23/11/2024
Had a fantastic time at OpenAI’s dev day in Singapore! So much energy here, and the depth of knowledge policy makers have about AI is incredible. Bumped into some old Singaporean classmates of mine too
042
Sebastien Bubeck @sbubeck.bsky.social · 23/11/2024
AGI is the mission 😁
130
Sebastien Bubeck @sbubeck.bsky.social · 22/11/2024
Today is the 6 months anniversary of phi-3, I wonder what these guys have been cooking up since then 🤔
1172
Sebastien Bubeck @sbubeck.bsky.social · 20/11/2024
So, this thing is for real? AI Twitter is going to reassemble here? Who should I follow?
3101