Lean, every mention
52 scenes across 5 shows · ← back to Lean
tap a year for its mentions
Latent Space 83
the MAD Podcast 38
the a16z Podcast 8
No Priors 6
TBPN 4
every year every show Latent Space 83 the MAD Podcast 38 the a16z Podcast 8 No Priors 6 TBPN 4
Verbatim, from the transcripts: passages where Lean comes up on Latent Space, the MAD Podcast, the a16z Podcast, No Priors, TBPN
Inside OpenAI’s Breakthroughs in Mathematical Reasoning
Can AI Learn Mathematical Intuition?
- ▶ 7:45 Lisha Li Obviously, we can't get too much information from the labs who are producing these models, like why, um, you know, how, of what the training recipes are, or how they're kind of advancing in reasoning, um, but at least one of the things we… 2 times in the scene
- ▶ 53:46 Daniel Litt Um, so, for example, uh, you know, with these, this recent list of 10 problems, uh, released by OpenAI, those were all formalized in Lean. 2 times in the scene
🔬 Why Transformers Hit a Wall the Moment Physics Shows Up — Anima Anandkumar, Caltech
- ▶ 5:08 Anima Anandkumar For instance, you know, if we claim that the theorem is correct, we have to go verify that, you know, that's where lean as a formal language can be useful for verification.
- ▶ 6:23 RJ Haneke We have already discussed on the podcast Lean and, uh, and everyone should probably be familiar with neural networks. 8 times in the scene
- ▶ 1:07:42 unnamed speaker Like for example, the, the torch lean work, but, um, the, the applications have really grown and I'm wondering what, what prompted that and like, um, what were some of the lessons you've learned since then?
OpenAI's Dan Roberts: Why AI Can Now Make Discoveries
- ▶ 10:38 Dan Roberts Take problems, present them in a formal language called lean, and then used methods to search for proofs in, in that language and some problems for problems to be representable. 2 times in the scene
Scaling Past Informal AI - Carina Hong, Axiom Math
- ▶ 0:07 Carina Hong Well, before blueprinting, that's human-human collaboration, and lean was a grounding, was a verification, formal language.
- ▶ 2:41 Carina Hong So, math startup, lean startup. 2 times in the scene
- ▶ 9:59 Carina Hong Okay, so, but proof assistants and, you know, formal proof checkers like Lean still found its place, right? 5 times in the scene
- ▶ 10:55 Carina Hong So Lean is a computer program, uh, a bit like for math proofs. 9 times in the scene
- ▶ 15:21 Carina Hong RL for Lean to, to have improvement because of seeing evidence of RL encoding.
- ▶ 15:47 Carina Hong So we heavily rely on kind of data called Lean data, and we kind of talked about Lean as all the, all the data that we have that's Lean proofs, you know, it's correct. 4 times in the scene
- ▶ 20:25 Carina Hong I think a Lean-based system will struggle in those very creative places, which is why we at Axiom actually also invest on something called mathematical discovery. 3 times in the scene
- ▶ 28:10 RJ Haneke So just help me map from the program to the proof, because like I could say, you know, this two line lean program verifies, you know, sort of like whatever
- ▶ 30:25 Carina Hong But if you have proof as Lean, and you have, you know, code, you can choose Rust, which is a strongly typed language. 2 times in the scene
- ▶ 35:13 RJ Haneke So I know I have to know, I'm going to give this input, I'm going to give this output, it has to have these characteristics, and so, and so I write test cases and I write a, so is there an equivalent in lean of this, right, where the… 3 times in the scene
- ▶ 38:47 RJ Haneke You know, when I'm writing code with cloud code, but I can imagine problems becoming big enough in a system like this where you have a gazillion lines of lean. 5 times in the scene
- ▶ 1:02:48 RJ Haneke I mean, I, that brings up also, I, I read somewhere that you guys have a really massive database of lean proofs that you've generated.
- ▶ 1:05:20 RJ Haneke I was sure they're going to announce that they're using lean to, to do like formal verification of proofs and actually generate proofs and then verify them so that they're grounding and reasoning. 2 times in the scene
- ▶ 1:06:58 RJ Haneke And, um, I actually tried it with Cloud Code, um, because it's easier than setting up, uh, you know, your own lean, um, tool chain. 7 times in the scene
- ▶ 1:11:38 Carina Hong I do actually have seen people use Ling and formalization, and they try to do it by hand, you know, not using any AI as a way to learn mathematics. 4 times in the scene
- ▶ 1:19:27 Carina Hong Well, before blueprinting, that's human-human collaboration, and Lean was the grounding, was the verification, formal language.
🔬How GPT‑5 derived new results in theoretical physics and quantum gravity — Alex Lupsasca, OpenAI
- ▶ 1:29:21 Alex Lupsasca We don't think the way lean, which is this language for formal verification reasons.
Mistral: Voxtral TTS, Forge, Leanstral, & Mistral 4 — w/ Pavan Kumar Reddy & Guillaume Lample
- ▶ 41:17 Guillaume Lample So, um, what's nice with Lean and with formal proving is that you don't have to worry about this whatsoever. 3 times in the scene
- ▶ 43:03 unnamed speaker It's, it's for lean.
Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI
- ▶ 30:22 Andrej Karpathy Like if you're a mathematician working in lean, I saw, for example, there's a few releases that really like target that as a domain.
AI That Can Prove It’s Right: Verification as the Missing Layer in AI — Carina Hong
- ▶ 7:33 Corinna Hong Okay, now we saw the Lean code, right, like thousands of lines of Lean code, and sort of read through it, understand it, and, uh, maybe we see some of the techniques as sort of standard, but I think the application of them and the… 4 times in the scene
- ▶ 8:59 Matt Turck So you mentioned lean a second ago, and I guess that's going to take us a little bit into, um, how the, the, the, the, the product and the model work. 10 times in the scene
- ▶ 14:15 Corinna Hong You know, we, we think that it's important for the system to be able to reason both informally and formally, and in a way sort of bridge across these different abstraction from high level intuitions to low level, like more like lean sort…
- ▶ 17:02 Corinna Hong Now, my question of that is, what if you also throw these vast amount of math text data to your AI, but you throw the lean version of them as well? 2 times in the scene
- ▶ 20:07 Corinna Hong And Lin, which we just talked about, the Carl Howard correspondence, exactly turned proof into a computer program.
- ▶ 22:22 Matt Turck And do you need a Lean equivalent for each one of those domains as you expand? 4 times in the scene
- ▶ 28:24 Corinna Hong Like, Action Prover has access, obviously, to a lot of the world's information, but because the lean data is so scarce, it's like a lot less than, say, the amount of code data out there.
- ▶ 36:27 Corinna Hong Um, I think that for the deterministic tooling, it's quite interesting because these are actually written for Lean in the language of Lean. 8 times in the scene
- ▶ 52:17 Corinna Hong Okay, so, so through, through all my good friends telling me about Lean, telling me about Howard correspondence, I believe math is code.
- ▶ 56:42 Corinna Hong Almost all of the new school of mathematicians are accepting Lean. 4 times in the scene
The AI Frontier: from Gemini 3 Deep Think distilling to Flash — Jeff Dean
Captaining IMO Gold, Deep Think, On-Policy RL, Feeling the AGI in Singapore — Yi Tay
Marc Andreessen & Amjad Masad on “Good Enough” AI, AGI, and the End of Coding
- ▶ 27:01 Amjad Masad Like the way that they're training on math, they're using this, uh, sort of like a program language, uh, provable language called lean for proofs, right? 2 times in the scene
Reid Hoffman on AI, Consciousness, and the Future of Labor
- ▶ 31:01 Alex Rampell This is like, if you're able to actually logically construct a proof for something and then validate it, um, there's a whole programming language called lean, which is for that, like that, that stuff is also fascinating.
Weekly Recap: Casey Neistat, OpenAI Cracks Math, The Future of ChatGPT, Apple x F1, Intel Layoffs
- ▶ 3:25 John Coogan Using this, this, uh, program called Lean that is a formal math proof verifier, and there's rumors that Google's building their system to leverage that a lot.
- ▶ 15:37 John Coogan Uh, Daniel lit shows Cheryl on the open AI team saying the model solves these problems without tools like lean, which is a math verifier or coding.
⚡️Math Olympiad gold medalist explains OpenAI and Google DeepMind IMO Gold Performances
- ▶ 6:25 Dr. Jasper Zhang Problems to Lean, and then they use Lean to kind of prove the, the, the problems, uh, to prove the, uh, solve the problems, and, and this year, uh, no longer need, need that, and the same for OpenAI, too. 2 times in the scene
- ▶ 19:23 Dr. Jasper Zhang So, for example, uh, for the logical thinking and reasoning, I think we can, uh, a lot of people trying to use lean, uh, because, like, 2 times in the scene
- ▶ 23:24 unnamed speaker Uh, and then I think the, the last followup I had was about a comment that you made, uh, maybe like three comments into your, your comments, uh, which is, uh, about, about the usage of lean and, uh, and then like sort of verifiability. 7 times in the scene
OpenAI Just Cracked the World’s Toughest Math Challenge — Here's How
No Priors Ep. 90 | With Google's DeepMind's AlphaProof Team
- ▶ 11:05 Laurent Sartran It can be stated in, in some obfuscated way, and how to, uh, translate them in Lean is a major difficulty, and then how to solve them in Lean is a bit, is a bit unwieldy. 2 times in the scene
- ▶ 35:13 Rishi Mehta And, you know, it's still a small minority of the mathematical community that operates in lean, but, um, it's a growing minority. 3 times in the scene
Agents @ Work: Dust.tt — with Stanislas Polu
- ▶ 8:21 Shawn Wang There's a bit of work with, like, Lean, and then with, uh, you know, more recently with, uh, with DeepMind doing, like, scoring, like, Silver on the IMO.