Entrepreneurial Thought Leaders (ETL)
00:32:30 · Ravi Belani, Carina Hong
[00:32:30] S1: Brain computer interface. [00:32:32] S2: Yeah. I think I'm a big nerd and like, I did a neuroscience master because of this, like, nerdy passion. So the reason I... So first principle is why I want to choose it. I want a few that's early enough that, like, no one really has no... Like, no one has any idea what it should be. There are many debates about which way it should be intrusive, non intrusive, a lot of technical problems. We don't really understand neuroscience. How can we understand brain computer interface? But I also wanted to be not just like a hard problem we have no idea about, that it's a problem that in... Involves every single person. Like there's something about every single person needs communication. And communication that isn't just too high. That's why Wispr Flow actually. I was very, very inspired when I first used Wispr Flow. Like, it allows me to type a lot faster and that typing is a way I communicate and that is a big sort of time sink in my life. And then the third thing why I think it's very interesting is because it ties into the past, the present, and the future. Right? Past being AI. Yes. Present being... I think BCI is the present Mhmm. In terms of something early enough. And then the future. The future is if you cannot crack communication between humans, then with AI, all these potential problems of being completely out of the loop. So that's what I would do.
00:32:32 · Carina Hong, Ravi Belani
[00:32:32] S2: Yeah. I think I'm a big nerd and like, I did a neuroscience master because of this, like, nerdy passion. So the reason I... So first principle is why I want to choose it. I want a few that's early enough that, like, no one really has no... Like, no one has any idea what it should be. There are many debates about which way it should be intrusive, non intrusive, a lot of technical problems. We don't really understand neuroscience. How can we understand brain computer interface? But I also wanted to be not just like a hard problem we have no idea about, that it's a problem that in... Involves every single person. Like there's something about every single person needs communication. And communication that isn't just too high. That's why Wispr Flow actually. I was very, very inspired when I first used Wispr Flow. Like, it allows me to type a lot faster and that typing is a way I communicate and that is a big sort of time sink in my life. And then the third thing why I think it's very interesting is because it ties into the past, the present, and the future. Right? Past being AI. Yes. Present being... I think BCI is the present Mhmm. In terms of something early enough. And then the future. The future is if you cannot crack communication between humans, then with AI, all these potential problems of being completely out of the loop. So that's what I would do. [00:33:56] S1: That's That's great. Thank you. What do you believe about building companies today that you wish you'd understood when you were sitting where these students are?
00:33:56 · Ravi Belani, Carina Hong
[00:33:56] S1: That's That's great. Thank you. What do you believe about building companies today that you wish you'd understood when you were sitting where these students are? [00:34:06] S2: It's really hard. Like, it's just really hard. I think, like... And and and I think... Oh, one thing that probably is worth worth mentioning is that it's it's very hard is not, like, that informative. It's it's very hard at the moments you least expect. So if you're someone who's very, like, I want my things, like, planned. I want my trips planned. Very, very hard to get past that sort of mental hurdles at times. Like, it will hit you as a weirdest day. Most of the hard moments will come after very good, like, achievements. After the grind, I'm now sick. Apologies about that. I think the uncertain pain and suffering is something I didn't understand. And I understood something about it's gonna be difficult or it's going to be spontaneous, but not both. [00:34:55] S1: It's going be difficult and spontaneous. Yeah. And the suffering. [00:34:58] S2: Yeah. And that makes it worse if you're like a planner. [00:35:00] S1: Yeah. Yeah. Because it's coming out of nowhere and you just have to build that resilience even more. Okay. Terrific. Can I ask what's the biggest professional failure you've had and what did you learn from it?
00:35:00 · Ravi Belani, Carina Hong
[00:35:00] S1: Yeah. Yeah. Because it's coming out of nowhere and you just have to build that resilience even more. Okay. Terrific. Can I ask what's the biggest professional failure you've had and what did you learn from it? [00:35:12] S2: Oh yeah. I got this... Okay, so there was this high school math competition And I remember I was devastated because I am so close to the place I wanted to be. And later on, I looked at why that is. I also realized that it's because of something that's my biggest strength. Mhmm. That's my biggest, like, that... In that context, failure. I... There are four problems, Chinese Math Olympiad. I, like, plunge into the last one, number theory, which is usually the hardest problem and also the one that I love the most. And I almost couldn't get out of it. I'm very obsessive. Right? Like, the sort of mental resilience. And I, like, ended up spending the entire time working on that problem. What I learned from it is that, like, a lot of things that is very context wise, like feature or bug. Like, I don't think I can be who I am today without that trait. But also in that local moment, it felt very demoralizing. So I think one thing is that maybe there is not a lot of right or wrong. What are... Like, time management is, of course, good and, like, lack of prioritization is, of course, bad. Not necessarily. [00:36:30] S1: Mhmm. [00:36:30] S2: Right?
00:36:30 · Carina Hong, Ravi Belani, S3
[00:36:30] S2: Right? [00:36:30] S1: Yeah. Yeah. It's like the ancient Greek heroes. Your strength is your... It can be also a tragic flaw. Awesome. Thank you. We'll we'll do student driven questions. Great. And please go close to the mic so they can hear in the back. [00:36:40] S3: Yep. This is good. That's perfect. If if I understand correctly, just very concisely, and you can correct me if I'm wrong, Axiom Math is basically building mathematician with the broader vision of building recursively improving AI systems. Why why do you see mathematical reasoning specifically as a particularly, like, effective path towards achieving recursively improving AI systems? [00:37:06] S2: Yeah. I probably would say formal mathematics specifically, which is the term math into machine checkable code. And so then basically by turning into code, you solve more math. So that's code for math. You also have math for code. So I think it's probably right to think of us as a company that's very close to coding. So we will be using the formal mathematical reasoning engine, the same stack that we built to prove software and also hardware. So that's, at the beginning, I had this idea where solve math, you solve everything. A lot of, I think, devils are in the details and the right sort of framing, I would say, is that every single part of the stack you build is now reused to to build coding, which I think is going to change the world. [00:37:54] S1: Thank you for the question.
00:37:54 · Ravi Belani, S4, Carina Hong
[00:37:54] S1: Thank you for the question. [00:37:56] S4: Some people will say before starting a company, you should essentially join, like, a big lab or a big company and kind of gain that training experience before starting. So in your opinion, do you think that's still relevant or would you advise someone to just jump into the deep waters jump into deep waters and just go for it? [00:38:17] S2: I think there's a lot of value in starting out at a big lab. If it's a big lab that you think allows you to learn and see the frontier. I'm in this weird situation where I think academia actually was the frontier of this emerging field, AI for formal math. I was reading all these papers that are academic publications because I think big labs were not very interested in. Well, that kind of like a spontaneous thing for like a year or a couple of quarters, and then they just dropped it. So I think if you're working in, for example, a field that is well developed, like mature, probably frontier the Frontier... Frontier has Labs have the Frontier, and it's good to start out that way. That would be my suggestion. [00:38:56] S4: But so now with, like, AI advancing so quickly, do you think there's still time to join, a Frontier Lab or, like, learn as you essentially get your training in before you start or do you capitalize now and just start?