Transcriptly

From Google DeepMind to a $8B Superintelligence Startup | ReflectionAI, Misha Laskin

EO31:49Added Aug 31, 2026

In 2025, investors backed Misha Laskin’s new company, Reflection AI, with two rapid rounds of funding totaling $2.1B. He’s the scientist who helped build Gem...

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format

  1. The moment I saw the possibility of superintelligence

  2. 00:00My co-founder was one of the people who made key contributions to the project of
  3. 00:05AlphaGo and he was one of the handful of people who flew out to the match with Lisa Doll. I think the thing that really
  4. 00:13made me internalize deeply there's a super intelligence here and [music] imagine what those things might look like in the future and I think this this
  5. 00:19was true for many other people as well was move 37. This was a famous move where Alph Go in its match against Lisa
  6. 00:26Doll made a move that looked like a mistake initially. It looked like potentially a bug. Some members of the
  7. 00:32team thought that in the same way that language models hallucinate like they'll sometimes make something up that this was game agent basically making up a
  8. 00:38move that was incorrect. It seemed like Lisa Doll or at least the commentators thought that as well. Few moves later,
  9. 00:43maybe 10 moves or something like that later. It turned out that this was actually a brilliant move that was smarter than anything that any of the
  10. 00:50humans who were viewing the match could have imagined. It was so smart that everyone thought it was dumb. What that
  11. 00:56meant was that an AI system had discovered a strategy that was fundamentally more creative. And it made me think about what will a world look
  12. 01:02like when you have move 37s across every category of knowledge work. You know, a mathematician asks an AI to do a certain
  13. 01:09task for it and comes back to it with a move 37. It comes back to it with a proof that the mathematician had never even [music] considered that was
  14. 01:14correct. What I think what end up happening is that in the same way that we're feeling some people [music] are starting to feel the AGI now where there
  15. 01:20are kind of semblances of sparks of intelligence. I think we'll start getting into this point in time in the
  16. 01:25notsodistant future, probably in the next couple of years, where we're starting to feel the ASI, the artificial super intelligence, where move 37s start
  17. 01:32popping up across different areas of knowledge work and doing things that basically expand our creativity, right?
  18. 01:38Because when move 37 happened, it actually expanded our knowledge of the game of go and what was possible. So, it
  19. 01:43was a very it happened and now there's this new strategy that people are aware. How do people who play chess or go, how
  20. 01:49do they interact with AIs? They actually learn from them. They learn to get better from interacting with these AIs and I think we'll start learning a lot
  21. 01:55from these systems in [music] the coming years and that's very exciting. I'm Misha. I'm the CEO and co-founder of

  22. Reflection AI’s mission and vision

  23. 02:03Reflection. At Reflection, we are building super intelligence. The question is how? Our belief is that if
  24. 02:09you solve the problem of autonomous coding, you will [music] solve the super intelligence problem more broadly. And
  25. 02:15that's kind of our path. Now in terms of you know how we got here the team [music] pioneered a lot of breakthroughs
  26. 02:21in AI over the last decade my co-founder Giannis was [music] one of the key architects of systems like deep Q
  27. 02:28networks alpho alpha zero and then Giannis and I worked together very closely on Gemini where we led a lot of
  28. 02:34the post- training work [music] for producing Gemini 1 and 1.5 we kind of realized that two ingredients had come
  29. 02:39together that would enable you to take these language models create not just useful co-pilots or chat assistants, but
  30. 02:47intelligent capable autonomous systems. [music] And these two ingredients, large language models are very broad in
  31. 02:54general. And the other ingredient was reinforcement learning as a technology which enables scaling up [music] the
  32. 03:01autonomy of language models. We thought these two ingredients had come together were kind of mature enough technologically that you could combine
  33. 03:08them and produce something that would be a highly capable super intelligent autonomous [music] system.

  34. From a boy who loved physics to teaching myself AI

  35. 03:30I was born in in Russia and when the Soviet Union collapsed my family and I I moved to Israel. I mean this was I I
  36. 03:36don't remember this. I was one and then grew up first half my childhood in Israel, second half of my childhood in
  37. 03:42Washington state because we moved around fairly frequently. I didn't have the same, let's say, long form bonds that
  38. 03:48some other people do when they grow up. Lifelong kind of friendships going from childhood all the way to adulthood. As a
  39. 03:54result, I actually ended up spending a lot of time kind of especially when we were in the states alone and with books.
  40. 04:00I mean we had my parents brought a lot of books to the states and I did have friends but I also spent a lot of time
  41. 04:06after school just looking at my parents' library reading various things at the time you know there's definitely you
  42. 04:12know as a kid you start feeling pretty lonely about that but looking back I don't think I would have cultivated the
  43. 04:17interest that I did if I didn't have a lot of time on my own to be bored and think in some sense like boredom is a
  44. 04:23gift that you only appreciate in retrospect I was interested in physics and in literature when I kind of had all
  45. 04:29this time on my hands hands but ended up kind of hard committing to physics and the reason was that when when I read
  46. 04:35about kind of all the kind of most impactful science that had been done and the technology that produced. I would
  47. 04:41look back at the like technological artifacts that we have today and try to derive how are those originated and so an example is is a computer obviously
  48. 04:48incredibly impactful technology everyone uses today and and I was asking myself well how was that invented and you can
  49. 04:54trace it back you can go much further back as well but really there are some core components are invented that
  50. 04:59enabled this and one of them was a transistor and a transistor was invented by a theoretical physicist named John Bardin similarly when you think of
  51. 05:05technologies like GPS and you trace back to what is the ingredient that enables those technologies to work. It turns out
  52. 05:12it's also physics. GPS relies heavily on Einstein's theory of special relativity. And so I wanted to work at that root
  53. 05:18node of the science that will enable everything else that comes after it. I wanted to work on the stuff that if we
  54. 05:24look at the technology we have a few decades from now and we trace back to what was the breakthrough that enabled
  55. 05:30that. I wanted to be working on those things. That's what got me interested in physics. What I learned when I was in
  56. 05:36the PhD was that you have to kind of think about the science but not just the
  57. 05:41impactful science at any time. You have to think about what is the impactful science of the time today. The work in
  58. 05:46physics that I was reading about was basically done anywhere from 60 to 100 years ago. That's when all of the these
  59. 05:52kind of impactful inventions were made. And I realized that the field at least for me had crystallized a bit. It was
  60. 05:58hard for me to see how like what foundational breakthroughs I could be a part of. and not necessarily
  61. 06:04individually but as a team that would enable the next generation of technology. And at the same time I saw
  62. 06:10deep learning as a field taking off. And right around this time Alph Go happened and Alph Go was the first I would say
  63. 06:17major worldwide proof point of super intelligence of a neural network being
  64. 06:22trained to master a very complex board game go at a level that was more
  65. 06:27intelligent than the most capable human player. And I thought there was something really fundamental going on
  66. 06:32here. and I had to understand it effectively inside out. So I actually ended up dropping what I was doing and
  67. 06:39self-eing AI for it must have been four or five months and made some progress there where I started doing some
  68. 06:45independent research that opened up some doors after that. But it was really the realization that AI and in particularly
  69. 06:53deep learning and reinforcement learning were these kinds of building blocks of foundational ingredients of the science
  70. 07:00of our time that will lead to the most impactful technologies in within the next few decades. Like there there are a
  71. 07:07lot of similarities I think between entrepreneurship and research and science. But it is this kind of ability
  72. 07:13to look at a problem that looks really complex and messy and be able to reduce
  73. 07:18it down to some core set of principles that are actually guiding basically the direction of that problem. You know for
  74. 07:24research it could be you know the problem we're interested in let's say is is autonomy. Like we really care about getting these large language models to
  75. 07:31be capable and autonomous. And the question is how do you do that? How do you train them to do this? There are all
  76. 07:36sorts of ways to pursue this question. You can go in many ways, but it turns out there are typically only one or two
  77. 07:42things that really move the needle in a very major way. This sort of a framework that physics gives you for thinking
  78. 07:47about things allows you to one come in with that assumption and rather than looking for hundreds of solutions,
  79. 07:53really try to find the most impactful ones, but then have some rigor in your thinking that allows you to reduce it to
  80. 07:59to those base components. And I'd say that's true for all aspects of company building. There's the research part,
  81. 08:05there's a product part, there's a customer part. And typically in any one of the major buckets of company
  82. 08:11building, there's one or two fundamental problems. And everything else doesn't really matter. And the question is, how
  83. 08:17do you identify those one or two fundamental problems that will move the needle, that will solve your customer's problems, that will be packaged in the
  84. 08:23right way as a product, that will make it easy to use, that will yield the research breakthroughs that you're looking for. I think physics is very
  85. 08:29helpful for thinking about these problems.

  86. Why I left DeepMind and the key lessons from building Gemini

  87. 08:36I think big labs have a lot of things going for them. There's a lot of compute. There are a lot of talented
  88. 08:41people. There are many problems that are suitable I think for solving in in a big lab. Ultimately we thought at the time
  89. 08:48this was you know after launch of Gemini 1 1.5 the paradigm for how people were thinking about things were basically
  90. 08:55building more capable chat bots. What we were deeply interested in since before this was kind of independent of our time
  91. 09:01at Deepmind. This is why we got into AI is the problem of autonomy and we really wanted to work on that and we felt that
  92. 09:06it is both a research problem and a product problem because suppose you build this really great highly capable
  93. 09:14autonomous intelligence. How do you know if it's actually working? How do you know if it's solving people's problems?
  94. 09:19One of the things we believe at reflection is that the evaluation that matters most is the real world
  95. 09:24evaluation. So if you're not working with customers and you're not building product, you're not actually evaluating
  96. 09:29your technology in the place that matters. We just felt that we'd be able to move faster on the research with a
  97. 09:37smaller, more focused team. And we wanted to be coupled very deeply with product and customers to make sure that
  98. 09:44we were steering our research in the right directions. And it's really hard to take a large organization that
  99. 09:50already has a product direction and is it's a big ship that is going in a certain direction. And if you internally
  100. 09:55believe that it should be going in a different direction, it's really hard to change to course correct. It's it's basically impossible. And so this was
  101. 10:02this was the main impetus for for starting as a company rather than doing it in a large lab. There are several
  102. 10:08insights that continue to resonate today from building Gemini and systems before that as well, but I can keep it specific
  103. 10:15to Gemini. One is that the things that tend to work at this level of scale,
  104. 10:20these are giant model. These are it's hard to comprehend how big these models are. Maybe to give a baseline the neural
  105. 10:26networks people were training say 5 years ago were 10 million parameter neural networks 100 million neural
  106. 10:32network was considered huge. Deep came out and it's a over 600 billion parameter neural network. These systems
  107. 10:37are massive. What ended up happening in AI in the before the era of scaling is that sophisticated complex ideas won
  108. 10:45like that you'd take a something small and you'd have like really complex kind of almost like mathematically sophisticated ideas and those seem to
  109. 10:52work and in the era of training these large systems and Gemini in particular
  110. 10:57it's the opposite. The simple ideas implemented at a great level of detail
  111. 11:02are the things that work. So you almost had to kind of flip a switch in your mind about how to approach research problems from adding increasing
  112. 11:09complexity until it works. Like an example of that is IBM blue like the system that beat Garrett Kasparov in
  113. 11:15chess. It was a very complex kind of uh basically treel like structure right
  114. 11:20that elicited all possible moves in chess then picked some best some of the best ones. The opposite is true for
  115. 11:25training these large language models. The objectives are very simple like predicting the next token or the next
  116. 11:30word is a very simple objective. The reinforcement learning algorithms tend to be pretty simple. My co-founder
  117. 11:36Giannis and I led a lot of the work in it's called reinforcement learning from human feedback or RLHF. And if you look
  118. 11:41at public kind of work on this that how other large scale models were trained like let's say Llama or DeepSeek,
  119. 11:48they're very simple algorithms like relative to what reinforcement learning researchers were thinking 5 years ago or
  120. 11:53a decade ago. These are very very simple algorithms. And so maybe that's a thing that stuck with me that doing simple
  121. 12:00things with a great deal of craft and attention to detail and building the right infrastructure to be able to
  122. 12:07support these large models and run them efficiently is probably the biggest takeaway.
  123. 12:13As you just heard from Misha, AI agents are freeing humans to focus on strategy, creativity, and real decision making.
  124. 12:20But here's the challenge for early stage founders. Where do I actually start? We suggest you to check out this AI
  125. 12:27adoption playbook from HubSpot for startups. This is the guide to systemically building real business value with AI without wasting time or
  126. 12:34money and hype. You'll understand how to start using AI without hiring a full tech team and automate repetitive work
  127. 12:40that's slowing your growth. It does this by providing a comprehensive 90-day roadmap that goes through the process
  128. 12:45week by week. We found the sections on every employee gets a chief of staff and the four layer agent architecture
  129. 12:52especially useful. The chief of staff idea shows how to start with personal AI tools before jumping into complex
  130. 12:58integrations and the agent architecture clearly explains what separates successful AI [music] systems from
  131. 13:03random experiment. This ebook was made by HubSpot for Startups which is today's video sponsor. A big shout [music] out
  132. 13:08to them for this free resource. Now back to the video.

  133. Why autonomous coding is the fastest path toward AGI

  134. 13:16There are some things that I have an unconventional maybe opinion on that maybe some other people might not think
  135. 13:21about. I won't speak for other people but I do think that one of the premise that we deeply believe in in reflection
  136. 13:28how fundamentally important the coding problem is and specifically autonomous coding the ability for an AI system to
  137. 13:34autonomously co code something on a computer and kind of go from task to something that's completed and give that
  138. 13:40to the user I think it's a common way to think about such systems is that they're going to be useful for software engineers that makes sense coding is
  139. 13:48that software engineers do a point of view that I have and it's not just myself but our team at reflection is
  140. 13:54that coding is going to transcend software engineering. It's going to go much further beyond that and touch
  141. 14:01basically every other piece of work category of work on a computer. And the reason for that is that when we think
  142. 14:06about how a language model is going to do work on a computer, we have to think about what is its embodiment? What is
  143. 14:12the natural way for a language model to interface with a computer? Effectively, what are its hands and legs? For people,
  144. 14:19we have really strong spatial priors that were evolved through millions of years of evolution. And so we have
  145. 14:24hands. We're dextrous and we have really good kind of innate spatial reasoning. We're born with it. Language models
  146. 14:30don't have that. They were never evolved. They were trained on the internet. And so what's intuitive to us is not intuitive to them. Like they
  147. 14:36don't have the spatial reasoning that we do. But what's intuitive to them is coding. There's a lot of code on the internet. And the same way that we can
  148. 14:44very easily almost trivially without thinking about it reason spatially about objects, language models are that way
  149. 14:49with code, it's just intuitive to them. And so when we think about like what is the way in which a language model interacts with any piece of software in
  150. 14:55the future, likely not going to be by moving a mouse around like humans do and using the human UI, it's probably going
  151. 15:01to be through code. Most piece of software we think will open up these language friendly UIs or interfaces and
  152. 15:07they're going to be mostly programmatic. A thing that we believe that maybe not deeply internalized yet but it is
  153. 15:12definitely internalized by some is that if you solve autonomous coding you solve intelligence on a computer and it
  154. 15:19transcends software engineering. I think as I got started spending more time in AI, I think the ambition of what we'll
  155. 15:27be able to achieve, not me personally necessarily, but as a field and and this is something that personally drives me
  156. 15:32to be a part of, is that we are on the cusp of building a general super intelligence. Even a few years ago, this
  157. 15:39would have sounded like complete science fiction, but this is going to be the most impactful technology of our time.
  158. 15:46And it's hard to say well what the world will look like after it. But it's hard for me to imagine being part of anything
  159. 15:52more impactful or or exciting from a scientific perspective. I think the this is not necessarily even a nice to have.
  160. 15:58This is kind of a a property of this that is pretty remarkable is that it's not just research. It's not just science
  161. 16:04in a vacuum. These systems are really useful today. And you kind of co-develop
  162. 16:09instead of having to wait three decades to see your science have the impact that you're looking for. You're kind of co-developing the science and the
  163. 16:16product together. That's what drives me that the mission of building super intelligence because it is the impactful
  164. 16:21science of our time and the luck I would say that we have it that this research is useful today. My worldview on how
  165. 16:29humans interact with AIs as these systems become super intelligent and start impacting the labor market are
  166. 16:36coming from a position that this is not a zero- sum game. It's not like there's a fixed quantity of labor that either a
  167. 16:43human does or someone else does. Each time there's been a technological advance. It actually just increased the
  168. 16:48amount of things we could produce. With intelligence, the that increase is the amount of ideas and theories and
  169. 16:56experiments and software that you can build. What I think the world looks like a few years from now as these systems
  170. 17:02start becoming extremely capable is that they kind of lift everything up and we
  171. 17:08end up creating in almost every field of computer-based work and then and and in
  172. 17:13the future also physical work. We end up creating an order of magnitude more or even more than that. You know, it's it's
  173. 17:19hard I think it's at least an order of magnitude more than we're capable of creating today. What that means though
  174. 17:25is that today in the same way that we collaborate with colleagues like we have colleagues and we work with teams and
  175. 17:30ambitious projects take big teams to I mean not big in the sense of thousands of people but you need a cohesive team
  176. 17:37to accomplish something big together. I think in the future it'll be that take
  177. 17:42example of an engineer I think a software engineer will become more of a software architect. they have these this
  178. 17:47AI workforce at their disposal. And the same thing will be true for other areas of knowledge work where we kind of
  179. 17:53become architects that manage an AI workforce. The thing that will still be
  180. 17:59really important is asking the right questions because basically if you have a really competent AI system, it will do
  181. 18:06more or less what you ask it to do. The challenge will be how do you pick the right problems to work on? How do you
  182. 18:11pick the right questions? Which by the way that is the whole challenge today with starting a company or pursuing a
  183. 18:16career in research. The fundamental thing to ask is like what is the right problem to solve and then you also have
  184. 18:21to then execute it right. So you have to put in a lot of work to execute it and imagine in the future most the burden
  185. 18:27will be on asking the right questions and designing them projects problems correctly and the execution will be done
  186. 18:35by an AI workforce for you. I think that's roughly the paradigm that we're going into.

  187. A framework for clear thinking and asking the right questions

  188. 18:42I think on asking the right questions maybe I mean it more in the concrete sense of like you suppose you're you
  189. 18:48have a job in creative pursuit and you want to do a good job at it right you have to there's some uncertainty on it
  190. 18:53either you have to build a new product or figure out some research like breakthrough or make a piece of art that
  191. 18:59actually resonates with people right if you're I mean there's one thing of like making art for yourself but if you want to make art for you know that will
  192. 19:04resonate with people that's kind of another thing so how do you know what to pick how do you know you're going to be right that's kind of what I mean around
  193. 19:10asking the right questions. So for example, like in the next year there's going to be one or two breakthroughs in AI. Every year there are basically one
  194. 19:16or two breakthroughs in AI. How do you discover one of them? Like what question should you be asking to discover one of
  195. 19:21them? It's really hard. I mean kind we were talking about this earlier. I don't have an answer to it. But I guess I mean it in in sort of that way kind of if you
  196. 19:28suppose you had these like really super intelligent AIs, you just needed to point them in the right direction. What direction would you point them at? And
  197. 19:34how would you ask them the questions to elicit those behaviors? Like another concrete example I'll give is that
  198. 19:39during reinforcement learning before language models, this is like AlphaGo days. It would take like billions of
  199. 19:44steps for a reinforcement learning training to get like an agent that was competent. And so people are looking at
  200. 19:50and saying, "Wow, billions of steps, that's so long. How do we make it more efficient?" That was the question people were asking. How do we go from billions
  201. 19:56to make it 10x more efficient? So now it's hundreds of millions and 10x more efficient. So it's tens of millions. I
  202. 20:01was asking that question myself and that was the wrong question to ask. But you're kind of saying like the reason you wanted to make it more data
  203. 20:07efficient is because you wanted to get these general agents. And the way people thought about getting general agents is just make them really fast to train. And
  204. 20:13it turned out that the right question was to ask was basically to invent language models because language models without any of this kind of
  205. 20:19reinforcement learning training became very general. If I would have thought about it that way then I would have picked a different research question to
  206. 20:26work on. So right there are plenty of examples where I made the wrong like even if it was locally the correct
  207. 20:31answer. I think there was a question on the list around one of my papers called curl which highly cited paper. It's like
  208. 20:37cited a thousand times and it was an impactful paper locally. So it got citations but it asked fundamentally the
  209. 20:43wrong question which is why it was cited a thousand times and not a 100 thousand times. The people who are asking the right questions write the papers that
  210. 20:50become the like first sentence in every other paper. Like the first sentence of every language model paper is language
  211. 20:56models have become very powerful. Cite GPD4. You know picking the right thing is the hardest thing. So I I definitely
  212. 21:01will not say that will not claim mastery over this. It's something I think about a lot but if I was a consistent picker
  213. 21:08of the right things I would have discovered Imagenet built Alph Go uh basically right built every single
  214. 21:14breakthrough that's come out in AI. So and and you know more widely, right? So picking the right thing is is really
  215. 21:19hard. But some frameworks that at least I use to to think about it is at least
  216. 21:25for me it comes down to clarity of thought. And it's hard to just have like off-the-cuff clarity of thought. You
  217. 21:32need to have some way of formalizing what it is that you're thinking. And at least and for me personally, it's
  218. 21:37writing. Oftentimes when I try to express what it is that I'm trying to achieve or like a method that I'm trying
  219. 21:43to kind of approach, I'll express it in writing, I'll write it down and then kind of almost in like short essay
  220. 21:49format and and revise it because writing often times exposes lack of clarity and
  221. 21:55thought like in when you're writing kind of every sentence should should have meaning and should have a reason for being there. When you do a first pass on
  222. 22:02how you're thinking about a problem and you write it down, you realize how many holes there are in your thinking or
  223. 22:07unnecessary parts of your thinking and you can kind of strip those out and iterate with yourself through a writing
  224. 22:12process. So for me at least, this has come through writing and maybe it's because I used to as as a kid when I was
  225. 22:18really interested in both literature and physics, I spent some time writing short stories. I wouldn't say those are any
  226. 22:24good, but I'll say that the version of the short story that I had written after several iterations was way better than
  227. 22:31the initial version. Making making the right decisions and asking the right questions is a very hard thing, but I
  228. 22:37think that writing is one thing. and then discussing with with other people that you think are very smart and trust
  229. 22:43but in a in a critical way like in a way that you're not kind of looking for someone who will just support your idea
  230. 22:50but you're looking for someone who will challenge find the holes with you. So I think those are the two the two main ways at least for me.

  231. How Reflection AI reached an $8B valuation in just a year

  232. 23:01So for any startup I would say there it's not like there's I think one particular hardest time. I think that
  233. 23:07hard times accumulate over different stages of the company have different hard times and they're probably kind of
  234. 23:12you know all equally hard. But in the very beginning, the things that are hard is you come in and you have a blank slate. And it's sort of reducing that
  235. 23:19blank slate into something that is much more directed and focused and having clarity around that with how it aligns
  236. 23:25with your long-term mission, what you're trying to achieve, but also that short term it's a thing that will work and get
  237. 23:32you to kind of the longerterm objective. Having clarity on that is really important. And it's quite hard uh in the
  238. 23:37beginning of startup to develop that clarity. This is why when people call it something a pivot, it's really, you
  239. 23:43know, a startup took a bet on something that uh did not work and right and so then they developed some clarity and
  240. 23:48then they they pivoted to something. I think that's the first thing that's uh for us and I think for other startups uh
  241. 23:54is sort of the first trial that you go through a startup of figuring out what exactly is it that you're doing today.
  242. 24:00You have your long-term mission. I know what you want to achieve. What are the first steps to that? The second thing is
  243. 24:06how do you get the best people in the world to work on this with you? I mean there's a simple answer that's hard to execute which is the best way to get the
  244. 24:13best people to work with you is by hiring the best people. And uh what I
  245. 24:18mean by this is that it's really hard to build out a stellar team if you don't already have stellar people. Given that
  246. 24:24there's so much uncertainty around startups, people who tend to be attracted to startups are ones that are interested in building something from
  247. 24:29scratch and kind of partaking in the growth and the upside of that. But why would they bet on you versus another
  248. 24:35company at a very early stage? Often times it's, you know, if you have an excellent team that you've assembled,
  249. 24:40even if it's a colonel, five really, really strong people. Good people beget good people. And so it's really important to make the first three hires,
  250. 24:48hire extremely caliber people who you have a great deal of trust with. And so I would say that was a challenge. But
  251. 24:54once that was solved, sort of good people attracted good people. And it has these sort of compounding effects. In
  252. 24:59terms of in terms of motivation, I think it also comes down to sort of what's your long-term strategy and what's your
  253. 25:04short-term strategy and are both of those things um compelling. It's kind of almost like uh in AI there's this idea
  254. 25:10of system one and system two thinking. System two being more highle abstract planning, system one being kind of local
  255. 25:17reactive. And I think you need both of these components to build a company. The nice thing is that when you set a really
  256. 25:23ambitious mission that's exciting, I think building super intelligence is exciting to um a lot of practitioners in
  257. 25:28AI. I mean, that's why I got into it. Like I would join a company that that was trying to solve super intelligence.
  258. 25:33Having a really ambitious mission helps attract really good people. But that's not enough. You have to have clarity on
  259. 25:39what is it that you're going to do today that will get you there. And how is your bet? It's not about the bet being
  260. 25:44different, but why is your bet right? Like why do you think you're correct when others are wrong, right? because the alternative is to stay at a big lab
  261. 25:52that is also pursuing general intelligence and take you know these labs have cast a bet and you can join a
  262. 25:58big lab and ride that bet out. So you have to have good reasons for why you believe your short-term uh wedge into
  263. 26:05the broader mission is compelling like what why is this correct and in our case it's focusing solely on autonomous
  264. 26:11coding um and nothing else and we have reasons to believe why that is the kind of correct bet if you want to aim at the
  265. 26:17problem of super intelligence and so I think those are the kind of the way you motivate people is by doing something
  266. 26:22really ambitious I had a previous startup before this and we did something much smaller and it was actually really
  267. 26:27hard to attract good people to work with us because it was sort of not you know you have one life and people want to
  268. 26:32work on the thing that will be most impactful to them building super intelligence is a pretty I would say
  269. 26:38it's at a similar level ambition of like taking people to Mars right of building rocket ships that go into space and take
  270. 26:44people to Mars and even though that seems really out there and difficult to achieve really talented people are
  271. 26:50attracted to very hard problems and very concrete approaches to solve those
  272. 26:56problems

  273. Three lessons that shaped who I am today

  274. 27:02the way to deal with setbacks. I think I mean there there are basically two main
  275. 27:08things. The first one is to deeply care what you're caring what you're working on. It depends on what inspires you but
  276. 27:14for us right it's kind it's the mission that we're going after and the approach. So we just deeply care and I I deeply
  277. 27:19care and that's really motivating. You have to deeply care about the problem and then you have to deeply care about
  278. 27:24the people who are working with you on the problem. I think with those two things, things that feel that would
  279. 27:30otherwise feel like setbacks, I don't know, don't really feel that way. Maybe it's because researchers have operated
  280. 27:36that their whole career is in uncertainty. That's the whole game is that you pick try to pick the right problem. There's a lot of uncertainty
  281. 27:42around it. There are a lot of setbacks and you persist so long as the problem is really interesting to you and you
  282. 27:47feel like the approach that you're taking is fruitful. Unless you learn, you know, there's some new evidence comes in that maybe you need to change
  283. 27:53your approach, then you need to do that. But then that's not really, if you're deeply interested in the problem, that's not really a setback. That's more of a
  284. 27:59learning. The setback would be doing the wrong thing, doing the incorrect thing forever. That's a setback. But doing the
  285. 28:05incorrect thing at first, then acquiring some evidence. Maybe, you know, from a company building perspective, it might
  286. 28:11be that you have an idea for what a product might look like. you show it to customers and then it turns out that
  287. 28:16they find something else valuable in it that you didn't think about. Some people consider that a setback, but that's actually you want to accelerate your
  288. 28:23time to that event. Like if you don't feel like you have like these sorts of setbacks, then you're probably not
  289. 28:28making progress. And so I think the important thing is to deeply care about what you're working on with the people
  290. 28:34that you're doing it with and be making and sort of have momentum like taking action and making progress. And so when
  291. 28:41I look at the past year, I don't really in in that sense like I I don't really see any setbacks like there were there's
  292. 28:47new information that was learned that change direction for us research and product. There are you know always setbacks in terms of you know maybe
  293. 28:54someone a really good candidate not really wanting to join. But you kind of have to take the aggregate view of like
  294. 28:59what is the sort of general vector that you're going on. Are you learning things constantly? Are you in aggregate hiring
  295. 29:04really good people even if some of them aren't converting? As long as there's that kind of momentum there, I think
  296. 29:10that setbacks don't really affect not necessarily just myself, but I think the setback doesn't feel as painful when you
  297. 29:16have kind of clarity of what you're pursuing. If I was giving advice to my younger
  298. 29:22self or you know I have two younger sisters as well. What advice would you be giving them or you know other people
  299. 29:28calling piece of advice might be sort of around picking the right thing, pursuing your passion like these sorts of things.
  300. 29:34But something that is underappreciated I think is sort of surrounding yourself with the right people. So if you have an
  301. 29:40internal kind of vector of interest you want to do something that like things that are impactful interesting to you.
  302. 29:46But so long as you have that in the same way that it's important in your personal life to surround yourself with very highquality friends. I think the most
  303. 29:53kind of when I when I look at how sort of my last decade has played out, the
  304. 29:58thing that has been most impactful to me was surrounding myself with right people who at that time maybe would have taken
  305. 30:04a chance on me like for example in in Berkeley I would say Peter Aiel when took me in as a postto lab I was a
  306. 30:11physicist I was not an AI person and AI was very competitive then but by being in that lab and surrounding myself with
  307. 30:16people like him and his PhD students that's what really enabled me to learn quickly and develop my thinking. I think
  308. 30:24often times a function of your what you're able to achieve is really who are the people who you're spending your time
  309. 30:31with. Very talented, ambitious people are also generally quite open to I think
  310. 30:36giving back. Now, it's of course hard to get in front of them and so it's not like I think just sending a cold email
  311. 30:41is not enough. You have to really demonstrate that you really want something badly and demonstrate it
  312. 30:47through not just words but actions. In in my case, it was I spent a few months and I went and did a research. I taught
  313. 30:54myself reinforcement learning. I did a research project. I had something clear to kind of concrete to bring to the
  314. 30:59table and get people's feedback on. But I think so long as you have that as like there's you're persistent, you're able
  315. 31:05to kind of show your desire to work on something through action and not words, that's a very rare thing for a person to
  316. 31:12do. People who have been successful in whatever industry that might care about, I think, will look positively on that.
  317. 31:18So there's a sort of I think you can get into almost any door that you want with sufficient effort and it's just really
  318. 31:24important to surround yourself with the right people to enable you.