The Breakthroughs Needed for AGI Have Already Been Made: OpenAI Former Research Head Bob McGrew

As OpenAI's former Head of Research, Bob McGrew witnessed the company's evolution from GPT-3’s breakthrough to today's reasoning models. He argues that there are three legs of the stool for AGI—Transformers, scaled pre-training, and reasoning—and that the fundamentals that will shape the next decade-plus are already in place. He thinks 2025 will be defined by reasoning while pre-training hits diminishing returns. Bob discusses why the agent economy will price services at compute costs due to near-infinite supply, fundamentally disrupting industries like law and medicine, and how his children use ChatGPT to spark curiosity and agency. From robotics breakthroughs to managing brilliant researchers, Bob offers a unique perspective on AI’s trajectory and where startups can still find defensible opportunities. Hosted by: Stephanie Zhan and Sonya Huang, Sequoia Capital Mentioned in this episode: Solving Rubik’s Cube with a robot hand: OpenAI’s original robotics research Computer Use and Operator: Anthropic and OpenAI reasoning breakthroughs that originated with OpenAi researchers Skild and Physical Intelligence: Robotics-oriented companies Bob sees as well-positioned now Distyl: AI company founded by ex-Palintir alums to create enterprise workflows driven by proprietary data Member of the technical staff: Title at OpenAI designed to break down barriers between AI researchers and engineers Howie.ai: Scheduling app that Bob uses

• 0:00Speaker 00

  I think what's really changed is that now that you have LLMs, you have this language interface to the robot so that now you can describe the tasks much more cheaply. And you have really strong vision encoders that are tied into that intelligence. So that gives the robots really a headshot at doing generic tasks. So we spent years solving one specific problem, teaching a robot to manipulate a Rubik's Cube. And now a company like, let's say, Physical Intelligence can spend months solving a huge variety of problems like laundry folding and cardboard and packing egg crates. And that's something that they can only have because they're building on top of existing frontier models

• 0:44Speaker 00

  and the entire tech and research stack that we've built over the last 10 years.

• 1:05Speaker 01

  Welcome to Training Data. Today, we're excited to welcome Bob McGrew, former chief research officer at OpenAI for a fascinating look behind the scenes of frontier AI development. Bob talks about the trifecta we have in AI: pre-training, post-training, and reasoning, and explains why we may have already discovered all the fundamental concepts needed for AGI. You'll learn why he thinks agents will be priced at the cost of compute, hence eroding traditional economic moats, and why even proprietary data will become less valuable when infinitely patient AI can recreate alternatives. Plus, Bob shares his contrarian take on where startup opportunities really lie and why robotics is finally having its moment after years of being too early.

• 1:50Speaker 01

  Enjoy the show.

• 1:53Speaker 02

  Bob, thank you so much for joining us today.

• 1:55Speaker 00

  Oh, it's great to be here.

• 1:56Speaker 02

  We're at a really interesting time in AI development. We have a beautiful new trifecta: pre-training, post-training, reasoning. Can you help us unpack what else is left? What alpha is there in each left?

• 2:07Speaker 00

  So I think we're going to continue to see capabilities increase. It's going to continue to feel like it's felt super fast, super exciting over the last even five years. And I think it's going to continue feeling like that. There's not a wall here. But what is going to be different is that 2025 is going to be the year of reasoning. So it makes a lot of sense. Reasoning is a new technique. When you have a new technique, you know, there's often an overhang of compute, of data, of algorithmic efficiency improvements that you can make. And so that's something that, you know, if you, if you look at just the incredible progress that we saw from 01 preview back in September, and then six months later, we go to 03 in April.

• 2:51Speaker 00

  And then, you know, at the same time, we also see diffusion of reasoning from OpenAI, who we've been working on that for years, out to Google, DeepSeek, Anthropic, again, just in a few months. And so this is really the, the, the right place where every lab is going to focus for the year. And just, just like a sort of a fun example of how low hanging the fruit is right now. If you look at the most interesting difference between 01 preview and 03, 01 preview is not able to use tools. 03 can use tools as part of the chain of thought. And this, this is pretty obvious, right? When we were training 01, we knew that this was a thing that we wanted to do, but it was difficult to implement. It took time.

• 3:32Speaker 00

  And so, you know, that took six months to get, to get, you know, done and released. The next step on reasoning is going to be a lot less obvious than that. It's going to be a lot harder. And so, you know, as reasoning continues to mature, we're going to see the overhang get eaten up, and it's going to start being slower and slower to make progress.

• 3:51Speaker 02

  You said there's not a wall. I think there's this meme in the Twitter sphere right now that pre-training is hitting a wall. Can you say more about that dynamic?

• 3:58Speaker 00

  Yeah, and I, and I think that's great. That, that's, that's a great question here because pre-training is not going away. But what we're seeing out of pre-training is that we are at the place where it's working really well, and we're hitting diminishing returns. And so, you know, Diminishing returns are baked in because the intelligence of a model is log linear in the amount of compute that you're using to train it, which means that you have to have exponential increases in compute to get each increment in intelligence. When you pre-train a model, that's a giant training, it takes all of your data center for a period of months. And when you go to pre-train the next model, you can't really do it on the same data center. You can rely a little bit on algorithmic efficiency to make it better,

• 4:41Speaker 00

  but fundamentally, you have to wait until you get a new data center. And that's not a matter of like, something you can do in six months, the way you can do improvements in reasoning right now. That's something that takes years. So that doesn't mean pre-training is useless, though, because the real lever for pre-training in 2025 is improving architectures. So even though you're working on reasoning, you want to improve pre-training so that you can have a better inference time efficiency, or so that you can have longer context, or better use of the context. And when you're doing that, you have to start back from the beginning, do pre-training on this new architecture, and then go through the whole reasoning process again. So that's the role of pre-training now. It's still important, it's just doing something different in the pipeline.

• 5:23Speaker 02

  Can you help us unpack what's left in post-training?

• 5:25Speaker 00

  Yeah. So post-training is pretty interesting, because both pre-training and reasoning are about increasing intelligence. And there's a very clear scaling law that you get, where you put in more compute and you get out increasing intelligence. And post-training isn't like that. Post-training is about model personality. And intelligence, it's a sort of a thin problem, right? If you can get better at it, it turns out to be very generalizable, and it applies to everything. So you can work on math, and you find that it makes you better at legal reasoning. But model personality, it's a thick problem. You actually need a lot of human effort to think about what makes a good personality? How do I want this agent to act? And it's much more of a training process, like you would go through over many years of interacting with people.

• 6:10Speaker 00

  And now, it's a very hard research problem to take... That specification for what the agent is and turn it into an actual appealing personality. But when you think about post-training, I think about people like Joanne Jiang at OpenAI or Amanda Askel at Anthropic, who really spend a lot of time crafting these model personalities, and they're not research practitioners.

• 6:32Speaker 02

  Interesting.

• 6:33Speaker 00

  Right? They're people with their product managers or they're people with a very deep understanding of human nature.

• 6:38Speaker 02

  And are there more legs to the stool?

• 6:40Speaker 00

  Well, OK. So I'm gonna say something potentially controversial, and I think actually there aren't. So I think if you go forward to 2030, or if you go forward to 2035, and you look back and you say, what were the fundamental concepts that you needed in order to create more and more intelligence? Maybe that's AGI, maybe it's something different at that point. I think you're gonna come up with the idea of language models with transformers, the idea of scaling the pre-training on those language models, so GPT-1 and GPT-2, basically, and then the idea of reasoning. And sort of woven throughout that, increasing more and more multimodal capabilities.

• 7:22Speaker 00

  And I think even in 2035, we're not gonna see any new trends beyond those. And the reason I think this is if you go back to 2020, so GPT-3 has just been trained. Imagine yourself sitting where at OpenAI, we haven't released this thing, but we know something, apocalypse has happened, and Dari Emadai, Ilya Sutskever, Alec Radford, we're all sitting there in the room, looking at this thing. And it was fairly obvious internally what the roadmap was. We knew that at this point, going from GPT-3 to GPT-4 by increasing pre-training was absolutely critical. We could see that we needed to increase multimodality, ultimately ending in a model that could use a computer.

• 8:05Speaker 00

  We were beginning to make experiments with test time

• 8:10Speaker 00

  compute. And in 2021, after the Anthropic people left, we really started developing the idea of reasoning at OpenAI. And it's funny, actually, if you if sometimes my friends asked me after Anthropic released computer use, they're like, did you see that coming? And I was like, well, we were working on that together back before they left. One of the people that did that project went to Anthropic and the other one went to OpenAI and developed operator. And it just took many years before, you know, the multimodality had matured enough to get to that point that was obvious to us way back then. And so that's why I think, you know, we're going to see from here on out. There's very important scaling. There's very important development and refinement of these ideas.

• 8:55Speaker 00

  That is extremely hard. It takes a lot of brain power. It's not going to be easy. But I think if we look back from 2035, we're not going to see anything new and fundamental. I think I'm right. I kind of hope I'm wrong. It would be a lot more fun if I'm wrong, but I think we'll have to see.

• 9:10Speaker 03

  That's a hot take. I'm glad we have it on the record. Yeah. Well, we'll see in 2035. That's amazing. I'm curious about reasoning. Like, it seems to me that OpenAI really leaned big behind this paradigm, probably before the others. And, you know, now everybody has reasoning models. What did you see in reasoning that caused you to lean so far in so quickly?

• 9:31Speaker 00

  Well, I mean, effectively, it really was sort of this missing piece where with pre-training, the model has an intuitive sense of how to answer the question. But if I ask you to multiply two five-digit numbers, this is something that's completely within your capability. But if I asked you to do it right now, you wouldn't be able to do it. Because it is just natural as a human capability to be able to think about something before we answer, to have a scratch pad, to be able to work through a problem. And that is something that, you know, the initial models, even GPT-3, really didn't have. And so we began to see, you know, glimmers of this publicly, things like, you know, thinking step by step. And the idea of having a chain of thought that you could train,

• 10:15Speaker 00

  that the model would learn itself how to guide a chain of thought, not just be guided by cloning from publicly available data on how humans think. That was very powerful. And we knew that it would be more powerful than pre-training because in fact, your thoughts are inside your head. They are not something that the model has access to. And so almost all the data that's out there is actually something that's just the final process. But you don't get to see that chain of thought. And so the model had to figure it out for itself. That's why reasoning mattered.

• 10:46Speaker 02

  You alluded to earlier that we probably still have to uncover more things in reasoning. Do you think we have a good sense of what those things are today? Or are we earlier in that R and D stage?

• 10:57Speaker 00

  I think at this point with reasoning, if you are at the coalface, then you're seeing a lot of ideas and refinements of things that you can do. I think we've gone past the point where if you're, if you're on the outside, if you're not at a frontier lab, you're probably not seeing them anymore. And you know, this is the same situation we saw where at one point academic labs could make huge amounts of progress. And then later, you know, I would begin to see academic papers and I think, oh, they rediscovered this thing that we found a long time ago. And so, uh, you know, now the level of effort that's being put into this, I think is actually quite intense. So there, there are definitely things to be discovered, but they're not sort of simple ideas that you and I could talk about.

• 11:42Speaker 02

  Cool. Switching gears a little bit, you tweeted recently about agents. Uh, I think a very, very interesting take that agents will be incredibly powerful, but priced at the cost of compute due to competition. If that's the case, where do you see the opportunities in new startups and companies that are now building agents?

• 12:00Speaker 00

  Yeah. So, I mean, I think that the, the, the thing about agents is people think, well, you know, I'm going to go develop an agent and they look at how much the job is worth out there by a human. So, you know, you want to develop an AI lawyer and you think lawyers get paid a lot of money. So I'll be able to develop an AI lawyer and the lawyers, I'm going to be able to charge huge amounts of money.

• 12:19Speaker 02

  Tens of a thousand dollars.

• 12:20Speaker 00

  Exactly, exactly. Right? But the reason lawyers are expensive is because their time is scarce. Because there's only so many people who have undergone that training. But by the time you've made an AI model out of it, well, now there's effectively an infinite number of lawyers. And so it's not scarce at all. And maybe you, with your AI lawyer startup, will be able to have a lead over other people. But it's the same frontier model underneath. And, you know, some other startup can come in and compete that away. And so we should expect to see it priced at some opportunity cost over the cost of compute.

• 12:56Speaker 02

  Interesting. Because you're changing. You now have a lot more supply, infinite supply of the highest capability intelligence in whatever domain you now have.

• 13:05Speaker 00

  And on the one hand, there's a story where this is, oh, this is bad because startups can't make money. But this is actually the future we want, right? We want services that don't require people to be extremely cheap. You want everyone to have access to a lawyer. What you want to be expensive and scarce are things that are actually about personal relationships. So, you know, maybe we won't be asking the human lawyers to write contracts because agents will be doing that for us. But we'll be asking them for sort of deep advice on, you know, how legal challenges affect the, you know, detailed challenges that I'm facing in my business. And I think, I think that's the world we want to live in.

• 13:43Speaker 03

  You think application companies will make any money selling agents, though? Like, where would you tell us to invest?

• 13:48Speaker 00

  Yes and no. So just to, to back up for a second, people often talk about where does the value accrue in the stack, right? Is it at the model layer? Is it at the application layer? And if you look at the model layer, it's very competitive. Every company has a frontier model. Some of the frontier models can do things other frontier models can't. But by and large, they're all really very good. And if you're in enterprise, you can swap them out very easily. And beyond the frontier, you know, all the models that are answering the bulk of the questions are distilled. They're very competitive. And so this isn't a very good business to be in when you consider the cost of training the models. So what's the point of training models in the first place? It's to give you an option. It's to give the frontier labs an option on the valuable places in the application layer that are coming up.

• 14:33Speaker 00

  So, you know, ChatGPT, that's a great business, right? There's a lot of competition over that. I think probably it's too late to replace ChatGPT. Maybe not. You'd have to do something very different.

• 14:46Speaker 00

  Coding, another place where all the frontier labs are eager to see right now. I think you can compete with the frontier labs, but you want to do something that's different, something that involves more than just, you know, you talking to your computer, you doing some sort of personal productivity task on your computer, something that involves other people, something that involves an enterprise. I think that's, you know, I think that the moats that you have for your business are going to be the same moats they always were. It'll be network effects, brand, economies of scale. And so you want to find an agent that allows you to have those network effects, not just something that, you know, would be high priced out in the world.

• 15:26Speaker 02

  Are there particular domains that you think are maybe outside of the scope of what frontier labs want to innovate in and build in that you think are interesting and that you've been mulling about? We've got, you know, scientists, lawyers, research analysts, agentic software engineers. What other domains have you been thinking about?

• 15:46Speaker 00

  I am, so personally, I'm very interested in robotics because I think robotics is something, I wouldn't actually say it's off the roadmap of the frontier labs right now, but I think it's something that's far enough away that, you know, to me, it feels like where AI was a few years ago. And so I think this is a very good time to be a company like skilled or a company like physical intelligence or, you know, to start a new robotics company, maybe not one that's competing with those two, but something that's doing, you know, something different, something on its own. I think it's, you know, at the end stages of being a research challenge and, you know, a matter of, you know, months or years, small digit years away from being commercialized. So I think that's really fun.

• 16:26Speaker 02

  Why now? What do you think has changed? Like OpenAI famously had a robotics effort for a long time. What do you think has changed?

• 16:31Speaker 00

  Well, you know, so in between Palantir and OpenAI, I actually wanted to start a robotics company myself. And I got to the point of teaching a robot to play checkers from Vision back in 2016.

• 16:44Speaker 02

  Wow.

• 16:45Speaker 00

  Yeah, it was very, very cool.

• 16:47Speaker 02

  The checkers pieces?

• 16:48Speaker 00

  It could pick up the checkers pieces and it could move them to a different place on the board.

• 16:51Speaker 02

  Nice.

• 16:52Speaker 00

  And my conclusion from this was that it was very fun and super cool and really far away from any form of commercialization.

• 17:01Speaker 00

  And when we pursued robotics at OpenAI, we didn't pursue it for commercial motives. It was really a demonstration of the power of machine learning. And some of the ideas we had there later played into large language models. But I think what's really changed is that now that you have LLMs, you have this language interface to the robot so that now you can describe the tasks much more cheaply. And you have really strong vision encoders, you know, that are tied into that intelligence. So that gives the robots really a headshot at doing generic tasks. So we spent years solving one specific problem, teaching a robot to manipulate a Rubik's Cube.

• 17:44Speaker 00

  And now a company like, let's say, Physical Intelligence can spend months solving a huge variety of problems like laundry folding and cardboard and packing egg crates. And that's something that they can only have because they're building on top of existing frontier models. And, you know, the entire tech and research stack that we've built over the last 10 years.

• 18:04Speaker 03

  Yeah, I want to go back to this point you had on, you know, where is the value? And I really liked your framing that the foundation models kind of have an option on whichever parts of the application stack they want to own. How much of the application market do you think the foundation models will win?

• 18:19Speaker 00

  I think I would look at this like in a slightly different direction, which is if you're a startup, you know, where is it safe to play? And where is it that you're going to get steamrolled by the frontier labs? And so I think the areas that I think are safe to play in are areas where you have to understand something very deeply outside the model. And so I think a lot of enterprise really has this flavor. So for example, you know, Palantir AIP actually really fits this where it's, you know, it's not a model company, but it's something that sits outside the model that interacts with the rest of the business. There's another company I'm an investor in and advisor in called Distill

• 19:01Speaker 00

  that builds AI systems that allow you, a business, to sort of extract the context from within the business, feed that to the models. And then use that to make decisions. And so, you know, these are things that the frontier labs don't want to do. The frontier labs see business problems as how do I train a model to do something new?

• 19:21Speaker 02

  Yep.

• 19:22Speaker 00

  And if you look at all these enterprises, each one of those is a very small problem. It's not worth OpenAI or Anthropic's time to train a model specifically for each one of them. If you put the problem, you think about what is the system that goes around the models? And how do I use the models to sort of ease the context in and, you know, get the outputs out? Then suddenly that's one problem, and I think it's a big opportunity.

• 19:44Speaker 02

  What are the specific use cases and problems that distill Palantir's effort solve for those enterprise companies?

• 19:52Speaker 00

  So, a lot of times right now, what you see is you're trying to automate some existing piece of work. And the easy cases are where that piece of work is in a regulated industry. And, you know, you're working on something like healthcare, maybe you're interacting with insurance companies. And you have a workflow that is extremely scripted, where the company cares a lot about fidelity to that workflow. And that doesn't mean you can just say, hey, you know, AI, go read the clinical guidelines and, you know, make these decisions. But with a process of transformation, you can get it to the point where the AI can do that. And that's sort of the low hanging fruit. And then the next level up, though, is that if, imagine you're working on something that

• 20:35Speaker 00

  isn't a regulated industry, or that isn't extremely scripted, and you want to ask someone, you want to automate some labor intensive process, well, the first thing you have to do is make that legible. And if you go to someone and you ask them to describe their job, a lot of times, you know, their manager doesn't know what they do, they don't even really know what they do, they can give you examples, but they can't say like, this is the workflow that I follow. Because in practice, they don't follow a single workflow. Right? And so I think that is what a lot of these problems look like. And you know, and for example, that's actually what Distill does is to, you know, work with companies, help them, you know, take the data they have, interview the people

• 21:16Speaker 00

  with AI, systematize all of that, and, you know, have it be something that an AI model can actually execute.

• 21:23Speaker 02

  That's really interesting. Huh. So this is also somewhat related to this other question I wanted to ask you about proprietary data. I was surprised to actually see you tweet this, but I very intrigued by this question that you posed, which was, how valuable will your proprietary data be compared to what your competitors infinitely smart, infinitely patient agents can estimate from public data? Can you unpack that for us a little bit?

• 21:48Speaker 00

  Yeah, so, you know, a starting point for this is, a few years ago, there was a lot of interest in training industry vertical-specific models, you know, that finance companies would say, we've got all of this data that no one else has, and we're going to train a finance model on top of GPTs or on top of Llama, and it's going to be so much better. And actually, all of those were worse than the next generation of GPT, because the power of intelligence and the ability to synthesize new information was bigger than the power of sort of memorizing the old information that you have. So that's, I think, what this theme looked like a couple of years ago. But, you know, fast forward a year or two years, you know, now the story is, I have

• 22:28Speaker 00

  all of this proprietary data, I've accumulated it over years, and in some sense, for a lot of times, you know, if that data is teaching the model a skill, or if it's meant to teach them all the skill, that data is sort of embodied labor, right? Someone worked through all these case studies, someone called all of these customers and found out all of this information. Well, that embodied labor is now free. AI can do all those things. And so now, there's an opportunity, you can have AI call all those customers, do a big survey, find out what they know. You can have AI work through all the case studies, just a lot of chats with O3, right? And then now, you've replicated that proprietary data.

• 23:10Speaker 00

  But without needing all of that work.

• 23:14Speaker 02

  How do you square that with the value of real world proprietary data? Say something like what Cursor gets from its developer community constantly? Or Tesla Autopilot over the last handful of years?

• 23:28Speaker 00

  So I think those are in the middle because they're really huge amounts of data. I think there are challenges sometimes to training on the data that you get from your users. A lot of time, you know, one thing models can't do is if you train data and you memorize data about a specific person, maybe that leaks out into the next person. So that's a real challenge to using these kinds of proprietary data. I think there is a kind of real world proprietary data that's very useful, which is data, very specific data about very specific customers that they trust you to use on their behalf. So to give an example, my financial advisor knows a lot about me. She knows my entire portfolio. She knows the kinds of objectives that I have.

• 24:10Speaker 02

  Your risk tolerance.

• 24:10Speaker 00

  My risk tolerance, right. And she uses all of that information to give me a better outcome, which is what is the next asset I should buy? And she doesn't do that, like the data doesn't make her a better financial advisor. It doesn't teach her skills, but it allows her an opportunity to use the skills she already has. And so that's the place where I think proprietary data is really useful.

• 24:29Speaker 03

  I want to switch gears a bit to coding. It feels like, you know, software engineering has just gone through this fast takeoff moment. And, you know, just judging from the pace of how quickly things are changing, you know, I think there's at least a certain subset of the market that thinks, you know, the super intelligence takeoff probability is a lot higher than it used, than than folks thought it was, just given how quickly coding has taken off. What's your view of what's happened in the coding space?

• 24:56Speaker 00

  So I think, you know, on the one hand, coding has taken off very quickly. On the other hand, way back in January 2020, as soon as we saw GPT-3, we launched a project to train GPT-3 how to code. And so, you know, when you look at an exponential curve, you know, the progress is actually the same the whole time. But the impact of that progress can become very nonlinear when it passes a threshold. And that's what's happened with coding in the last couple of years.

• 25:26Speaker 00

  And so, you know, my, my take on where coding will go is that you're going to continue to see a mix of, you know, coding with the user in an IDE, you know, traditional cursor style work, and coding in the background as an agent, you know, something like Devon style work. And that, you know, this is going to continue for a long time. It as, you know, a year or two, maybe is a long time in AI adoption.

• 25:51Speaker 03

  That is, that's forever in AI years.

• 25:54Speaker 00

  But that, um, you know, if you think about something like vibe coding, right? Like the, the story you hear with vibe coding is you can, you know, if you have a PM, right? And you want to create a demonstration project. I think that you're going to see PMs vibe coding, really cool prototypes, really cool demos that they can use to get user feedback. But then those things are going to get thrown away. And that they're going to get rebuilt with professional software engineers. Because, you know, if you are given a codebase that you don't understand, this is a classic software engineering question, is that a liability or is it an asset? Right? And the, the classic answer is that it's a liability. Like you have to maintain this thing. You don't know how it works. No one knows how it works. That's terrible. Usually the answer is it's actually cheaper to rewrite it from scratch.

• 26:35Speaker 00

  And so we don't yet have a way that we're comfortable with agents being the ones that understand the codebase right now. I think, I think the liability has gone down, but it's still net a liability. You need humans to do the design, to understand the codebase at a high level. So that when something breaks, when the project itself becomes too complicated for the AI to understand, you can have a human do a problem decomposition and break it down into problems that are small enough for the AI.

• 27:04Speaker 03

  What do you think happens after that one or two years, though?

• 27:07Speaker 00

  Oh, I don't know. We're gonna have to find out.

• 27:09Speaker 02

  I love your bifurcation, though, of, you know, on one side, agentic software engineers that handle tasks autonomously in the background, and on the other side, human programmers who code in an IDE with the help of AI. I don't think that most of the mainstream actually believe that, realize that. Can you maybe unpack that for us a little bit? What would the agentic software engineers who handle these tasks autonomously handle? And then where do you see this other end of the spectrum go? Do they collide at some point? Do you think they remain separate things over time in the long term?

• 27:43Speaker 00

  I think it is already a spectrum, where, you know, the things that your agentic software engineers can do, you can say, well, fix a bug, do a refactor, something that, you know, requires relatively little taste and has a clear outcome. Another great use case I've heard is translate software from COBOL into Python, right? It's very clear when you've done this correctly, but it's a lot of work. It's very boring, and you can't get smart people who want to work on this and do a good job on it. On the flip side, if you're doing something where it requires a lot of taste, and taste in how it's implemented, where there will be non-obvious consequences to how the implementation works.

• 28:27Speaker 00

  Maybe there's non-obvious performance consequences. Maybe there's non-obvious consequences in how the user interface is going to evolve, and therefore how that needs to change the abstractions deeper in the system. Those are places where, right now, we have no alternative but to have humans do that work. And I do think this is very interesting. Is there a way, you know, is there a sufficiently detailed spec, or a sufficiently detailed architecture diagram, that the agents can be writing for us? That means that when you take work from one agent, and you put it into another agent, which could just be the same agent the next day with a different context window, that it's able to actually make progress on the codebase. So, these are the kinds of questions I want to see the answers to over the next couple of years.

• 29:11Speaker 02

  Love it. It's exactly what we're working on at Reflection.

• 29:13Speaker 00

  Perfect.

• 29:14Speaker 03

  Why is it called member of the technical staff?

• 29:16Speaker 00

  That's a great, yeah, that's a great question. So... For a long time, this was true even before I joined OpenAI, by the way. I believe this was Greg Brockman's idea. But we really wanted there not to be a distinction between engineers and researchers. If you look at a classic lab, a place like Google Brain, for example, where a lot of the people who started OpenAI came from. At the time, and maybe still today, there was a big differentiation between whether you had a PhD and you were a researcher, or whether you were a software engineer and you had, you did data, you did implementation. And it was bad because the researchers didn't feel like they could get their hands dirty

• 30:00Speaker 00

  writing data code or writing implementation code. And you can't understand the systems aspects of your research unless you're writing the code. If you think about what makes Alec Radford the genius researcher that he is, it's each time he does something, it's that he looked very closely at the data and he thought, what are the possibilities of this data? He wrote his own data scraping code from the very beginning. And so if you want to have someone who really understands the full stack, I think Paul Graham has this great analogy to painting, where the resistance of the medium dictates the kind of painting that you're able to make. Research is very much like that. It's very much an artistic endeavor, and researchers themselves are artists and should act like artists.

• 30:43Speaker 00

  And so by not having that to think, just by calling everyone member of the technical staff, we were able to have a much more level playing field. And later, that really came in handy when we had people who didn't have PhDs. Many of the great researchers at OpenAI, you know, Aditya Ramash, Alec Radford, you know, many of these people don't have PhDs and in fact, learned their trade by working at OpenAI.

• 31:06Speaker 03

  That's a great answer. It's a random throwaway question. I love that answer.

• 31:11Speaker 02

  So what AI sent recently, Sam Altman left us with some interesting fodder, which was how different generations use ChatGPT. He said, if you're old, you tend to use it as a Google replacement. If you're in your twenties and thirties, you use ChatGPT as a life coach or a life advisor. And if you're in high school or younger, then you're using it as your operating system. How do you see people use ChatGPT around you? How do you have your kids use ChatGPT?

• 31:38Speaker 00

  Yeah, so, look, let, let, think about that operating system comment for a second. At the very highest level. The total addressable market for ChatGPT is every user intent that requires thought or action that you don't want to do yourself. Anything that you wish got done, but you didn't have to do, is something that you might want to use AI for. And so, there's, I mean, if you think about that, there's a version of that that feels very scary, right? It's like people don't do anything for themselves anymore. There's a de-skilling. No one learns how to do hard things. We're all just zombies, watching our VR headsets, you know, watching movies. But I don't think that's actually what people want out of AI. And I don't mean this, this isn't the world we want to live in, I think that's true.

• 32:23Speaker 00

  But that's not what I want out of my relationship with AI. And that's not what I see people doing now. And partly, this is because the technology for ChatGPT as an operating system isn't actually there yet. Pretty famously, you cannot use ChatGPT to control your iPhone. But it's also not what people want. And so, I see this with my son. He's eight years old. He's been using ChatGPT from a pretty young age. I used to ask him to test the models before they were publicly released. And he always gave pretty good feedback, actually. And he spends a lot of time with ChatGPT. He knows it is not his friend. It is not his companion. It is an expert, someone he can talk to, to explain things to him.

• 33:06Speaker 00

  And if you are eight years old, having someone who can explain things to you correctly, in great detail, and with a lot of patience, is a very valuable thing. And so, he has, he has like curiosity, he has enthusiasms. And one day, he decided he wanted to be a coin collector. And so, he collected all the coins in the house, sorted through all the ones that were from before 1970, went to ChatGPT, started typing, and just asked, took pictures and just asked questions about every single one of the coins from before 1970. And he's, you know, what's this worth? Well, what would make this worth more? You know, how can I test? What, what is a mint mark? You know, all these different questions. And if you think about this, this is something, you know, when I was a kid, I, I probably

• 33:49Speaker 00

  could have learned this, like, maybe there were books, there were magazines, maybe I could have looked in an encyclopedia. But all of this is just so accessible now. And it's accessible to an eight-year-old. And so, um, when we went on vacation, we took him to a coin shop. And the staff at the coin shop were just shocked at how much this eight-year-old knew. And the, the very detailed question, he was, "Show me all your coins." "No, I don't want that one. I want one that has a San Francisco mint mark. I want one from this year. This is the year that they were all made out of silver." And the, the coin shop owner was just very surprised. He doesn't deal with kids that have that level of detail, at least not until now.

• 34:30Speaker 00

  And so, this is, I think, actually what we want out of AI, is that AI should make you an expert at the things you want to do, and it should remove the burden of doing the things, that, the boring things, that you don't want to have to do.

• 34:44Speaker 02

  Yeah. On the topic of the next generation, how else are you preparing that next generation for all the capabilities to come in AI?

• 34:53Speaker 00

  Oh, I think this is a super, super tough question. The, uh, if you think about any particular field, you know, should you teach your son how to code, right? Like, think about my eight-year-old, you know, um, you know, my daughter is writing essays. My eldest son is really excited about math. All of those things are going to be automated. And so, it's clearly not some specific skill that you have to teach them. I think there's really two things that I want my kids to understand. The first is the process of learning and figuring things out. So that's the value in the math and the essay writing and the coding. It's, it's sort of this, this process of being able to, of learning to learn. The second thing is the, you know, having ideas and projects,

• 35:37Speaker 00

  and the belief that you can do it, and the ability to use whatever tools are at your disposal to figure it out. So this is, this is agency, right? And so, that's where, I think that's the right way to have kids use AI right now. And there's, there's always a trade-off. I'm, I'm often very torn between, so I, my eight-year-old uses, uh, ChatGPT for a lot of things, but I don't let him use it to code, because he's trying to learn to code. And if he sees that he doesn't have to use it to code, then it's going to be very hard for him to do that work, to get all the way there. I don't let my other kids use it to do their school assignments, of course. Why would you do that? Um, but-

• 36:16Speaker 00

  I want them to have those basics. And then, once they have the basics, once they understand things one level down, to then be able to use it to extend their capabilities. And, you know, here's another fun story about my eight-year-old. Last week, he decided he wanted to build a project where the grandparents who are coming to visit could press a button and he could go, it would ring a buzzer in a different room, and he could go get breakfast in bed. And he asked ChatGPT for help. He, I mean, he asked ChatGPT for help. It said, okay, you need jumper wires, you need, you know, two Arduino boards, and, you know, just a sort of list of things. And he asked a lot of questions, you know, how is this gonna work? He asked it to give us a list of Amazon links for us to buy.

• 37:01Speaker 02

  Wow.

• 37:02Speaker 00

  I reviewed this, made sure he wouldn't get electrocuted, bought the items on Amazon for him. And now we're putting it together. And my approach in this is I'm gonna let him put it together, everything he can. I'm gonna install the software, because, you know, his computer's locked down, he can't install software. And this is gonna be his project.

• 37:23Speaker 02

  That's amazing.

• 37:24Speaker 00

  Who could have, none of us could have done that at eight years old. And he has learned so much in doing this. It's not just that he outsourced it all to ChatGPT. Now he understands what Arduino is, he understands what the circuit boards, you know, what is going, what happens when I hit this pin? Why is this pin named, you know, GRP1? You know, these are, these are all, I mean, I don't know the answers to these things either. So it's really, you know, it's just this huge help that ChatGPT is able to do all these things for him.

• 37:50Speaker 02

  That's amazing. Sparking curiosity and then agency, I love it. And it's also just the time to impact. And that just feeds more and more curiosity and agency.

• 38:00Speaker 00

  Yeah, that's right. I mean, if, if, if you think back, you know, well, you wanna do this project, well, here's a book on Arduino, and, you know, you're gonna have to write the code yourself, and, you know...

• 38:12Speaker 00

  What circuit boards am I supposed to do? I don't even know how to do that.

• 38:16Speaker 02

  Yeah.

• 38:16Speaker 00

  You know, probably this project just dies on the vine. And you know, there's a, there's a truism in education theory that when someone asks a question, that's the time when they're ready to learn the thing that they're asking the question about. And so you want to, you know, it's, it's worth going off script to answer someone's question because you're, you're doing a huge service to them in teaching them that thing right then. And that's, you know, now you have that, you have the ability to get your questions answered on demand at the right time for you, when you are mentally ready to do it. Not when maybe you're tired and you're in school and you're thinking about all sorts of other things, just right then, when you actually want to know the answer. And I think that's hugely powerful.

• 38:57Speaker 02

  So how else are you using AI in your daily life? Chat GPT, Deep Research, I'm sure, Howie AI for scheduling, maybe autopilot. What else?

• 39:08Speaker 00

  So, yeah, so I am pretty much exclusively use O3 at this point. Once you use a good model, I think it's very hard to go back. I think I could probably use Gemini 2.5. I hear it's really good. But of course, as we've talked about, if it's good enough, why switch? And I use Deep Research about five times a week. And, you know, it's hugely helpful. And I, I think even one time that it saves you a few hours of doing work sort of repays the cost, absolutely makes sense.

• 39:37Speaker 02

  What do you use Deep Research for?

• 39:40Speaker 00

  It's a mix. One answer is I'm batting around something with my kids. And it's a question that no one has ever asked before, probably. And I want to know the answer. For example, what happens when you compress wood? You know, it starts off, it, it, you know, it's elastic compression and then it starts deforming and then you go a little further and it becomes diamond. And then you go a little further from that and it becomes a black hole. But actually, there's like a dozen steps. And so that's a really fun topic to dive into and just, you know, this is the kind of thing that would have been an ex KCD comic 15 years ago and would have taken him weeks to figure out. And now you can get an answer just in a few seconds.

• 40:20Speaker 00

  Also, I use it when I'm thinking about a new domain or a new startup opportunity. You know, well, if I'm interested, if I was interested in robotics, tell me everything there is to know about a particular company or about a particular market.

• 40:37Speaker 00

  And

• 40:38Speaker 02

  That's our daily life.

• 40:39Speaker 00

  Yeah, yeah, yeah, yeah.

• 40:41Speaker 02

  Any other new products?

• 40:43Speaker 00

  Well, like you mentioned, I use an AI assistant for scheduling, which is great. I mean, I'm solo right now. You know, I could hire an assistant, but it's just actually more fun to do things myself. But calendaring, it's really boring and it's just very nice and very pleasant to have, be able to CC, you know, an AI agent and have it do the calendaring for me.

• 41:03Speaker 03

  I'd love to hear a little bit about managing, you know, OpenAI, the research org. It's such a, you know, collection of insanely smart individuals, creative, I'm sure. Like what, and you know, the feedback we have on you is exceptional in terms of, you know, what a fair and what a great manager and leader you've been for the organization. I guess what have been some of your lessons, uh, leading an organization like that?

• 41:27Speaker 00

  So this, this sounds sort of boring, but like the core thing that you have to do as a manager is you have to really care about the people you're managing. And this, this maybe isn't relevant a lot of the time. A lot of the time as a manager, your day-to-day job, you know, you're coordinating, you're, you know, helping people understand things. And, you know, loyalty doesn't really matter that much. But there comes a time as a manager when you have to ask someone to do something hard. Early on in your career, this is when you have to ask someone to come in and work on Sunday when they'd rather be playing basketball. But later in their career, it's, you know, working with someone and you have to ask them

• 42:07Speaker 00

  to give up a project they really care about and give it to someone else. Or share credit for a research breakthrough that they know they could get to by themselves, but that they, that, you know, a team of people, you know, not just this one talented person, but two very talented people or three very talented people working together could get done even faster. And one thing I learned from working with Alex Carpe at Palantir is that

• 42:35Speaker 00

  Very talented people have superpowers, but they also have debilitating weaknesses. And for people who are at the very edge of these capabilities, they often don't even understand what their weaknesses are. But it's extremely apparent to everyone around them. And for me, as a manager, it's something that I could see very easily. And at this level of capability, when people fail, it's almost always a form of self-destruction.

• 43:05Speaker 02

  Wow.

• 43:05Speaker 00

  That there's a choice that they could have made. And I don't mean little failures. I don't mean like, oh, I had a bad day. I mean, when someone makes a career-altering choice in a bad way, it's almost always a matter of self-destruction. Because they had to do something that was very difficult for them. They had to confront something that was extremely scary for them to do. That to everyone else is kind of obviously the right answer. It's obviously the right thing for the company. But it's emotionally extremely hard for them. And if you, going back to being a manager, if you as a manager, if people know that you're in it for yourself, when you tell them to do something, they won't trust you.

• 43:48Speaker 00

  But if they know that you were doing what's best for them, then when you tell them to do that thing that is super hard and extremely scary for them,

• 43:57Speaker 00

  sometimes you can help them across the chasm and you can solve the problem and prevent them from doing something really stupid and end up with something that works out really well. And I hold this bar even for firing people. For me, it is always, when I am talking to someone, I have to be talking to them, giving them advice, helping them do the thing that is best for them and for the company. Even if you're firing someone, if they're not going to succeed in this role and I have invested enough time to make sure that they won't succeed in this role, then it is in their own best interests for me to tell them that they're not succeeding

• 44:41Speaker 00

  and give them the opportunity to find somewhere else. Loyalty, in the end, is the thing that I think unlocks all of the other things that you want in management.

• 44:50Speaker 02

  I really, really love that. There was a nuance there that you said in the middle around, you know, working with a ton of high performing individuals who are really excited about a particular research direction that they want to break through. They know they can get there, potentially by themselves, essentially with one or two others. They all have a good dose of confidence, maybe sometimes ego. How do you actually get them or convince them to embrace that, you know, effort of working together to get there?

• 45:21Speaker 00

  Yeah, it's very hard. And I think this is actually one of the things that's very different about a research lab from an engineering culture. Because in an engineering culture, it's sort of an assumption that we're all working together, we're all building one product. But research often comes out of academia, which has this very negative culture of, you know, it's a P I, you know, it's his team, who's going to be the first author, who's going to be the last author? None of the other people in the middle matter. And we struggled with this a lot. And I don't think there is any one answer. One thing we tried, which worked well for a time, we published some papers where we actually had OpenAI be the first author. So that there wouldn't be the fight over who's the first

• 46:02Speaker 00

  author. That was, you know, one technique. It didn't, you know, we couldn't always do that. It didn't always make sense. But, you know, in the end, the key is really, when you work with people, you understand there's something they want. And you have to find a way to give them the thing they want, and let them do the thing they want to do, the art that they're trying to create, while also letting all the other people do that, and having it all add up to one big hole. And just, just spend time over and over again, making sure you're solving that problem.

• 46:36Speaker 02

  Security, I know, is an interesting topic to you. In an increasingly agentic world, what kinds of security issues do you think we should be aware of? And where do you see potential opportunities?

• 46:47Speaker 00

  When I think about how AI impacts security, for me, the first order is the much easier ability to do offensive work than you could do previously. And so the number of threats have gone up, the time to execute on a threat has gone up. And so that then pushes the defense to be much more agentic. So there's a company I'm an investor in, it's called Outtake. I met the team, they're a group of ex-Pallanier folks. And we also ended up using them very successfully. At OpenAI, and what they've done is that they have made an agentic stack for doing cybersecurity that uses very little human input.

• 47:30Speaker 00

  And I think this is, you know, right now, we're at a place where the models can actually do all of these things. You know, if there's something that a human could do that's sort of these, one of these bulk operations, you know, if you can't make the model do it, that's your fault. It's not the model's fault. But the barrier then is that businesses and organizations aren't set up to do this. They have to go change their business processes in order to make this happen. And so I think that's an opportunity for startups where similar, you know, as, as big as, you know, this shift from web to mobile is just disrupting the existing businesses because it may be faster for you to replicate their technology and their distribution than for them to be able to change

• 48:11Speaker 00

  the way they operate, to get rid of or reduce the number of humans they need.

• 48:16Speaker 02

  Awesome. Bob, thank you so much for joining us.

• 48:19Speaker 00

  This has been really fun.

• 48:19Speaker 02

  It's a pleasure to have you here.