Reid riffs on fostering trust with founders, in boardrooms, and in AI

In this week's episode, Reid and Aria unpack why trust is a practical skill, not just a feeling. Reid riffs on the importance of hands-on investor-founder relationships, boardroom decisions that can make or break a company, and investor behaviors that build durable trust. They also discuss what product teams and policymakers must do to earn user confidence in an age of AI. Along the way, Reid shares personal anecdotes regarding his work with OpenAI, crisis leadership, and bridging Silicon Valley policy with the rest of the globe. For more info on the podcast and transcripts of all the episodes, visit https://www.possible.fm/podcast/

• 0:00Speaker 00

  Elon had decided that OpenAI should only be a company that he owned the majority of. After a town hall saying this is what he wanted, and the company said, no, he said, well, you're all a bunch of jackasses and left. He asked me a question out of the blue, no prep, he's like, so-

• 0:11Speaker 01

  And this is in front of the whole OpenAI staff?

• 0:13Speaker 00

  Entire, it's an all hands interview. Started very early with Brian, because when he and his co-founders were sitting down with me and pitching me on Airbnb, by minute three, I was like, okay, I know I want to offer this.

• 0:26Speaker 01

  Reed, it is so thrilling to be here at the Masters of Scale Summit. We get to do Reid Riffs in person, which we don't always get to do. And today, we're going to talk about trust. You have invested in some of the most important technology companies, Airbnb, Figma, Facebook. They are about building trust with the consumer, trust as a social platform. And actually, just a few weeks ago, we were talking to an entrepreneur, and they were talking about the entrepreneur-investor relationship. Is it contentious? Are the investors and the board members pushing the entrepreneur to do things? Should the entrepreneurs share everything with their investors? Or should they keep things close to the vest? So talk to me, how do you build trust with those entrepreneurs? So they are open and honest with you about the state of the company and how you can help?

• 1:10Speaker 00

  So the first thing is general advice, both for the entrepreneur and for the investor, you know, person X, person Y, is you have to first start with a working assessment of how much trust can the other person hold? Because building trust that's not held in trust is actually, in fact, can be bad and can be disastrous. And that's part of the reason why frequently, in places where there's conflict in a relationship, people tend to start with, oh, let's not go into the trust thing, because since we have potentially interests that are in conflict with each other, okay, we'll start from a non-trust perspective. But of course, that collaboration is super important. And without trust, you can't make decisions effectively, you can't communicate effectively.

• 1:55Speaker 00

  You might very much miss out on thinking, problem-solving. You might not be able to survive as well. You know, most companies go through difficult times. If you have no trust, or mistrust, or distrust, as opposed to trust, that's all like very challenging. So it's very important, first, as both founders, CEOs, and investors, to try to choose investors, try to choose board members, try to choose entrepreneurs, try to choose CEOs. That you can build a deep trust relationship. Because otherwise, it's basically non-functional.

• 2:34Speaker 00

  Right, so now, but trust, like trust building is, you know, you have to have, and by the way, you may not go like, okay, I think I can, and now I need to go through steps doing it. And all trust building is a journey. Right, it's never like, oh, I just, like, I met you in five seconds from now, I like, I would trust you with my children and all, it doesn't work that way. And part of that journey starts with, like, for example, one of the things, the reason why, um, the entrepreneur you referenced, mentioned this to us, is, um, that, you know, one of the things that he was noticing was that I tend to be pretty explicit. I go, look, here's where my interests are in this stuff. Right, and here's the advice I'm giving you that's in my interest.

• 3:18Speaker 00

  Here's the advice I'd give you that's outside my interests. Here's the advice I give you as part of our partnering interest. And then you, by being explicit and putting it all on the table, people go, oh, you're not trying to hot, like masquerade everything as interest, that your interests are my interests. You're not trying to hide the fact that you have some interests that are not my interests in this. Um, that you're putting it on the table explicitly because you want to share the navigation of it. Like, kind of a classic issue that when you have, because, you know, some parts of, of life and work and investing are zero-sum games. We try to play as many non-zero-sum games, as many positive-sum games as we can. But like, for example, when you're doing an investment, it's like, okay,

• 4:03Speaker 00

  I'm putting in a hundred bucks, what percentage of the company do I get for a hundred bucks? That's a zero-sum negotiation. And so, uh, how do you navigate that? Part of the thing I actually use that kind of negotiation for is say, can we build trust in each other? Can we identify what our interests are? Can we navigate them? Can we deal with the fact that we have an area of conflict? That's actually of conflict between us, to be positive in how we navigate the relationship. Now, there's actually, as you know, Parallels between this and the other thing that I thought deeply about for decades, which is friendship, because you know, part of the thing that like too often, for example, in friendship, people say, oh, I try to avoid conflict because I want any conflict in the friendship.

• 4:44Speaker 00

  And actually, in fact, what you want is you want to steer into the conflicts that are potentially real, because part of what happens when you do that, and you of course try to do it gracefully, and try to do it constructively, and try to do what you're learning about yourself and about the other person in terms of doing it, but when you steer into that, you either become closer friends or not as close friends. But if you're not as close friends, that's actually better outcome, right, in terms of doing it. So when you find an area, and I do the same thing in, with people I'm working with, with, with, you know, entrepreneurs, with investors, I go, okay, let me steer into the area of conflict, let's see how we're navigating it. Now, one

• 5:25Speaker 00

  of the challenges that you have, that's more high stakes in the investor relationship than in like, look, friendship, obviously seriously deeply important in the meaning of life, but in friends, you can kind of go a little closer, you could say, hey, we're not gonna, we're not gonna cover this area, and we'll still be friends over here, and that's fine. And depending if it's not too deep of a moral topic, or other kinds of things, that that works just fine. One of the things that I kind of most often tell founders, young founders doing this stuff, is one of the weird things about taking an investment in a, in a board seat from somebody is it's, it's kind of like getting married on like two power, yes, well, two PowerPoint presentations and one or two dinners. And because once you're in it, you're like, you're

• 6:08Speaker 00

  all you're in it for a while. Yes. And so it's actually one of the reasons why I highly recommend both investors and founders, reference check, right? Because it's like, how do I, how do I make that proxy work? These questions around trust building are actually extremely important. It's like, one of the things that like I most often, like, like I surprise, you know, founders when they're calling me and asking me about different, you know, possible investors, like, look, actually, in fact, there's a huge virtue in that that's someone you can go on a trust journey with, because all startups go through moments where we're like, why did the hell we think this was a good idea? And you want someone who goes through that with you. And by the way, now, then founders are like, oh, I want someone who just blindly follow what I do. It's like, well, you're losing out on tons, like.

• 6:54Speaker 00

  Like, I don't want friends who, who, who say, oh yeah, read whatever you are, that's fine. It's the no, I want friends who go, hey, you'll be a much better person. You'll be more capable in the world, you'll be a better friend, you'll be happier if you start doing more of X and less of Y as an instance of doing this. That's the kind of partners you want in business. And, and, and you have to show it. This gets back to that kind of trust building.

• 7:20Speaker 01

  Can you give us specific examples of ways you have built trust with founders that you've invested in? So Brian Chesky from Airbnb, or Sam Altman and Greg Brockman from OpenAI, like, how do you set that relationship on the right foot?

• 7:33Speaker 00

  So the first thing, just as we were saying before, is it's important to be genuine. Right? So like, this is one of the, the things that's a little key. When someone comes up and says, I'm a networker, I want to have you in my network. It's like, no, no, it's, it's a relationship. It's, it's a dance, it's bidirectional. It's not, it's actually important to be intentional, but it's intentional in the same way that like, for example, if you're learning to dance the swing with someone or dance the tango, you're collaborating in it. It's not the way that you're manipulating them, it's the way that you're going into the dance. And that's, that's, that's an important first part of the mindset on, on part of both parties. But then, of course, people say, oh, no, it should be all the natural, you're not, you're not exposed. No, no, you learn how to dance better. You learn how to do this stuff better. So that's, that's an important sub-part of it.

• 8:15Speaker 00

  So now part of it, like, for example, it started very early with Brian, because when he and his co-founders were sitting down with me and pitching me on Airbnb, by about minute three, I was like, okay, I know I want to offer to invest. I was like, let's change this in a working session. Let's start working on some of your hardest problems, because both I will get a sense of you and you will get a sense of me. So we'll start working on this problem and that, that's how we're going to go through this. And they were like really surprised because no investor had ever, ever really done that. It was like, okay, let's work through them and like, like, how are you trying to solve this problem? And isn't this an issue? And not, isn't this an issue? Prove it to me. It's like, how would we solve this? Like, what were the, what are the things that you're thinking about doing? Here's something I could think of, you know, da, da, da. Boy, this will be hard.

• 8:58Speaker 00

  We're going to have to stay focused on this as we're going on, et cetera. So that's, that's one, where you kind of roll up your sleeves and you're kind of working with them and you're showing that you're doing that and you're showing that you, you know, you're, you're capable of throwing ideas against the wall and. When you know, like you threw a good idea, I threw, remember I threw an idea in someone's wall and then five seconds I was like, oh, that was a dumb idea. Let's move to the next one.

• 9:18Speaker 01

  Well, you're also not like afraid of the problems. You're saying, tell me the problems because I'm actually here to help.

• 9:23Speaker 00

  Yes, exactly. And that's one of the pieces of advice I give entrepreneurs and dealing with investors is don't try to say, we have no risks. Right. If like, duh, you have risks. So if you're not telling me what the risk, you're either dumb, bad, or lying to me. Also bad. Right. So, so it's like, no, no, tell me what the risk and tell me what your theory of the game is with these risks. So that was, that was with Brian. Now, the other ones with like Sam and Greg, there were, I mean, there were lots with Brian too, but you know, maybe the, the, the interesting thing kind of came from when, you know, basically Elon had decided to open AI. It should only be a company that he owned the majority of.

• 10:03Speaker 01

  Yep.

• 10:03Speaker 00

  So after a town hall saying this is what I wanted and the company said, no, he said, well, you're all a bunch of jackasses and, and left. And you know, Sam called me and said, what we do is, and I said, look, I'll step in, I'll help. And you know, part of that was to say, hey, I'm, I'm not just there for you in the good times. I'm there for you in the bad times. It was literally like the very first reaction. Greg didn't actually know me that well. And so Sam said, well, look, I'd like to join, have you join the board, but would be okay if Greg and Ilya interviewed you. And I was like, of course it is. Right. Like it's not like I'm Reid Hoffman. How could you possibly interview me? It's like, no, that'd be great. I would love that. I would get to know them better, that kind of thing.

• 10:43Speaker 00

  And I understand that like they're just like, they're just coming out of having had an investor who was like, you're all a bunch of jackasses. And I think you're going to die and I'm, I'm leaving the building. And it's like, okay, so like I'll come do that. And so in sitting down, I was like, you know, like for example, and as part of it, I was like, okay. Like, look, for example, one of the things I was like, one of the natural questions you guys should ask that you might feel intimidated asking, how do I think that I'm different than Elon as a partner? Like as as someone who's collaborating, working with you? And they said, yeah. I was like, yeah, let's start with that question. Right. And here's one of the things.

• 11:22Speaker 00

  And by the way, part of the thing that I, you know, like what I always do when I'm talking to to investees and so I say, look, I'm going to give you a bunch of references. I would like you to find anybody that I perfectly work with. You're allowed to talk to anyone and you can tell them they'll know this, that they that they should tell you my strengths and weaknesses and the way. I will identify people to you that I think are my my negative references that I've that I've worked with before. And one of those people got so sick of it that one, when one person called them was like, well, Reid Coleman called you for negative references. It's like, can you tell him to stop calling people for negative references? Right. You know, as an instance, because it's like, because look, it's I want to be in truth with you.

• 12:04Speaker 01

  Absolutely.

• 12:05Speaker 00

  Right. And so they're like, oh, and even before they'd done the calls, and I know what they went out and did phone calls because I usually get a call saying, oh, you know, hey, Greg called me and blah, blah, blah. It was like, oh, that's good. Great. Right. Let's have a real conversation. You want to have that real conversation and and part of it is, and again, like the earlier things, was by being there and present with them. Now, obviously, you know, part of most even amazing companies like OpenAI, like Airbnb, you know, all, you know, and I've had the honor of being a part of many of these journeys is they all go through value of shadow moments.

• 12:46Speaker 00

  And a trust is intensely built when you show that you are there to share the pain and help with it in navigation. Because until you get pressured that way, people go, well, no, no, actually, no, you take all the pain. Now this is all your fault, you know, you fucked it up. And I've seen a lot of investors lose their cool.

• 13:06Speaker 01

  Yep.

• 13:07Speaker 00

  Uh, and like just heap garbage on the entrepreneur when you're like, well, this is a risky game. You encounter problems. It's, you know, um, and the point is not to panic. The point is not to blame the entrepreneur, um, unless the entrepreneur really like, oh, you were embezzling. Blame the entrepreneur, but you know, like, but like, it's like the, the nature of building these things. And so, um, like, you know, I, like, I would get these calls like, okay, is, is, you know, Sam Altman a good guy? I'm like, yes, he is, right? And yes, he is because, um, it doesn't mean he's perfect.

• 13:41Speaker 01

  Of course.

• 13:41Speaker 00

  Like saying Reed is perfect. That's dumb, right? Ari is perfect. That's dumb. But it's, but it's the, no, no, he actually in fact has, um, humanity and what's going on for how AI affects, um, like not just industry, but society at fundamental. And part of like, again, like Sam sees me do that and he, he knows that I'm like, he, he hears echoes of it. Cause I don't just then call and say, oh, Sam, I gave a positive reference. It's like, no, no, no. Like it's part of what that is. And, um, and I think that that's, that, that's where people learn that you're trustworthy.

• 14:20Speaker 01

  Absolutely.

• 14:20Speaker 00

  Right? In, in the relationship.

• 14:22Speaker 01

  What do people say about you when you're not in the room?

• 14:23Speaker 00

  Yes.

• 14:24Speaker 01

  Like that's what you want to know, that you are the person you're there with during the board meetings. But you're also there repping them when you're talking to other investors, when you're talking to other founders.

• 14:32Speaker 00

  Journalists, founders, the whole thing.

• 14:34Speaker 02

  This podcast is supported by Google. Hey everyone. David here, one of the product leads for Google Gemini. If you dream it and describe it, VO3 in Gemini can help you bring it to life as a video. Now with incredible sound effects, background noise and even dialogue. Try it with a Google AI Pro plan or get the highest access with the Ultra Plan. Sign up at Gemini dot Google to get started and show us what you create.

• 14:59Speaker 01

  You told me a story about you going into an OpenAI meeting where Sam was interviewing for a fireside chat. Can you, can you say a little bit more about that meeting?

• 15:09Speaker 00

  So this was a classic Sam thing, which was hilarious because part of, uh, again, trust maintenance, actually. Which is one of the reasons it was unusual. Trust maintenance is. You have to have CEOs that trust that you as a board member will not operationally undermine them, except by direct discussion, either one on one or in the board meeting. Like if I'm undermining you, it's because I'm doing it to your face. Right. It's part of what we're like, the thing that we're doing for the responsibility of the company. And so one of the things I've learned over time, one of the reasons why like now, literally for well, maybe even two decades at this point, CEOs trust me to talk to anyone in the company, trust me to talk to the exec staff.

• 15:53Speaker 00

  They're like, I'm like, hey, I want to call the, you know, these exec staff, I want to talk to them about this. They're like, great, let me know what's going. Because I know I don't undermine them. Absolutely. You have to be careful because like, oh, you know, Ari is the CEO, do something, and I'm sitting with your exec staff and I'm saying, so how good of a job do you think Ari is doing? Right. That question itself is undermining. You do not ask that question. Totally. Right? Now, say for example, that's a question I'm interested in, the way that I would ask that question is say, so tell me a little bit of how do you think the org is doing? Right. Right? Like, what are the things? And as you go through it, you'll get the echoes of the things that you need in order to get an assessment of that. And that's the way to do that and to help the organization, to help Ari and et cetera, without actually in fact, doing the undermining thing.

• 16:36Speaker 00

  So I'm very careful about not undermining the CEO. So we're sitting down and Sam asks a question out of the blue, no prep, he's like, so... And this is in front of the whole opening. Entire, it's an all hands interview. Because again, like Sam was like, would you do this to help build confidence in companies? Absolutely, of course. And he goes, so what are the conditions on what you'd fire me? Right. You're like, what do you want me to say? Like, I don't undermine confidence. It's like, well, you know, it's, it's, you know, obviously what I would start with is, you know, kind of like I wouldn't just do that, I would actually be working to try to help the whole company and all the rest. He's, no, no, no, no, answer the question, what would you do to fire me? And I'm like...

• 17:19Speaker 00

  Well, okay, right, like, look, if you're, um, look, there's the obvious things of

• 17:25Speaker 01

  Right, you're stealing money.

• 17:26Speaker 00

  You're stealing money, all those things, those are all things that, that, that are like, you're gone the next minute, right?

• 17:32Speaker 01

  Totally.

• 17:32Speaker 00

  But there's also things around like, okay, you know, working in this company is, is, is an instance of great trust. And so if you're not holding, you know, AI's benefit for humanity, if you're not holding that in, in subs- like, in, in the deep trust of, and, you know, obviously, there's certain things that could break it, that's equivalent of a minute, but there's also other things that'd be like, hey, I'm working with you, and if it's not working, I'm going to be escalating the, this isn't working.

• 18:02Speaker 01

  Yep.

• 18:03Speaker 00

  Right? And one of the weird things that has happened as part of, you know, functionally, every board I've been on, when we've had to fire a CEO, I'm always the person who's asked to fire the CEO. Right? And so I, I actually do do that. But by the way, I've maintained a relationship with every single CEO because I do it in a human way. I explain it, I don't ask them to agree with the judgment, but like, here's how we got to this judgment, these are the kinds of things we're doing. And it's like, look, if we got there, we would get there because it's not just about, like, like, your performance is your performance for the company, but it's also your performance for society, for humanity. And are you, like, delivering on the right way? And obviously, I want you to. My first job is to help you succeed.

• 18:48Speaker 01

  Absolutely.

• 18:48Speaker 00

  But my next job is to see the transition when it's important to happen.

• 18:53Speaker 01

  Well, continuing on the theme of trust, a lot of people, when they talk about AI, they talk about trust. They say, I don't trust these companies. They say, I don't trust these models. I don't trust the outputs because of hallucinations. Like, how are we going to get to a place where we're building trust with AI and society?

• 19:13Speaker 00

  Well, there's a couple of things as baseline. So one is, um, you know, a lot of times people say, I mistrust it. Look, there's the journalists and everything else is generally, oh, mistrust this, mistrust this, mistrust this, it's bad because it was using copyrighted material to get caught, you know, to be made, or, you know, blah, blah, blah, blah. And so, but the first is, use it. For individuals, use it, right? Get a sense of it. They say, "Well, but I don't know what's going on with my data that's going into this." And you're like, "Look, these companies are trying to build trillion-dollar franchises." They care a lot about not screwing up you with your data, right? So like, people say, "Well, but I have a, I have this genius secret idea." It's like, they're not looking through their chat logs to mine for your idea. Right, just steal your idea and run with it, yeah.

• 19:57Speaker 00

  You know, and if you say, "Well, I say something personal, like I'm worried about my friend's mental health and so forth," it's like, they're not gonna go publish that. You know, like, it's like, there's much more natural alignment of incentives there. So try it, and try it for serious things. Now, that's on the individual side for realizing what kinds of things they could all trust on, and the users. Now, the companies obviously need to do a lot of things too. Part of those things is to hold that trust. So for example, you say, "Well, I've got a whole bunch of chat logs with a bunch of people." Well, make sure that you're using modern, the best that you possibly can in cybersecurity. Make sure hackers can't get access to it. Make sure it doesn't get accidentally revealed, or, you know, like, published to the internet

• 20:42Speaker 00

  by the dark web and so forth. You know, like, you have to, you have a responsibility, and it's commit to that, and engage in it, and take your lumps if you get it wrong, right? And, so, you know, that kind of thing, and that's important. And then, also, you know, like, okay, I think one of the things I've been saying at least pre-COVID for some number of years, and obviously still now, is as technology becomes more and more part of our everyday fabric, the kind of our entire life around us, as you're like, mobile phones and AI agents on them, et cetera. It's, I think companies should be very explicit about, "Here are the things that we think we are holding and trust for you.

• 21:25Speaker 00

  Here is what we are building towards, here is what we're doing, here's what we're trying to, the changes we're trying to be in the world, here's what we're trying to be good for you, and here's what we're signing up for, that you can hold us accountable for." Yep. And I think that...

• 21:40Speaker 00

  Dialogue is very important. Now, it's of course still important that they have like a 35-page Terms of Service, which is, you know, the legal gotcha because this person goes, "Oh, I want to sue you for this." And so you read through a Terms of Service, they're always like, "We're not responsible for this, and we're not responsible for this, and you're responsible for that." And you have to because of a legalistic society. But nevertheless, I think it's important to say, "Look, here is what our theory of what's good for you is. Here is what we're building towards. Here is what we're seriously investing in, and this is why we're doing it." And I think being in that dialogue with the individual consumer, with the communities, with the societies, with governments, with NGOs, it's one of the reasons why, like,

• 22:23Speaker 00

  people are surprised when I'm like, "No, no, I like journalists playing critical functions. It's actually important." And then like, we're like, "Oh, well, why did you provide funding for co-founders?" We're like, "No, actually, in fact, like, it's important that all of us are engaged in this dialogue to make this technology more human. That's an critical part of it." And so as opposed to like, "You're not allowed to criticize me." That's bad. That's trust, trust eroding. It's like, "No, I welcome your criticism. I may disagree with it. I may say, 'Hey, I can't get to that yet,' etc., but that's an important part of the dialogue for how we improve."

• 23:00Speaker 01

  Right. I mean, to your point on dialogue, the politicians who do town halls versus the politicians who say, "I'm never doing a town hall because I don't want any questions from my constituents." Who do you trust more?

• 23:10Speaker 00

  Yes, exactly.

• 23:11Speaker 01

  So we're in the midst of like a video model renaissance. We've had so many video models come out in the past few weeks, and Sora 2 is everywhere. And last week, our, my Twitter was full of Sam Altman stealing things from Target, and our internal Slack was full of our coworker hitting game-winning shots as a 76ers basketball player. It is everywhere. And so some people are saying, "This is amazing. This is..." So many people, like, the burst of creativity that people are getting from these new video models is incredible. Some people are saying, "This is great because this is getting more people using them." And the more you use them, the more you use them as a social platform, the more you trust them. Other people are saying...

• 23:55Speaker 01

  I thought these companies were gonna like get to AGI and cure cancer. Like, why are they giving me second, seven second videos of my face on them? How do you see sort of the new release of Sora either playing into the trust? Or maybe even if you don't, what does this new release and amazing adoption tell us about this moment in AI?

• 24:16Speaker 00

  Well, maybe there's three things. And by the way, all technology, iterative deployments are important. There will be things that are, that are fail moments in it. They'll have to iterate and change it. That's, by the way, not a fail for them. That's called iterative deployment and learning. And so the feedback is really important in terms of what are the things happening? And you cannot get it perfect before you get it. Like, you tried to get a car perfect before you put the car on the road. You'd have no cars. And so, so it's kind of like, oh, and by the way, some of it comes from, you know, regulation, you know, which is like, you know, questions about seat belts. And some of it comes from, oh, shoot, airbags, you know, like, like just industry kind of kind of making work.

• 24:52Speaker 01

  Absolutely. Gotta improve as we go.

• 24:53Speaker 00

  Yes. And so, so part of it is a very strong plus one to launch and iterate and engage. Now, obviously, try to avoid catastrophic things. Try to avoid things that break systems, break financial systems, you know, cause serious harm across a large number of human beings, or even a few human beings if it's very serious harm. And obviously, they did a whole bunch of that, like trying their best they can before they got out. They did a lot of alignment testing, a lot of like, what if things happen? But that doesn't mean it'll be perfect and it'll be iterated. Now, that being said, I think there's a couple of things. So one is, again, it gives another surface for people to interact with this thing.

• 25:32Speaker 01

  Yep.

• 25:33Speaker 00

  I think part of this surface is to enable people to engage with their creativity, like, you know, pictures worth a thousand words, you know, as part of the image generators, but also now video and, you know, like our coworker, you know, kind of going, oh, my God, I always wanted to be a 76er, you know, basketball player. No, it's like, it's like, it's a good way. That's, oh, it's sick of panic. No, it's fun. It's play.

• 25:54Speaker 01

  Totally. Having a good time.

• 25:55Speaker 00

  Yes, as a way of doing this. And I think. It creates a palette. I mean, also, for example, you know, the Brobot stuff that Greg's doing. It's hilarious. Yes, it's hilarious. And, and, you know, if people haven't checked out Brobots, they obviously should. Absolutely. Right? Because it's, it's, it's, um, you know, part of his thesis of that was to say, look, part of what Super Agency, and this is Greg Beato, co-author in Super Agency, is say, I have zero experience with making any video, any music, and look at what I can do because of this. And I can express my kind of funny, like you, you get like Greg's sense of humor and commentary on AI systems and all the rest in a fun music video.

• 26:39Speaker 00

  That's the way of doing it. So, so that creativity landscape, I think, is important. Now, I say, I thought you were going to work on curing cancer, which of course is in IR. But what people don't realize is, look, there's a bunch of different iterations for the evolutionist models. Part of the reason why they work on images, part of the reason why they work on video, is because the multimodal side of these things, both in comprehension and understanding, but also output, are actually, in fact, part of what gives us this massive elevation and functionality across all this. And so, you know, one of the simplest lightweight ways to say, hey, we're going to have this thing, you know, start understanding,

• 27:19Speaker 00

  you know, kind of real world through interaction of consuming and understanding videos, through generating videos, through getting feedback on the videos, is the way they're doing it, is actually a part It's on the path to getting to. Of the AGI mission. And a path onto the mission of, like, for example, one of the things that, you know, companies like Aurora and others, you know, Google and MIMO are doing intensely is creating simulation with AI data to try to be, because you, you like, it's like, well, can we have these trucks drive at every possible scenario and circumstances? No, no, you can't do that. It just doesn't happen that way, because some of these circumstances are very rare. So what are the things you use? Generated data for simulated data, etc. It's like, let's go through every possible really

• 28:04Speaker 00

  bad circumstance.

• 28:05Speaker 01

  What's the rarest of edge cases? We would not see in the real world, even in a million instances, if we can generate.

• 28:10Speaker 00

  And part of the reason you do that and train them all is that it's part of how you get in substantive ways safer than human drivers. Because the human drivers never encountered those circumstances. Maybe they have some more natural biologically inspired reflexes in terms of how you're driving, but they haven't been trained on all the what happens when this and that and the others coming together? And what do you do to recognize you're in that circumstance, to interact on it, to do that? And that's, again, part of like the kind of the world generation part of these things. And that's part of where the going into, you know, video generation is steps in that direction.

• 28:46Speaker 01

  We talk about trust. We've spoken about trust in entrepreneurs, trust in AI. And then if we actually zoom out, I feel like geopolitical trust is also a big issue, whether it's the US and China, or what is the potential AI collaboration? You helped set up the UK Safety Institute. And most of your focus has been on American AI. And specifically, actually, how AI is American intelligence. And we want to imbue AI with sort of Western values, democratic values. But you've also spent some time recently in Europe, and you were talking to leaders in London, in Paris, etc., about their AI efforts. So why do you think it's also important to engage with some of these European leaders?

• 29:27Speaker 00

  I think there's a number of reasons. First, I actually think the cross-cultural, cross-border dialogue with other cultures that are taking kind of the key values that we find is very essential, human rights, individual rights, autonomy, ability to participate in a democracy and have a voice, ability to say, how do we, what are the ways that we should empower, you know, individuals in society where, you know, governments are very important, but governments should have, should help shape, enable those values versus suppress them. And that's part of what we mean by Western democratic values. And this may shock some people, it's like the US is, you know, one of the oldest amazing

• 30:11Speaker 00

  Democracies on the planet, but there's others too, and we can learn from them. And the fact that we would not say, hey, we should learn from them, we should collaborate on this together, we should learn about these things is just super important. And I think it's it's better in partnership, better in alliance, better in collaboration. And so obviously, I've got years of connectivity with the Brits, you know, went to school at Oxford, Marshall Scholarship, which is part of the thing they gave as part of the thank you plan for the Marshall Plan and kind of, you know, building bridges. But also, one of the things I learned about being in Oxford was, ah, here are the things that we can learn from more global society that I hadn't learned as just by being an American. And by the way, of course, that's part of the the wonder of America is like being a

• 30:56Speaker 00

  country based on immigration from many different places is there's a number of things where like immigrants bring in some of like their best things, like hard work, cultural values and certain things. And then we we come into

• 31:08Speaker 01

  Not to mention food.

• 31:09Speaker 00

  Not to mention food. Yeah, it's like, you know, like, you know, people like we people say, what's your favorite food? They usually say something like Mexican or, you know, or sushi or, you know, as an answer. And so, and, you know, part of it, and I started this decades ago here here in the valley, and now it's kind of just also a little bit more on the road is when a minister from a from a from a Western democratic country comes to Silicon Valley says, I want to learn from it. I want to understand how to how to help my country with this. If I can, I will always meet with them because the more Silicon Valleys there are, you know, around the world, especially around the Western world and creating technology, having humanistic values in terms of what you're

• 31:54Speaker 00

  building, etc., the better we all are.

• 31:57Speaker 01

  I mean, it harkens back to your point. We are not playing a zero sum game, actually, when it relates to building up amazing Western values with AI talent other places. This is positive sum, and so we all win.

• 32:08Speaker 00

  Exactly. And so the first time I met Macron was actually when he was a minister, not not the president, because he came over here with a, hey, I'm trying to figure this out. And so I get a call from a friend saying, hey, this minister from France, Macron, who you know, I'd heard of, I'd read some articles about, was here. I'm like, great. And he's like, can you come by and say some things about like entrepreneurship, investing, technology, risk taking? What are things he wants to figure out how to help amplify this within France? It's like, great, I'd be delighted. We spent two hours talking. Matter of fact, I think he's the first minister that ever went, I want to take a selfie with you. He was like, I got my phone, let's take a selfie.

• 32:48Speaker 01

  That's hilarious.

• 32:48Speaker 00

  Like I'm tech forward as a way of doing this. And so, and it's been, you know, Singapore, Italy, you know, France, the UK, like all of these folks and, and, and part of it is, like, like I actually even tell a whole bunch of, of kind of US government people because it's very funny because too few of the US people come to Silicon Valley and say, look, how do I learn from here about what I can do to help the rest of the country?

• 33:17Speaker 01

  Right, absolutely.

• 33:18Speaker 00

  Like, how do I, how do I bring economic prosperity, technological transformation, invent innovation, invention? What are the, it's not to say, like, do it exactly Silicon Valley style. No, that would be better to have diversity, right? And so, you know, sometimes as a governor from Colorado and so on, which I've done, but it's like, that kind of thing is actually super important.

• 33:39Speaker 01

  So you mentioned in your most recent meeting with Macron, you talked about the Lafayette Fellowship, which is actually having Americans go to study in France. So when you support something like this, like, why do you do it? Is it like for a tech pipeline? Is it exchange of cultural values? Like, what do you think the essence of the Lafayette Fellowship is?

• 33:58Speaker 00

  Well, I'm not adverse to a tech pipeline, right? I mean, win-win is, is, is, is business slogan, but it's a good principle.

• 34:06Speaker 01

  Sure.

• 34:06Speaker 00

  And so, but in this case, it was as simple as, like, I've also endowed some of the Marshall scholarships because I think this kind of alliance and bridge building is, is massive wins on both sides. And so for the, the countries to whom we'd go, hey, we are allies in the world. We are trying to create an elevation of the human condition. We are, we have these principles about how we, how we deal with children. These principles are how we deal with human rights. These principles about how we deal with political, you know, kind of systems and enfranchisement. We should be allies on these things. And so when I was asked, I was like, yeah, no, I think this kind of

• 34:48Speaker 00

  bridge building, because it's one of the things that certainly helped me open my eyes, because I, I had, until I studied Oxford, I hadn't realized both how much of an American I am, and also how much of a Californian I am.

• 35:00Speaker 01

  Yep.

• 35:01Speaker 00

  And it was, and it was like, oh, that's really interesting. And by the way, it's both the, the delight of that, and also then evolving and learning from there.

• 35:10Speaker 01

  I love it. Reid, thank you so much. Appreciate it.

• 35:12Speaker 00

  Always a pleasure.

• 35:13Speaker 01

  And a big thanks to Dan Nealon, Eve Tre, Brian Pugh, Aaron Bostinelli, Jeff Berman, and the frickin' awesome teams at Wait What and Earth Species Project.

• 35:24Speaker 03

  This podcast is supported by Google. Hi folks, Paige Bailey here from the Google DeepMind DevRel team. For our developers out there, we know there's a constant trade off between model intelligence, speed, and cost. Gemini 2.5 Flash aims right at that challenge. It's got the speed you expect from Flash, but with upgraded reasoning power. And crucially, we've added controls, like setting thinking budgets, so you can decide how much reasoning to apply, optimizing for latency and costs. So try out Gemini 2.5 Flash at ai studio dot google dot com and let us know what you built.