Elad Gil, Jared Kushner & Eric Wu on AI in Business | Brain Co

0:00Sometimes founders ask me, what's the biggest way to create a good culture?

0:31Welcome to Facebook. Vanka's husband. Vanka's husband. Which Ivanka? Jared came out to kind of talk to different people about AI and what was happening. We started just kind of comparing notes on what he'd been seeing from the corporate and government world in terms of people trying to adopt AI and not really knowing what to do. And I'd been talking to a lot of Fortune 100 CEOs and execs about sort of a similar thread. And so we kind of realized that the world's biggest institutions need to be able to have an easy way to adopt AI. When I met Jared Kushner, I think the media had portrayed him in a way that was very different what they like in person in terms of being very smart, driven, wanting to do the right thing for the world.

1:04One thing I've always wanted to do is like imagine that I, well, I actually don't want to talk about this. I can tell you after. I just don't want to put it on there. Off the record, we have an exclusive. Yeah, exactly. Print it. With Eric Wu, are you guys going to do anything with Open Door? You mean like a take private or something? No, but that would be interesting.

1:32One of the things that when I was doing research, I was listening to previous podcasts you were on, and there was a really good one with Patrick O'Shaughnessy. And you were talking about markets and how they collapse usually to two to three winners. So can you explain how that works and like what the framework is around that? Oh, I think there's lots of different market structures. So there's some things that form natural monopolies. Right. And so that for a long period of time was like the Microsoft OS or other things like that. And then the internet displaced that and allowed new entrants. Then mobile created another entry point.

2:07So some things actually naturally are just one player really dominates. It's kind of like Uber in the US or whatever now. There's many markets. And this was something I had to learn because I didn't invest in things. I'm like, it's not a monopoly, but it turned out multiple players could win. That are actually oligopoly market structures where you have a handful of very large players. Now, sometimes you have these large players and then you have a lot of fragmentation, you know, the three big players on 50 % of the market and everything's fragmented. And that traditionally has been payroll, ADP, paychecks, and then a bunch of stuff.

2:34And then hopefully now Rippling and other sort of gusto and related companies. And then I guess deal is another one. And then there's markets where things are just default fragmented. It's like the restaurant business, right? There's thousands of restaurants in San Francisco. Um, and so depending on different aspects of the market structure, what sort of customers you have, how dependent are they on your good or product or their network effects or their scale effects, et cetera, that drives which market structure you end up with. Um, so I think there's lots of different ones in tech and consumer tech, the monopoly markets were more common.

3:15That was Google to some extent in search. That was Microsoft with this OS, et cetera. um and then in many many other markets it's more an oligopoly structure payments is a good example where there's adgen and there's stripe and there's paypal and you know there's a variety of players do you think moats change over time i think that for some companies they build new moats over time for sure i think there's a handful of ways to have a moat and i think there um things things just tend to collapse into these sets of things usually i find it's kind of the opposite of there's no good generic startup advice you know kind of thing but the reality is there are patterns to.

3:49And so, um, you know, one mode is a scale effect. So for example, um, if you're a payments company, the more scale you have, the cheaper your interchange fee is on the backend, which means you can charge less, which means you can get more scale. And so you have the cycle, right? That'd be one example of a scale effect. Um, there's ecosystem effects. Salesforce is a good example. Everybody's built an app on top of Salesforce. Everybody's selling those apps are integrated in a deep way that helps create defensibility because if you're trying to build a Salesforce competitor, you may also have to build all these apps at once, right?

4:19So you have more to build against and it's tougher. So there's these ecosystem or platform effects. Some businesses have long-term contracts. So for example, in the medical distribution world, McKesson and that sort of world, all the pharma distribution, people end up in these like three to seven year contracts. So the entire industry is effectively locked up between a handful of players because they each on a chunk of it. So there's a lot of different types of, I mean, I'm not going to keep going because it's like listing all my companies alphabetically. Right. But fundamentally, there's only a handful of ways to do this and they tend to be very consistent, but they tend to be very strong.

4:54Network effects is another one, you know, shifting into one of your other strategies. We covered investing macro and everything like that, but I'm really curious how you shift from investing directly into companies to incubating them. There's a couple companies that you started that start with brain. So what's the deal with these brain companies? I mean, so Ankur came up, Ankur Goyal, who's the CEO and founder of Brain Trust, came up with Brain Trust, which I thought was a great name. And I helped him really early get that up and running. I did a lot of the early customer calls. And initially we talked about just doing that as like an open source project and all this stuff, and then it kind of converted into a company.

5:34So we were just doing it for fun. It's like a kind of a side project. and then when I helped get this other company up and running called Branco we were just calling it Branco as a placeholder yeah it wasn't meant to be the long-term name and hopefully Ankur isn't upset you know that we have another brain thing in the family but it was meant to be like kind of funny and quirky and it was an internal placeholder and then we're going to come up with something better and like Apple and like Stripe and a few other companies we just could never come up with anything better and the people who are working on it liked it so it kind of stuck So it was inadvertent.

6:06How did you recruit Ivanka's husband? Ivanka's husband. Which Ivanka? Ivanka Trump. Oh, yeah. So I met Jared when he was. So Jared came out to kind of talk to different people about AI and what was happening. And we started just kind of comparing notes on what he'd been seeing from the sort of of corporate and government world in terms of people trying to adopt AI and not really knowing what to do. And I've been talking to a lot of Fortune 100 CEOs and execs about sort of a similar thread. And there's Luis who works with Jared, who similarly was having sorts of conversations. And so we kind of realized that there was a very clear need and pinpoint around very large scale enterprise, government, like the world's biggest institutions need to be able to have an easy way to adopt AI for various workflows and purposes.

7:02And so we thought it would make a lot of sense to kind of build out a company that could have a core platform that you could then build different applications on top of that would serve those needs. And so Eric Wu came in as sort of a founder and chairman. And then Dan Ashton, who you know, and Marcia and others came on as sort of the initial founding team. And we kind of just were off into the races from then. And Clemens recently joined as CEO of the company. So how do you think about I and I think this company is really interesting because you're going after one end of the market that has been underserved so far.

7:37Like we're seeing a lot of value accretion in the hardware layer and then now in the LLMs. But like the services really has not been tapped into yet because it's just a downstream effect, I think, if downstream is the right word. But I'm curious from your standpoint, like how large is this opportunity? You talked about TAMs before, but like how large is this and who are they going to be? Yeah. I mean, I think this is a truly massive, massive TAM. And so, you know, it really is. If you look at a startup's life cycle, normally that's doing B2B, you tend to start off with smaller businesses you're selling to.

8:11And then you realize that's a tough market and you go mid-market and you spend a couple of years in mid-market and then eventually do big enterprise. And you land your first multimillion dollar deal and you're ecstatic or you have a$10 million deal. It takes you like three to seven years to get there. And so let's just jump straight to those customers. We have access to them. We know their needs. We can build really aggressively against it, but we can also attract a caliber of engineer to work with them that they just won't necessarily have access to in any other way. And that's because we have created a team or brought on a team or hired a team that are the types of people that normally be working at the foundation labs, right?

8:49They'd be working at OpenAI or they'd be working at Google or they'd be working at some of these great companies. And we've really taken the approach of saying we're both going to build out this platform. So it's a common piece of infrastructure that everybody can use. But then on top of that, we can build bespoke apps and then resell those apps to other people in the same vertical. And so if you're dealing with one of the world's biggest healthcare chains, you can repeatedly provide that to the world's biggest healthcare organizations because they all have similar needs. and so we think that strategy has actually worked really well and there's enormous pent-up demand because every ceo is asking what is my ai approach what's my ai strategy i know it can impact my margin or my productivity or my my sales output you know maybe i can double sales through it but they may not necessarily have the resources to to go right at it immediately or they'll have a research team internally that's working on it but that team can use extra resources or they can use some of the product work that we provide sorcery is brought to you by brex the financial stack trusted by more than 30 ,000 companies, including one in three venture-backed startups in the U.S.

9:49Nearly 40 % of startups bail because they run out of cash. Rex is literally built to help founders avoid that. Unlike traditional banks that let your money sit idle, chipping away at it with fees, Rex is designed to help you spend smarter and move faster. Their all-in-one solution combines checking, treasury, and FDIC protection into one powerful account. You can send and receive money globally at lightning speeds, get 20 times the standard FDIC coverage through their partner banks, and even high yield from day one. With same day and even same hour liquidity, access your funds anytime. Companies like Scale AI, DoorDash, Service Titan, HIMSS, Anthropic, Flexport, Robinhood, and Plaid trust and use Brex.

10:34Start today at brex.com slash sorcery. That's B-R-E-X dot com slash sorcery. Do you think that the MIT study was correct? I think it had aspects that were correct. So basically, the study you're referring to is the one that says that AI doesn't really help anything or whatever. Yeah, it said 95 % of AI pilots are garbage. But if you read it, it says 63 % of ones that were internally created were actually successful. Yeah. I don't know how much I believe that kind of data. I think it depends on what they mean by pilots and garbage and what does it mean. And a lot of people mess around with things and try it, right, the first time.

11:22And so often the AI adoption cycle to big enterprises, you have a small team that's doing almost like skunk works or research or they suddenly become the AI team. and they go and they try some stuff and they prototype it and they mess around with it and they're just learning how to use it, right? It's a new technology and it takes time to learn how to use it. And then usually there's three places you can implement it as an enterprise. You can buy something from a vendor. So you're like, I'm going to buy Decagon for customer support, or I'm going to buy a bridge for medical scribing or whatever it may be.

11:50And so then there's your vendor selection process and who you work with and all this stuff. Harvey for legal would be an example of that. there's a second set of things that you do, which is internal tooling. How do I make my internal teams dramatically more effective? How do I build apps for them internally to do things? My ops team is doing this repetitive work and use AI for that workflow. And then there's stuff that gets it into your own product that you're offering externally to your customers. Those are sort of the three levels of AI for an enterprise. And so when I see a study like that, I'm like, okay, which of these things are they doing?

12:22and what did they try and who was actually working on it and is it the right talent and all this stuff. And the reality is, if you look at mobile as an example, when mobile first came out, everybody said the exact same thing or the internet, right? The internet was this giant transformative wave. People forget that some companies literally spun out their internet division as a tracking stock because it would trade so highly. They could pull in a billion dollars through an IPO of some random piece of whatever business they had. And then when those stocks collapsed a few years later as a dot-com bubble collapsed, they rolled them back in super cheap.

12:53And so it's just a way to ARB getting some money by branding something as, as an internet company. Um, again, I think the nineties is a very interesting lesson for today in terms of what might be coming. So if you look at the early mobile wave as an example, um, the first B of A, like mobile website was really bad, right? You'd go to this janky kind of, uh, HTML site that was on your phone, on your browser, and you couldn't really do anything and it would crash half the time. And then 10 years later, you can do these amazing things in the app, right? You can scan checks and OCR them. You can take out money by tapping it against the ATM in case anybody ever uses cash for whatever nefarious things people do.

13:32Or if I'm just kidding about that, or, you know, you can send money really easily with Zelle or Venmo or other things. And so suddenly the BVA mobile app is really useful, right? But it took a decade and it just takes time to adopt these things, to learn new patterns of usage, to understand how to use it properly for users to adapt themselves to it. So I just think it's going to be one of these long arcs and to say that things should immediately be working is not reflective of any technology wave that's ever happened. And in the internet, people said the internet doesn't matter. And in the early crypto days, people said crypto doesn't matter.

14:03And people always say these things don't matter. Oh, I remember when people said that SaaS was stupid and cloud was stupid because why would an enterprise ever move their data to a third party cloud that wasn't secure and it's always always the same thing. So I just view it as like, um, there's that old saying of, um, old wine and new bottles. And I think that's what this is. Old wine and new bottles. How do you think about, and I'm sure you've talked to like many fortune 500 CEOs, teams about this. How do you think about how different companies are approaching AI and which ones are doing it right?

14:35Like, what are they looking to do? What are the key components of that? And then how do you decipher, Or how do you know which teams are just not going to make it because they're just not thinking about it correctly? Yeah, I mean, I think these things are 10-year journeys, right? So the migration of the cloud has taken 15 years for many companies. Some companies are still using COBOL and very old programming languages for the services they run, right? You'll literally find a COBOL installation running some core server for a company doing payment processing or whatever, and it hasn't been touched in 40 years, right?

15:08So I think people misunderstand the state of software and sophistication. And they may be very sophisticated. They must have these old legacy things. And so it's kind of hard to say, hey, we're just going to grade somebody on something, right? There's all these other things, moving pieces, how do they deal with data? You know, the whole migration to a Databricks would be an example of something that some companies are still doing, right? To get to data lakes and modern data infrastructure. So it's a very uneven world around these things. And so often what you need to look for is early adopters, and then you can work with them to really help implement something that's very useful for them or to build on top of this platform that, as an example, BrainCo has.

15:47The thing I found striking having worked in different businesses for a long time is in every industry, there's a set of early adopters, and they're trying to be the key influencers or decision leaders or opinion leaders in that area. So for example, if you're selling HR software, I think it's Costco and Home Depot and a few others are great at adopting new benefits, new HR software, et cetera, which you wouldn't expect, right? These are big, you know, impressive enterprises, but they often adopt things before younger companies will or smaller companies will. And so often what you ask yourself for any new product is who are the tastemakers or who are the people who are in general tend to adopt things more?

16:24And can I work with them first because they'll be fastest, they'll be easiest, they'll be the most thoughtful, they'll give good feedback. but they'll also spread it to other people because other HR departments call them for advice and view them as sort of a leader. We should talk about the$30 million Series A. How did that come together? Yeah, I think what we decided to do was there's almost two ways you see people incubate things. One is they take a lot of common stock and they try to control the company. And we did the opposite where we said, we'll just put money in to buy our ownership stakes.

16:56and, you know, the equity should really go to the employees and the founders and the people running it day to day. So we kind of approached it as a more traditional financing structure. And who was involved? Yeah. I mean, we had a great group. So I, my fund and Jared's, which is Affinity, co-led it. So we each put it in half and then we had a great sort of bevy of angels across AI, by, you know, Andre Carpathi. Can I read them off? Yeah, please. I have them. Yeah, that'd be amazing. It's a better clip from you. I was trying to tee you up for this. Yeah, please. Where? Oh, here it is. Yeah, please.

17:32I've memorized in alphabetical order where you should do it. This is going to take a while. Yeah. It's going to be hard for me to pronounce all of their names. We have the CEO of Databricks. We have Andre Carpathi. We have former chairman and CEO of Dow Chemical. We have the CEO of Perplexity, CEO of Coinbase, CEO of Applications at OpenAI, Chief Product Officer at OpenAI, former chairman and CEO of EY, global co-head of Blackstone Real Estate, founder of AngelList. That's Naval. Yeah, Naval's great. If you didn't know. Yeah, I love Naval. He's so good. Noam Brown, AI researcher at OpenAI. We have Patrick Collison, CEO of Stripe.

18:09Reid Hoffman. We have Sarah Gua. Is that how you pronounce it? I think so. I think it's Gua. Well, I've been listening to like a lot of different things and pronunciation. We should ask her. I should call her. Sarah Guo. Yeah. And then we also have the CEO of Together AI. How do you get all of these people on one cap table? Well, I mean, I think a number of people helped really pull this around together. Eric was very involved. Jared was involved. Dan, the president of the company, you know, really drove the fundraise day to day. So I think it's one of those things where people were just very excited about what we were doing.

18:45It was one of those moments where everybody you talked to was like, of course, this is needed. Of course, I've seen use cases. Of course, I've seen people ask me about this. So I just think it's one of those things where we thought it was a clear indicator of just like how interesting and exciting what we're doing is. And then as people met, Dan and Clemens and the team, I would get texts where they're like, oh, my God, these are really impressive people, right? Like, I want to back whatever these people do. So I think it's also the caliber of the people involved operating the company really matters.

19:14And with Eric Wu, are you guys going to do anything with Opendoor? Oh, you mean like a take private or something? I mean, no, but that would be interesting. I know where that question was leading. Partner with them. Maybe they'll be a customer. Maybe they'll transform Opendoor and help them, you know, fill in that market cap that they just got up to. Yeah, Eric is awesome. He's so good. It's one of those things where he's so impressive in terms of how good he is at operating things and getting things done. and thinking about things strategically. Like, you know, I backed his very first company, Movity, like years ago.

19:49And then I backed Opendoor and I've worked with them over the years. And then I started working with them on BrainCo. And I was just so impressed. I already knew him and I already knew he was very good, but he's really taken things to the next level. In today's high-speed business world, staying ahead means using the smartest tools possible, including the powerful capabilities of artificial intelligence. Meet Turing Intelligence. Turing builds customizable AI systems designed to solve your mission-critical challenges, no matter your industry. From expert guidance to tailored projects, Turing helps top companies realize AI that's more capable, more adaptable, and more effective.

20:25With Turing, discover how AI can accelerate your business growth. To learn more, visit Turing.com slash sorcery. Spelt S-O-U-R-C-E-R-Y. That's Turing.com slash sorcery. Between all of your investments and the companies that you've funded, what is the best redemption story that you've seen? A redemption story. How do you define redemption? Coming back from the ashes. Oh, I mean, I think there's a couple examples of that and it depends on what's meant by the ashes, right? There's a few founders that I've backed where there was some, you know, bumps along the way with their first company in one form or another, even if the company was successful.

21:06So an example would be Parker Conrad from Rippling. I backed him at Zenefits. I was an angel in that company. And then I invested in the first round of Rippling. And that was a good example where there was obviously controversy around the company. I always thought he was amazing. I always thought he would have some amazing redemption arc. And so I got involved with Rippling in the very first round. It's sort of a vote of confidence behind him and what he was doing. Palmer Lucky is another one where I backed Andrew in the very first round. and he was controversial given what happened at Meta early on with him.

21:37I think in both cases, they basically got thrown under the bus in one form or another. So I think they deserve that arc. So I think they've done amazing things. I mean, when I met Jared Kushner, I think similarly, the media had portrayed him in a way that was him and Ivanka, both actually in a way that was very different from what they're like in person in terms of being very smart, driven, wanting to do the right thing for the world. Um, so at this point, I think I've worked with lots of people who at one, in one form or another, at one time or another, uh, had gone a little bit through some firestorm and came out the other side, you know, stronger and more motivated and crisp on, um, what they care about and what's important in life.

22:18And, you know, cause I think, um, experiences like that for many people, and I, I, I'm not saying anything specific to any of these individuals. I've just observed in general that sometimes the controversy also just makes you crisper about what's important to you, like as a person. Most of those do run through the media cycles and the harsh reality of what can happen to you in the news. I'm curious from your standpoint, you have podcasts, you put out a book, you're like close to all these worlds, these personal stories. What do you think people most get wrong about headlines versus reality? Oh, I mean, for certain media, the headline is literally the opposite of reality.

22:57And we've seen that a lot throughout COVID and some of the claims that were made and then reversed a year or two later. I mean, literally reversed. It wasn't like, oh, we were slightly wrong. It was we were completely wrong, right, on so many topics. And so I think it's very hard to trust reasonably large swaths of the media. And you've probably heard or talked about like the Marie Gell-Mann effect. Gell-Mann, I never know how to say his name. Do you know this whole story? Okay, so Michael Crichton, the author who wrote Jurassic Park and all these things, was trying to coin a phrase that would reflect a situation where, say you're reading the New York Times and you read an article about something that you're an expert in.

23:36You're like, oh my God, they got all this wrong. It's complete bullshit. And then you flip the page and you read an article about something that you don't know anything about and you believe it. Right? But you just read something that they wrote that was completely false. So he called this the Murray-Gell-Mann effect because Murray-Gell-Mann was a two-time noble physicist. And so he's trying to choose the name of somebody really smart and attribute it to them to make the effects seem more weighty or important. So it was kind of like a branding exercise to name it after this physicist. But I think everybody has that, right?