Keith Rabois: Israel, OpenAI, Opendoor, and DOGE

23:52But like Microsoft Word has overshot the market. But they do have the distribution, which I think is the hardest part. But ChatGPT is going to be on everybody's phone or device. And like, so for example, I was considering writing a book. Everybody's always like, you should write a book. So I was like, Chet Chippity, write me a book. And it was a pretty good, you know, sort of first draft. That book's not going to get written in Microsoft Word, as far as I can tell, if it's ever written. It's crazy to think that for a day, Sam Allman was conceivably going to join Microsoft. Yeah, for like 24 hours until Vinod and some other friends intervened a fair amount.

24:25I guess I asked that to ask, did Microsoft sort of drop the ball on their relationship with it? Could they have capitalized in a much bigger way, or was it just... I don't know. So, you know, the talent, there's a complexity around talent. Like, you know, individual people, if there's only 150 people that can build research models successfully, talent's going to have a lot of power. And if the talent doesn't want to work for Microsoft, there may not be economics there. I mean, Mehta's trying this of, like, can you use economic leverage to force people to work there? You know, and we'll see what the results are.

24:56It's not a bad idea. Like, I actually think Mark's clever. He means more money than most. take the money and use it. It's like, imagine you're competing in sports. Mark doesn't really have a salary cap. So why not try to pay the highest salaries? Now, there are issues when you do that, and he's running into some of those, but it's a relatively coherent strategy for Meta to try it. I just worry, I wonder, I shouldn't even say I worry, like the moral hazard of it's like, I mean, this is kind of what went wrong with like crypto 1.0, where it's like, I'm going to give you tons of money up front. And normally the way that you build a startup is you build a startup, you work really hard, and then at the end you get the Ferrari.

25:33And if you get the Ferrari first, will you still build the hardest part of it? And like the people that are in it for the love of the game, they will. They will. But then can you tell the difference? Can you discern the difference? In sports, it kind of works where you guarantee contracts in some sports. But maybe that's because there's such statistical measurements of people's performance. And a lot of people who are in sports are just so dedicated, love of the game, whatever. They're so passionate about what they do. They can't imagine doing anything else. But I think this could backfire, but it is a coherent strategy for a company that was losing AI race to try something different, leveraging that they make probably other, you know, the Apple could have done this, but so culturally foreign.

26:12Apple could have competed on dollars. They made more money than anybody else. Like their profits are greater than like Google's revenues or something. It's something like that. I haven't done the math recently, but it's pretty comparable. But culturally, that just wouldn't work at Apple. And so it became a non-starter, but they're going to have to figure out something or they're going to be irrelevant at some point too. once the device question comes into play. Yes, well, that's what's saving Apple right now. The vertical integration required to build a device successfully is something that no other company, at least in the West, knows how to do.

26:43Because you do need software. You need hardware expertise. You also need the ability to compete on batteries and chips. And there's four or five different things you have to be pretty excellent at. But someone will figure out a solution at some point. Well, it feels like they're looking at it like, every Samsung phone when it comes out has a better camera for like a week or a year or whatever than Apple or like now the Samsung phone folds in half or it has all these advanced features that really aren't for the mainstream and Apple has always been like we're going to wait we're going to make it better but like they're not the first and like AI strikes me as a little bit different than that because number one it's software so it's not like we have to go get the world's greatest AI right now and put it in in day zero and this whole we're going to wait three years I mean, it's just such a lost opportunity.

27:28I'm convinced that, like, traffic fatalities would go down in this country if Siri actually worked. Yeah. Right? Like, there are so many things where they could make it better, but they're treating it, I mean, at least from the outside in, like, oh, this is like, you know, Samsung has a 400 megapixel camera or something. We're going to wait on that one and not get it right. And, like, they just announced, like, you know, why doesn't Siri work? It's an embarrassment in 2025 when ChatGPT is, like, AGI. Like, I mean, maybe not by, like, current researchers, but, like, 10 years ago, we would have said this is AGI.

27:56And like Siri is just as bad as it was in 2015. It's just remarkable. Yeah. Just to close the loop on this, Keith, what's your perspective on Meta's sort of durability or defensibility or strategy going forward? Well, as I mentioned, I think the approach is interesting because when you want any strategy, whether you're a venture capitalist or an entrepreneur, you want to leverage the advantages you have and maximize those. And so cash is one that Meta has. And so trying to leverage that's smart, but they still need the ability to ship products. For Google's weakness has always been, they just don't ship products.

28:34Like they have ideas, they have researchers, they have lots of money flushing around. But when was the last time Google launched an interesting product? Gmail was pretty good. Yeah, 2008. 2004. Yeah, actually, that's right. So, yeah. And, you know, when did they do Maps? Like 2005? Something like that. When did Brett do Maps? Like there's something like that. And then they acquired some things like YouTube, whatever, but launched a coherent product. Now, actually, Waymo would in some ways count, but it actually won't help with this problem as far as I can tell. But a true sort of consumer product, I'm not sure they know how to do it.

29:15Right. That's Google. And the meta of the question is, how defensible is the graph that they've built? Well, I don't think the graph is defensible. I think there are other assets that are. I think the graph has kind of been weakened by, like TikTok has kind of shown that on social products, the graph certainly is not indispensable to success and may even be a handicap. X, as you've moved from people you follow to a for you feed, which is the default and probably how all of us engage with X, has also made the graph fairly irrelevant. Well, there's a question people are asking with Sora, which is, hey, if your friends can make as good content as professional content makers, which is a big if, but maybe you want the best of both worlds, which is you want the social and you want the highest quality stuff.

29:59Ideally, but also Reddit. Like Reddit's a$40 billion market cap company, no graph. So I think the graph was very valuable, to be clear. I'm just not sure it's valuable in the future. Let's get into real estate. Let's get into Opendoor. We recently had Kaz on as well. Why did Opendoor lose its way? What were the mistakes that were made? And let's talk about how we're going to figure it out. Well, there was two. One was predictable. Vinod actually warned Eric and me about this in 2015, that real estate has a cyclical nature to it. And you need to make sure that your cost structure will be acceptable when you hit the lows of that cycle.

30:42to basically get as many variable costs built into your company culture and as low fixed costs as possible. And then you don't care. Like, so real estate, residential real estate, people's intuition on this is very, very off. At the low market, like when people think real estate's dead, 4 million homes transact a year. High in the market is 6. So you have 4 to 6 million transactions a year. You need to build your cost structure that you can be break-even or not, you know, incredibly unprofitable at 4 million transactions a year. and Eric and me collectively did not do that from 2015, 2019. So when the Fed started raising interest rates and it actually did raise interest rates six times in a very compressed period of time, which is the fastest rate of interest rate hikes, the company then went from five and a half million transactions a year as a market as a TAM to four and immediately started burning money.

31:38An extreme example of this that clarifies my point, I think, is what happened to Airbnb COVID the first month. Airbnb is a wonderful business. One of the best network effect businesses, maybe the best of all time. 13 % kind of margin, take rate, whatever. And all of a sudden, COVID happens, and nobody's traveling anywhere. Companies had a very bloated cost structure, too, but nobody noticed because they were minting money. Nobody travels. Guess what? All that G &A is pure burn, and they almost went bankrupt until they structured this very complicated deal with Silver Lake, which saved the company.

32:10Open Door had that problem, but less excusable. COVID is kind of these true black swans that you really don't have to typically build your business taking account for. In real estate, you should take account for the idea that this would happen. So anyway, basically that's what happened in Open Door. Then secondly, as Eric decided you don't want to be CEO anymore, the board for a variety of reasons promoted this completely mediocre CFO, which is almost always a bad idea, to be CEO. And she went from mediocre to like the worst CEO on the planet. So the company for three years did every possible thing you could do wrong.

32:44It still survived, barely, but like literally the dumbest possible thing. Partnering with agents, shutting down innovation, hiring people overseas, adopting DEI writ large. Every possible mistake, stupid ass capital markets decisions, underinvesting in what we call AVM, automated valuation. Every possible mistake. So the good news is you can get like 10x value just by unwinding those mistakes, which Kaz is definitely doing. You know, even his first month, I think he'll get half of those unwound, maybe even more, like just in one month because, you know, faster. But it's just like literally the more you dig in, it's kind of like Biden, you know, trying to cover Biden.

33:24It's like if I just stop doing every stupid thing that Biden did, like the world would be so much better. And like I don't have to really do any innovation. I just have to stop doing stupid shit. The immigration border is a perfect example. Like literally on day one, he's just, we're going to stop like enticing people to come here. And all of a sudden, like border crossings go from like here to here, like in 24 hours. So like, that's like Opendoor. It's like basically border crossings. Like we had Biden running the company about as competently. So what's your wish? You know, in our last episode with Kaz, Alex fleshed out kind of his Amazon for homes.

33:56He says why he got so excited about, you know, leading our investment in the company. What's your, where are you excited? Well, here's my, I'll give you the top down thesis, which I've shared publicly before, is in 2017, I had dinner with Yuri and Max Levton. And Yuri asked this question. He's like, what's the largest market cap company on the planet in residential real estate? And I assumed there must be something outside the United States. And that's why he was asking the question. Turns out at the time, the answer was Zillow at 18 billion. And it's insane. The largest asset class, period, that of 30, we've basically been innovating in technology, consumer technology, we've called it 30 years, like 1995, I'd say, to now, 30 years.

34:33that in the largest asset class in the world for consumers, that the largest market cap company that would take advantage of any technology would be the$18 billion makes absolutely no sense. Like, that's like insane. So our goal is to reinvent the process of buying and selling a home. And if you do that, you're going to be worth hundreds of billions of dollars. Now, it's difficult to do that successfully, completely reinventing the process of buying and selling a large asset, but it's not impossible. things that you know eBay did back in 1995 to 2003 apply to homes Carvana has mastered this applied to automotive which for a normal American there's actually sometimes people are like oh there's nothing like in common between autos and homes but that's often very wealthy people who have like a 40 million dollar home in Palo Alto or Atherton and they drive you know if they drive at all, like a 100K car or something like that, Tesla or something.

35:32Most Americans have a 5 to 10 ratio, more like 10 now, but like a reasonable ratio between the cost of the value of their car and their home. So you're only talking about an order magnitude difference. And I think that we should be as valuable. And some of the things we do are very comparable to what Carvana does. We actually have less competition than Carvana. Carvana actually has been incredibly successful, given that there were very significant competitors that were innovative that they were competing with. But it's a plus or minus a$40 billion company. And if we just get the same multiples on what they do, we'll be worth tens of billions to 50 billion.

36:10That's like the base case. I want to segue to fintech. You guys are among two of the best fintech investors and founders, operators in the world. What is the state of fintech in terms of where are we actively looking for investments and how we think about it, Keith, when you start? Well, I think there's lots of promising fintech companies. It's always been difficult. Like the reality is I personally have a view that to be successful as a fintech innovation, you want an underwriting advantage and a distribution advantage. If you have both, then you can build a really epic company. If you have one, you might be able to build a pretty solid company.

36:43A firm, which we were both involved in, had both, underwriting, mispriced. Certain kinds of people were mispriced. We're going to figure out a different way to underwrite based upon other things. and a distribution, a very clever distribution hack that other people have now copied, but basically was very innovative at the time. And that's why a firm's, you know, $25 billion company was still upside. But I think those are rare to find. If you find them, the world's still open for innovation. It's not like the world's incredibly efficient around financial services. We have some very successful ones that have been built over the last five years that now need to infuse more AI, but they're not predicated on AI.

37:19And I actually kind of like that because you could argue there's an AI overinvestment sort of cycle going on. And arguably one way to kind of create a portfolio that's not as sensitive to AI valuations is financial services innovation, which still requires distribution and innovation around underwriting. There's also this kind of question of like, there's bits and atoms, like you mentioned, Vommel's cost disease, right? It's like, you know, plumbing, healthcare, like all of these things that are so atom-centric that will be atom-centric for a long time until we have sentient robots or something.

37:52Whereas every financial services product that I know of, except for like the delivery of physical gold ingots, like it's all bits. So there probably is a fair amount of upside for AI, but you still have to get the distribution, which is really hard. Like consumer fintech products are so hard because like all of the, I mean, right now just all of the economic rent goes to Google or wherever you find it from. But for the ones that unlock distribution, it's like number one, you can bring down the variable cost that you were talking about with AI, and then number two, it's like there are, I think there are so many things right now in the world, this is kind of my view on AI, where it's like cost is here just for any object, any delivery of service, and value is here.

38:31So if you just keep value here, like you're not trying to change the value equation, but you can bring the cost down, now you've unlocked a new market that just didn't exist before. And it's almost hard to articulate what those are until you see them. But like, it's like, oh yeah, like people wanted that. It's like, you know. Well, Square was a quintessential example of that. Yeah. Everybody thought there was no market for Square. Like, including very smart people, like lots of very smart people that we both know. Because you couldn't really prove that everybody really wanted to accept like credit card-based payments.

39:01And so you actually gave them, real people, the option to do it. It made it so simple, easy, intuitive that they didn't have to think about it. And then it's a massive market. Right. Or like, so I'm on the board of a company called Wise. And Wise has shown that like the market, this kind of makes sense in retrospect, but like the market for real-time delivery of cash is just so much bigger than the market for like, you get it two days later. Yeah, well, Cash App, this is the monetization strategy behind Cash App. Cash App makes money now. The reason why is people will pay for instant quote-unquote delivery.

39:33This has always been true in financial services. By the way, this is the magic behind PayPal's revenue model. We basically created an instant debit, effectively an instant debit or ACH product back before people knew that language. and then we just charge the fee for the instant receipt of the availability of the capital. It's like FedEx type. FedEx to financial services. It's like, you know, Frederick Smith proved this with FedEx. Yeah, versus the UPS. You pay 32 cents versus$10. It's probably already overnight. It's like some people need it before 10 a.m. And like it turns out that's true for like a large, like countably infinite number of things.

40:06But I think just in general for like fintech, there probably are going to be a lot of other services. Like the banks, I love something that you tweeted. I think it was your pinned tweet for a long time. It's like how to disrupt a market. Pick something with very low NPS, and you don't get better by delivering one particular service of that. It's like, oh, here's a camera that attaches to your phone. No, nobody wants that. You want the better, fully integrated service, so you vertically integrate, and this was Opendoor. You could just say, here's a better AVM, automated valuation model. Okay, if you really want to get the whole value, I mean, would you rather be, at the heart of it, It's like, would you rather be Amazon, which has a very low gross margin on products that they actually, you know, store and sell, or would you rather be eBay, which has amazing high gross margins but is the shittiest service in the world?

40:53Nobody wants to buy from eBay because you don't know if you're going to get the product on cloud. You don't know if Eric Thornberg is going to ship it to me. Like, you want to buy from Amazon, and Amazon is worth$2 trillion, and eBay is worth, like,$40 billion. So, like, obviously Amazon is better, but you would misjudge that based on the gross margin classification. Like, oh, VC is only, like, doing high margin things. And like, I think that's one of the things that I really respect about you. It's like, do something really hard. It doesn't matter what the gross margin profile is. It matters like how much gross profit are you generating in the net.

41:21This is the worst thing. There's lots of bad things you learn as an MBA. Like, wow, it's like in the laundry list. This is the worst one. It's like the gross margin, especially percentage. It's one of the reasons why, A, Amazon is successful. Stripe actually back in the day took this model and back before more people understood this and said, we only care about the sliver of contribution dollars. not like the percentages. And everybody else was like, oh, you got to worry about this margin or that margin percentage. It's just like, are we adding incremental layers of dollars, actual cash? And that actually works really well.

41:52Yeah. And so to the extent that VCs have soured on the, or been not as excited about the category in recent years. Just send them all to us. Keep being soured. Financial service, terrible innovation. It's a very small market. Nobody likes it. Well, this is the point that I was thinking. It's like, how many people like their bank? And like a lot of these companies actually work because of regulatory capture. Yeah. Right? Like if it was very easy for Keith and I to start a bank tomorrow, yes, distribution is very hard, right? And underwriting is very hard. I'm pretty confident, at least in your ability and to a lesser extent mine, like we would be able to get, we could build a very valuable company.

42:26But it's so hard. Like the regulatory overlay on these things is so high. And that's part of what makes it challenging because nobody likes Chase. Yeah. And there's a saying that I love, which you've heard me use a lot, which is the best companies have hostages, not customers. And, like, all the financial services companies, all the incumbent financial services companies, they have hostages. They don't have customers. And it just, it turns out, it's very, even if you build a better product, a 10 times better product, it's still, like, the distribution is very, very challenging. But this is what the killer entrepreneurs can do.

42:54It's like you find, and this is why, like, you know, Nubank did this for Brazil. Like, in many cases, it's geo by geo just because the regulatory overlay is so high versus you build a software product. Like, Microsoft Excel is Microsoft Excel for the world. There's no, like, Portuguese version of Microsoft Excel that only is sold in Brazil and Portugal. Yeah, but this is a great opportunity. One of the best investments I've ever made is a company in Europe called Trade Republic. It's actually better than New Bank on any metric, actually. But it could only be done in Europe. It's not easy to translate what they do and why it's successful here.

43:28Robinhood's a poor comparison here, but, like, is the closest. But so you have to kind of arbitrage the GO, too. because it's because of the law and regulation. There's things that are easier actually to do in Europe. There's a reason why Trader Republic's better than Robinhood because in Europe, actually, it's more permissive certain things. And then there's things that are much easier to do here. And so you need someone who understands how to create a product and value prop given the constraints and where are their gray zones and how does it translate to customer value. And so that's very difficult.

43:59But there's a lot. There's a modern generation of financial service innovation. It's a little under the radar. I mean, people definitely are paying attention to RAMP. So, you know, like high profile company, but even a company in our portfolio that I've invested in many times called Avon, most people still don't really know about. It's a great company. Another company is called Imprint. Great company, still fairly under the radar, competes with two public companies. Actually, we sort of helped seed fund it back at Affirm. Back when they were starting the company, Affirm did a corporate, you know, sort of investment.

44:29Cool. Well, and that's the thing. It's like, you know, software, Like, I'm very confident that all of the banks cannot build software, will not be able to build software, will not build software. Let me use Devin. Yeah. So that's their best shot. That's their best shot. But they can't really, like, if it turns out that software is at the core of everything, if software is going to eat the world, like, why hasn't it eaten financial services? It has around the margin. It's like the cool thing for Ramp is, like, it's like a net new market where it was kind of like there was a shitty version of some corporate card something something.

44:56But it's like that wasn't the software. That was a financial product. Yeah. Once you marry it to a software product, it's 10 times better. Like you have vertically integrated that. And the incumbents can't do this. They're still like literally in COBOL often. And their ability to attract talent is just like zero. That's why Devin is actually arguably over time a save for them because they won't need to attract talent. They can use software to compete with software. And the software can write COBOL. Yep. We'll take the conversation full circle. We were talking about real estate. A real estate expert solved the Middle East.

45:29What does it say about the nature of expertise that, you know, a guy who is a tycoon in real estate and private equity, but didn't have a ton of, you know, foreign policy expertise, you know, initially or a trained background in it could be more effective than, you know, trained experts? Well, at KV, you know, from Vinod on down, we don't believe in experts. You know, Vinod has a whole speech about how no expert has ever created fundamental disruption in any field. You could argue that maybe in the last 90 years, there's one or two exceptions. But the fact that you have to argue about the one or two exceptions suggests he's right.

46:02So, like, for example, I don't like experts. Like, hiring domain experts in the companies I work with is pretty much a non-starter. At PayPal, we had, of the 254 employees we had in Mountain View when we went public, two maybe three had any expertise in financial services at square if you included me which is i'm certainly not a prototypical financial services guy of the first 300 employees maybe two had any financial services experience so it's a pretty common formulation airbnb i'm pretty sure brian nate and joe did not have any hospitality experience so i think if you're going to reinvent in industry you don't want experts so the same thing is true in politics if you want to solve a problem that's been intractable, at least since World War II, arguably for hundreds of years, but at least the modern era have been intractable.

46:50You probably don't want anybody who's got expertise in the Middle East. You know, as Jared, you know, was maligned for, he just read books, which is a great idea. Actually, I talk about this all the time. Like, people are like, where do you get your contrarian ideas? I was like, I read books. Like, there's actually all the greatest thinking of all time is available to anybody. You just have to read. And the more you read, the sharper you are because what it what jared talks about and i think is the common formulation we definitely apply this at kv specifically explicitly is asking the right questions now so what books prepare you to do is to ask questions so when jared went to the middle east and got dispatched the first time he actually asked the people who are influential and important in the middle east like what do you care about and why once you understand that you can navigate and find solutions and that's the same thing we teach all of our investments at kv is if you ask the right questions, you'll make the right investment decisions.

47:42So when people join the team at KV, we don't actually look at their output, like, are the companies great? We look at the quality of their questions. Vinod is adamant about this, been adamant about this for like 40 years. Just the quality of questions tells you everything you need to know. That's a good place to wrap. You can just do things. You can just read books. Keith, thanks so much. I'll give you book recommendations next time. Amazing. Perfect. Awesome. Thanks for having me. Thanks for listening to this episode of the A16Z Podcast. If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family.

48:17For more episodes, go to YouTube, Apple Podcasts, and Spotify. Follow us on X at A16Z and subscribe to our Substack at a16z.substack.com. Thanks again for listening, and I'll see you in the next episode. As a reminder, the content here is for informational purposes only. should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see A16Z.com forward slash disclosures.

49:04Thank you.