The Exchanges

Q Keith, did you not want to roll in the administration yourself?

A No, you know, the thinking, I love politics. If you, you know, follow my Twitter feed, I pay a lot of attention. I used to be involved in politics before I got into tech. However, what I realized about where I am in my career in tech is if I stop doing what I do, I'm never going to come back. Like technology is rapidly emerging. We're going to talk about all the latest developments this week. Like you can't take your foot off the gas in the network building parts of venture for two to five years and come back when you're like 50 years old. And so I felt like I'm not ready to give up on venture. I'm in like the prime of my venture career. I'm only 12 years in actually, Chamath. So figure five, 10 more years, like is the sweet spot. And so I'd like to see the companies that I was involved in grow up, become public companies, et cetera. And I didn't feel like I could ever come back if I quit at some point when I like to get involved in politics, probably yes, but it's a decade out.

AI assessment note: “I felt like I'm not ready to give up on venture.”

Q even Trump himself flip-flopped a little bit on that. So what are your thoughts as we get started here, just on that assembly of people, you know, including Saks, obviously who couldn't be here this week, but will be on future episodes. There's your announcement, folks. What are your thoughts on that? The sort of diversity and opinion in the administration and how that all sorts out? Cause some people.

A I think it's extremely exciting. I think it's very obvious watching from afar that the way Trump's makes decision is he likes to ask a lot of people, a lot of different questions, and then he makes the decision. That's why he's to some people, the media and he's very unpredictable is he doesn't just take one source of input. And so you can never totally predict the output, but he arrays an interesting cast of characters and listens to them. Like, so for example, I haven't spent that much time with him, but insofar as I have, he would go around the room and ask every single person at dinner, what's your view on X and literally go around the room of 28 people and listen to every single person. So I think that's how he makes decisions.

AI assessment note: “he likes to ask a lot of people, a lot of different questions”

Q column of somebody who wants to allow business to occur and wants a free market. He wants to do basically everything Lena Kahn, uh, didn't do, which is allow some M&A. Raboy, what's your pulse like here, uh, And what's your take on the venture industry? And then we'll go into M&A and exits. We'll start with just investing. Is it, is it ramping up? You're seeing high quality companies.

A Well, I'd say that you should cut that data by AI and non AI, and you might see a tell to different cities. My anecdotal experience is there's AI companies where the market's pretty hot, maybe cooling a little bit, but hot and AI companies with the right team are getting funded, you know, frequently, quickly, et cetera. And then there's non AI companies. And I think you'll see a very different, you know, sort of chart there. I do think net, net, you're probably at a city state that looks reasonable across 40 years, you know, like, et cetera. But, um, it would be interesting and, you know, you have to make some methodological decisions about what's an AI company, what's not, but if you could do that, it'd be interesting to see if the lines, you know, look similar or not. But it's pretty hot. Like people starting companies, founders are optimistic. Crypto companies also, you know, are back in vogue, obviously due to the change in administration. I think a lot of people had been hesitant to start new crypto companies and, you know, there's a belief and confidence in the new administration, the SEC, et cetera. So we'll see if the innovation accelerates, you know, with all the new capital and all the new founders back in crypto. In enterprise software, It had been pretty cool, non-AI based enterprise software, but, you know, like, as you mentioned with the IPO this week, trading ver…

AI assessment note: “My anecdotal experience is there's AI companies where the market's pretty hot”

Q you know, AI being a disruptive force at the interface level, i.e. your phone or desktop. What are your thoughts on, uh, This new story about them doing more chips. They've obviously had great success with the processors and phones and now the M four incredible. If you haven't tried the Mac mini best computer for the dollar in the world right now, but what are your thoughts on Apple?

A So the most important thing about Apple is to remember it's vertically integrated and vertically integrated companies. When you construct them properly, have a competitive advantage that really cannot be assaulted for a decade, 2030, 4050 years. And so chips, classic illustration go all the way down to the metal. Build a chip that's perfect for your desired interface, your desired use cases, your desired UI, and nobody's going to be able to compete with you. And if you have the resources that, you know, because you need balance sheet resources to go in the chip direction, um, it just gives you another five to 10 year sort of competitive advantage. And so I love vertically integrated companies. Uh, you know, I posted a pinned tweet. I think it's still my pinned tweet about vertically integrate is the solution to the best possible companies. Uh, but it's very difficult. You need different teams with different skill sets, and you need probably more money, truthfully, more capital, but Apple's just going to keep going down the vertical integration, software, hardware, you know, all day long, and there's nobody else who does hardware and software together in the planet, which is kind of shocking in some ways. Is there a world-class company, a company that's world-class, it's both software and hardware other than Tesla. Yeah, maybe.

AI assessment note: “chips, classic illustration go all the way down to the metal.”

Q Can you tell us what your point of view is?

A Well, I mean, I tweeted in October, but, uh, you know, that we were at the height of the market. I tweeted last January that we're going to see 2000 all over again. And so privately, internally, I've been arguing this internally that this is exactly what's going to happen. And so, you know, my behavior should reflect my views. I believe in some consistency and harmonization. So if I believe tech stocks and tech companies aren't worth that much, I can't be investing until they reset. And so I don't want to spend money and invest in companies that aren't going to make me money. My job is to ultimately return billions of dollars To my LPs. And if I can't do that, I shouldn't be giving anybody any money.

AI assessment note: “if I believe tech stocks and tech companies aren't worth that much, I can't be investing”

Q Keith, why did you leave Founders Fund to go to COSA?

A So COSA was great. I spent six years there. The truthful reason why I left, it's kind of funny given COVID and how history changes, I hated commuting to Sand Hill Road every day. We were one office period in office every single day. And I felt like the future of investing was more in San Francisco than in Palo Alto at the time. And I just despise sitting in a car, 45 minutes each direction. Turns out, you know, COVID changed everything, how people do their work, like we're recording this by Zoom. Before COVID, we'd probably all be in the same studio recording a podcast like this. And so, but Vinod and the team was very inflexible about it, and Founders Fund was located in the city. Obviously, I knew Peter since college, as Jason alluded to, and I decided that, you know, it was better for me. I remember talking to Sam Altman, And I said, am I crazy for changing funds mostly on a commute basis? And Sam said, you're human, and every single study of human happiness is, it's inversely correlated to your commute time. It's like, there's nothing wrong with being human. In any event, there are a lot of similarities between FF and KV. Both are great funds that have put up, you know, incredible returns, have funded iconic founders and companies, but they're very different. KV is involved as early as possible. And FF is a momentum investor and is maybe the best on the planet as at being i…

AI assessment note: “The truthful reason why I left... I hated commuting to Sand Hill Road every day.”

Q Keith, do you think that the Pegasus spyware that can infiltrate WhatsApp so that you can turn on the mic and listen remotely And it has no fingerprints. It's very difficult to detect. Do you think an equivalent, let's call it a backdoor, exists inside of TikTok?

A Yes. So my evidence for this that I believe is an interpolation of what's in the public domain is if you looked at that vote to ban TikTok, it was extremely bipartisan despite, you know, controversy. And what happened was that vote was taken a week after there was an intelligence briefing to both the House and Senate Intelligence Committees that was, that's confidential. And the votes, you know, all of a sudden flowed through. I think there's things that are not in the public domain about TikTok that spooked a lot of elected officials and led to this bipartisan consensus. How many bipartisan votes do we see on an allegedly controversial issue that, you know, basically the vote was like, you know, like not even close in either house.

AI assessment note: “Yes. So my evidence for this that I believe is an interpolation”

Q Why? Explain to people who are like, it's just an app. It's just an app. Why? Why?

A Why? So I think there's different dimensions. One of the problems is there's so many things wrong with TikTok that actually sometimes people get confused because there's not just one. It'd be easier. Sometimes it's just one. So they are definitely using the app to track data about Americans. And there's evidence that regardless of what alleged protections exist, There are people in China monitoring what certain people in the United States are doing, and the CEO lied under oath to congressional committees about this, and the evidence is now in the public domain. Hopefully, this Justice Department will prosecute him for lying under oath. I think setting a really good example that you cannot just purge yourself before Congress.

AI assessment note: “they are definitely using the app to track data about Americans”

Q Walk us through that. Why? Tell us about that. Brain damage.

A So typically if a found, you think there's an efficient market of pricing, right? Like I need this much capital and the market's going to float with the price of that capital is. In private capital, it's not really true. Like, so if someone comes to me and says, you know, my last round was done at three hundred million, you know, nine months ago, and today it'd probably get priced at, let's say, a hundred twenty million. I'm more likely to say no than to give them an offer at a 120 because I know their prior investors and their prior employees are going to be mad at me and furious at me. And I don't want a lot of founders and people annoyed at me. Um, and so that brain damage isn't worth it. So I'm more likely in our fund is more likely to say no than try to find whether 80, a hundred, a 120, a 140 is the appropriate price, which is very bad for the company in some ways because you, They, they might need the capital.

AI assessment note: “so that brain damage isn't worth it.”

Q Well, the household's involved, uh, big announcement. Your husband, Jacob, is joining the administration. Can you tell us a little bit about that? Obviously, we're all very proud of that.

A Yeah, it's extremely exciting for him, obviously, for the country, I think, which is, he's going to be the chief economic officer, really, for the, for the, for the country. His job is to build foreign policy from the business standpoint. Which if you think about it, what's the foundation of power in the world? It's economic success. Why did the United States win World War II? Is we had an economic engine that could outcompete Germany plus Russia plus Japan. We could build more tanks, you know, blah, blah, blah, blah, et cetera. And so the economic engine is critical to this administration. Obviously, Trump understands that. We had a great three years under his first administration, as he likes to say, best economy ever before COVID, which may or, you know, may be true. And we need to rebuild American strength. And Jacob's job is to export that, you know, philosophy. And sometimes you can build economic strength through working through foreign affairs. And so that's his main job is to be the primary point person, undersecretary of economic affairs. And then they've got a bunch, the Democrats and the woke people added a bunch of other things to the title. It used to be just undersecretary of state for economic affairs. Um, and they added like a, You know, environment and all these politically correct things. So hopefully they'll subtract all that stuff and just go back to unders…

AI assessment note: “that's his main job is to be the primary point person, undersecretary of economic affairs.”

Q lot of product work. I'm curious your thoughts on it, is you don't have to do this in the cloud. You're authenticated already into a lot of your accounts, nor do you have to worry about being blocked by these services because it doesn't look like a scraper or a bot. It just, it's your browser doing the work. Your thoughts on this? Have you played with it at all?

A Yep. I think it's a great Hail Mary attempt by perplexity. I think opting something like this perplexity is toast. Like I've heard a stat about ChatGPT is going to a billion users. Like, it's becoming the verb, you know, the way you describe using AI for a normal consumer. There's nothing left of perplexity if they can't pull this off. So it's a great idea because, like, the history of, like, consumer technology companies is whoever's up, has uphill ground, like, in a military sense, whoever's first has a lot of control. This is actually what Google should be doing, truthfully. Like, I think Google's also, Google search, cross-search, is toast. And since they have Chrome and they theoretically have a quality team in Gemini, they should be putting these two things together and hoping to compete with Chad GPT. They're going to lose the search game. Like the assets that are best at Google right now have nothing to do with search. It's every other product is the only thing that's going to save that company if they can figure out how to use them.

AI assessment note: “I think it's a great Hail Mary attempt by perplexity.”

Q What do you think about all this crazy crypto going on here, Keith? Everything's spiking back up. XRP, Bitcoin. What do you think of, uh, Sailor? I know the Sailor fans are like obsessing about him coming on the pod. What do you think of Sailor? Taking these loans to buy Bitcoin and then other people are following suit to put it in their treasury.

A Well, short version on crypto generally is I still think the primary use case is speculation and not that there's anything wrong with that. People speculate on stocks, retail trading, people speculate on football games, gambling, like the natural tendency for humans to want to speculate is very real. The, the art to me is because everybody's now a node connected into these crypto, let's call Bitcoin, you know, Can you build an application on top that would have too much inertia, too much friction to start from scratch, but it has real value. And I think it's possible because so many people are now connected based upon speculation initially. So that's what's exciting to me. But, you know, we've watched speculation before in crypto markets and, you know, it's very volatile. So hopefully someone builds real layers of Productivity or utility on top, but takes advantage of the network. It's a little bit like back in your days at Facebook, Chamath, you know, you had the social graph and you'd say, you know, you can build applications on top of the social graph, but if you had to create a social graph from scratch, that was a heroic effort and nobody else was going to do it. That was like literally the Facebook monopoly story. And so I think that's true of crypto. There's lots of nodes. Someone's got a social, your social graph.

AI assessment note: “short version on crypto generally is I still think the primary use case is speculation”

Q perplexing to people in finance. One is tariffs and how that's going to work, and the other is inflation and what the impact would be on fifteen million people being shipped out of the country and the impact that would have on Inflation vis-a-vis the unemployment rate getting even lower than the historic low it's been in our lifetimes. What do you think of those two issues? Take either one.

A Well, I think, first of all, the treasury secretary understands this. Like, I think Scott, actually, from everybody I've talked to, I don't know him, but I, I've interviewed about 10 or 12 people that are very successful on Wall Street, and every single one of them universally has a claim for him, and they've all made points to me that these kind of subtleties, he really does grok. And so when he, when he says you can raise tariffs without inflation, he understands how all this substitution works and is running an equation in his brain. He's a very successful trader, and that's a really good skill. Also Chamath's earlier points about Darwinistic evolution of brilliant people being successful, understanding how the economy actually works. He's perfect. So I think you're going to see real tariffs, like, especially against China, And the Trump administration is completely committed to reducing inflation, the cost of eggs, you know, and groceries, and I think those things can be reconciled. I know amateur economists, you know, with random degrees from Wharton don't think so, but as Trump pointed out in the debate and, you know, JD Vance pointed out with some research, his first administration imposed all these tariffs and there was no inflation. And then JD Vance pointed out that the Fed Reserve has a study that says, you know, housing prices, which are primary driver of affordabil…

AI assessment note: “his first administration imposed all these tariffs and there was no inflation”

Q What, how does it save the world, do you think?

A Well, we need a free speech platform where people can debate ideas, and the left wing of Twitter, the employee base, has completely suppressed ideas. For example, my husband, I happen to know this, wrote an article in Foreign Policy magazine, like, the most prestigious publication in the entire planet for foreign policy debate about the CCP. Twitter refused for years to allow them to advertise that article published in foreign policy magazine. So there's clearly someone at Twitter suppressing content that's critical of the CCP. And we tried appealing to everybody and they wouldn't change this. So there's either Chinese spies there or a left-wing culture that, you know, suppresses debate. This is foreign policy magazine. We can't get any more prestigious than that. It's absurd. Let alone the fact that I have 300,000 followers and do not have a blue check. I must have the largest follower. Of anybody who doesn't have a blue check, and it's all because I have views that are unacceptable.

AI assessment note: “we need a free speech platform where people can debate ideas”

Q How does that happen, Keith? How do you, you're at Founders, sorry, and then you get What, like, what pulled you to go and work with Vinod, and then what pulled you back? Like, how does that process work? Because these things are typically meant to be sort of forever jobs.

A That's true. So I, you know, had the benefit of having Vinod on the board of Square, which I think is typically how executives, uh, wind up turning into VCs, is you forge a relationship with board members. So like, for example, Roloff Bote, who runs Sequoia, had Mike Moritz on his board. Roloff was our CFO at PayPal, and Mike recruited him. So that's a very common, same thing, Ravi at Sequoia today was the COO and CFO of Instacart. Again, same thing, Mike recruited him into Sequoia. So I think that's typically how people become VCs. I always knew I wanted to be a VC since 2003. I was a very active angel investor, as you know, even when I was, you know, concentrating on all these other jobs, I was writing a lot of checks. And so anybody who's writing a lot of active angel checks probably has the back of their mind. One day I might want to be a professional investor. We could talk about the merits or demerits of that, but like the goal was pretty clear in my mind. And then at some point, I think you have to make a decision. What do you want to be in life? Um, you know, venture has long time horizons. It is a job for life. Like 1520 years is pretty much what you have to commit to. So you don't really want to start venture when you get too old because 20 years, you know, you'd be like Donald Trump age. Yeah. So I can run for president. I can run for president in 25 years or somethi…

AI assessment note: “having Vinod on the board of Square, which I think is typically how executives”

Q Well, it's also very hard in some circumstances where a car is driving down the road, and it's learning in that process, which is why you need a safety driver, and I think Elon made the right decision to put one in there. Keith, your thoughts?

A Yeah, a couple points. It's not quite that binary, Chamath. I generally agree with your arc, but like, if you think about LLMs being the most important unlock in AI, LLMs are all trained on human writing. So someone wrote every piece of data that every LLM used a human wrote at some point in history. So, yes, it's true that they've shocked everybody, including OpenAI's, you know, original team, on the implications, the broad implications, the general applicability to almost every problem, but it's not like there was some tablets floating in space that weren't drafted by humans that we've trained on. As you get in non-LLM based models, you may be totally right, but almost no one's really using non-LLM based models at scale. I'm driving specifically. Travis is totally right. The humans are actually really good drivers, except when they get distracted, they get distracted by drugs or alcohol. They get distracted by being tired. They get distracted by turning the radio. They get distracted by chatting with their passenger. So trading against human behaviors actually turned out to be a great decision because what, for whatever sort of Darwinistic reasons, humans are pretty ideal drivers. And so you don't have to reason from first principles. This is a much better path. And I think, again, there may be a, A broad, uh, sort of lesson there. The most important thing I think as a VC tha…

AI assessment note: “On driving specifically... trading against human behaviors actually turned out to be a great decision”