E66: Why Inception Investing Beats Traditional Seed with Ed Sim of Boldstart

22:28I mean, the height of Zerp, Ford multiples were at 50 times. Now they're at like eight times. And maybe for the best companies like Palantir, they're at 20, but let's just say they're 10 to 12 times for next year's revenue. Think about that. And once upon a time, we're like, if we get a billion dollar company and we own a big chunk of that, we're going to generate great returns. Then it was like during Zerp, it was like, oh my God, these companies can be 20 or 30 billion. Let's model it that way. And then now moving forward, I think we're back to an environment that is kind of more sane and rational.

22:57And what it means is that because these multiples have been compressed, and maybe some of these valuations are a little bit higher now. If we look at the numbers, they are actually going up as round sizes go up. We feel like we need more shots on goal because the idea of having a 5 billion or 10 billion dollar company, maybe you need two$5 billion companies or two$4 billion companies versus one$10 billion company. So you need more shots on goal. So how we've modeled it is we still want core portfolio companies to run the board. So we move that from 20 to 25. However, we put what we call discovery checks, which I mentioned earlier, we're going to do 12 to 15 of those, right?

23:29So let's just say you 15 of those checks. And let's just say a third of those graduate to core. So you end up with about 35 portfolio companies, which is almost double the shots on goal that we had before. And I think that's important is that we can find the balance of being on the boards, but also have the balance of making sure we have enough shots on goal because I just don't know what's going to work. We pray that everyone's going to work, but not everyone will. So you want to make sure that you have the shots. And so therefore, that's the adaptation that we've made around it. And you've got to constantly adjust because if you just keep doing the same thing over and over again, you can get your ass wiped.

24:02I'm just telling you. Yeah, you alluded to earlier the fact that it's hard to return capital or a lot of capital in this business, especially now that there's kind of a flood of capital. Let's talk just briefly about the asset class or zoom out. Obviously, there's a lot of capital that is flooded in. I'm curious how durable or sustainable you think that capital is. And let's say we're 2024 right now towards the end of it. if we're talking of five years now, 10 years now, is there more capital in the ecosystem? Is there less capital in the ecosystem? How do you think the ecosystem matures? Does it continue to have this barbillification that you described where the multi-state aggregator firms just get bigger and bigger and bigger?

24:43And then there's this increasing niche and you're either in one of those two poles and there's sort of LPs that differentiate accordingly, or is this sort of like LPs that are putting money into the big aggregators, but not getting as big of return, even though they're putting a lot of money, is that unsustainable? Yeah. So great question. So, and I think about it a lot. So let's just start with the first part. I've been doing this for 28 years. And frankly, I think this is probably one of the greatest platform shifts I've seen in history. I mean, I have more energy now than I did even when I saw the first internet boom.

25:16And think about this, at the height of internet and telco kind of bandwidth investment. It was about$120 billion in 2000, which is about$200 billion in real money terms. We're sitting here in Gen AI, the same thing's happening,$200 billion from the big fives, converted to 250. And we all know how much value up the stack was created. And what I really think, by the way, is that yes, venture actually dollars will actually increase because there'll be so many freaking losers. The losers will outnumber the amount of winners, but the winners that win are going to make so much capital and then dwarf the amount of money lost.

25:49so the capital is going to keep coming in the venture side. As far as how people think about the world, I think you really nailed it. And I mentioned earlier, it's go big, go niche, or go home. And I think either you... Actually, if I'm an allocator and I have a hundred billion bucks, I'm going to want to make sure I invest in a bunch of the platform plays. It's an easy way to get a hundred million dollar check in,$150 million check. Get exposure. I know it's not true early stage anymore. It used to be early stage, right? They used to be 10 years ago when I started, they were early stage. They're not early stage anymore.

26:16They're a multi-strategy fund, right? In fact, the growth investments are running their bigger checks in the portfolio. So maybe that's a premium on PE. So then the question is, how do you go earlier? And the question there is, if you go with funds that are sub-150, then sure, you're probably going earlier. But once again, you're going to miss out on that stage that I talked about, the jumbo stage. Let's just say you've got to be, in my opinion, a 150 to$300 million fund to do those jumbo rounds. And so then you're saying, okay, let me go fund these smaller funds that are in the$50 million range, maybe they're tagging on to other people's investments, right?

26:51Or maybe they're leading some very, very small things. But I think you're eventually going to need funds of our size. You're going to have to look at the 150 to 300 range and say, holy crap, you know what? I don't want to lop off a third of the greatest founders in history that could actually start companies right now. It might be second-time founders or a third-time founder. And so I think that you're going to see this stratification. And then within the early stage bucket, you're going to see kind of these smaller funds where you're just kind of tagging on. And then you can look at the funds like ours that focus on ball control.

27:20It's all about ball control, by the way. When you meet these founders, I don't want to wait for someone else. If I see something I love, whether it's a $20 million round or a$3 million round, I want to make sure that I can write the check. I can move fast enough and write the check that's big enough to lead that round, get the ownership that we want and kind of move forward. So I think you nailed it. That's what's going to happen. And I can see it happening now. I mean, people are coming to us LP-wise. And by the way, LPs never used to come to us. So I just want to be really clear. Even from 2010 to 2019, it was brutal.

27:48I can tell you stories all about that. But LPs are coming to us like people that never would have come to us before saying, I've got exposure from the multi-stage firms and yet I want to get early, but some of these other early funds are too small. But you guys are just been round long enough. You're at the right stage, right size. We're kind of like Goldilocks in a way. You know what I mean? It's the right porridge at the right time, but it took a long time to get there. And what size is too small or what's the outer limit? I wouldn't say too small, but the question I have is too small to lead the jumbo rounds, right?

28:15And so that's the question. I tell founders of VC funds, do you want to do those rounds or not? And if you do want to do those rounds, then do you feel like you have the brand to do those rounds? Do you feel like you have the expertise to lead those rounds? Do you want to put on your helmet, your cleats and whatever, and go head to head on a daily basis with some of these funds with the hopes of either winning or potentially getting a split with them. And if that's a model that you want to actually go under, then go for it. If not, and by the way, if you can raise the capital. And if you can't, then I would stick with the classics and the discoveries or maybe latching onto some of these other bigger rounds and just write smaller checks.

28:49But that's a decision. That's a really big decision people have to make. A lot of people aren't willing to make that or afraid to make that decision, but we've decided we're doing it. We're playing ball and we're actually doing okay so far. We'll see. We'll see if I still have my hair in five years. And the way that you justify doing these rounds at 50 or 100 or a post is basically, these are repeat entrepreneurs. This is enterprise. It's not consumer. It's not like a ton of randomness in it in the same way. And it's just de-risked based on their pedigree and based on the space that it's a safer 10X, or even if the upside is somewhat bounded, given it's not at five posts, it's at 50 or 100.

29:28Hopefully, they're less than$50 million. I would say those are the rare exceptions that are above 30. I'm calling jumbo rounds. It could be six or seven or$8 million rounds. They're typically 30 and below, but let's just say there are some that we've paid higher. The question for us is, what is the potential outcome? I talked about earlier, the most you lose is 1X. What is the undounded opportunity moving forward? Are they really going to transform the world? Do they have a better shot of doing it because of who they are and how they're thinking about it? What do they learned from their past experiences that's going to drive them to move faster, build faster, to attract a better team.

30:00And so that's one, how big can it be? And then two is it's going to be like, do I want to play in that game or do I want to play in that space or not? Because it's going to be massive. And I think it's going to drive the returns of the fund. Let's think about it. People really want to have from early stage, like a 5X gross, 4X net for something like that. So I think something can even deliver 20X or 30X, even though the price is higher, then I'm going to go do it. If I have faith in that founder, if I've worked with that founder before. The second thing is that the capital they raised today, it's not like they spend all of it tomorrow.

30:30Just to be clear, I'm not funding foundational models where the money goes all to NVIDIA because guess what? You can just invest in NVIDIA if you want to. You don't need me with the fee. But the money's there. It's not like they're spending it. I literally got this board deck for tomorrow. The founder had raised 10 million bucks and they're spending 160 a month right now after the last year and they have a team of six engineers. It's not like they're looking to ramp it up. Why did the Devil's Advocate, why did they raise the money then? They raised the money because they actually were very successful founders and they felt like the type of team they wanted to hire, they wanted to make sure that the runway was available for them.

31:06And it's like, hey guys, at least I raised$10 million. I have to risk you because a lot of the best talent now, I mean, do you see the numbers right now? I mean, opening eyes paying$500 ,000, $600 ,000 for people. I'm not saying that we want to pay that number, but at the same time, you have to have more than three or four million bucks for someone even to jump in because now you've basically collapsed. If you think about the rounds, you've collapsed the first check, the inception round, maybe the post seed round and part of the A all together into one stack. And you have to believe, by the way, if you do this jumbo round, that that company, if they're at the 75th or 90th percentile valuation pre-money wise, that they're able to hit the 75th or 90th percentile in the A and the B round.

31:45I'll give you a good example. Protect AI certainly did that. We came in it in the 30-ish post range. Next round was at like 116 posts. The next round was at 460 posts. So they did hit their milestones. They do have the customers. They have hired the people. So you really have to believe that this team will be able to execute and maintain that curve and be on that higher path across the board. If they're not, or if you have reservations about it, then you shouldn't participate in it. Yeah, that makes a lot of sense. Hey, we'll continue our interview in a moment after a word from our sponsors. Going back to what you said about fundraising, why was it so hard for you in the beginning?

32:20Why didn't LPs see the vision at least a few years in when things seem to be trending well? Man, it takes a long time to actually, when you invest in Inception where people just have an idea and they're just assembling their team, it probably takes 12 to 18 months to even get the product out the door in a way that you're happy about. Then it takes another year to figure out what am I selling and how to get there. So it's like two and a half to three years to get the product market fit. Okay. We have this slide on our AGM, which I'm prepping for right now and says path from zero to 5 million and path from zero to one.

32:49Look at even like Clay in the portfolio, which people talk a lot about clay.com. It took them six years to get to their first million and now it's grown like wildfire right now. And the last announced round was at a 500 million post in like May. But they had built this more programmable Airtable and guess what? Airtable won. So people didn't want a more programmable Airtable. And then people started asking them, hey, can you start hanging all these data feeds in? Can I start using this for sales prospecting. Then they kind of did that and the company just took off from there. So it takes a very long time to kind of see what happens.

33:19Two is we're non-conventional, right? We're out of New York. We're writing first checks in enterprise where everyone thought the Valley was the only place to write checks out starting in 2010. And I didn't come from a pedigreed firm where it was easier for me where LPs knew me. So this was just building it brick by brick with my partners. And now we have Alan out of Boston, a former founder. We have Shomik out in the West Coast, who you may know, and then Elliot. So it's four of us now. We have an operating partner team, which I can tell you more about, but man, it just took a while. Yeah. And that this is the most exciting platform shift of your career.

33:53Do you think the internet bubble is a good analogy in that we will have this sort of either downturn or drop of excitement where people are taking money out or questioning, but then it will lead to sort of the opposite or what happened in the early 2000s? Or do you think we'll just kind of skip that and just go straight to people will see, learn from mistake last time, not pull out and just say, Hey, we're just going to, there will be companies that lose money, but this people, we won't lose faith in the trend and in real businesses being formed. I think it's a little of both. I think the downturn, when I watched the internet bubble, I mean, I had, I was actually on the board of a company that was when the last companies went public called live person.

34:33It was a, at the time it was called ASP. It wasn't called SAS or cloud for a live chat. And it's still public out there right now. And the founder of Al Pacasio is a good buddy. But that downturn lasted a long time because people really did not know. There was no prior analogy other than maybe railroads or telephone, right? That kind of you could point to. And I now think that everything has been compressed because of information and because of the internet, things are compressed. The speed at which decisions are made, the speed at which people make decisions has changed. So yes, there's going to be an absolute mega crash of some of these companies out there.

35:07But at the same time, it's not going to be a downturn that's going to last forever because as some of these companies are crashing, there's going to be tons of pockets of companies in the enterprise that are making shit tons of money. And so people might be scratching their heads like saying, huh. And so yeah, it just means that people are going to be more judicious about what they invest in and how they invest in. And perhaps maybe some of these round sizes may get constrained again, but I don't know because these multi-stage funds keep investing more and it's easy for them to say now, once it used to be a$5 million option check, I can write a$10 million option check right now because now my fund is$40 billion in size and really doesn't matter anyway.

35:42Guess what? That screws up some product models and screws up some of the founders, right? So those are some of the things that we think about. But yeah, there's going to be a lot of companies that actually go, big companies that people know about that have raised a ton that probably will not make it. But at the same time, the speed at which growth will happen on the other companies is going to also show and people are not going to stop investing. They're going to continue to plow money into this because this frankly is, if you believe, even if there's AI scaling law issues that we have, this actually is something that I'm watching the enterprises, the largest enterprises in the world, deploying this, starting to move from pilot to production.

36:15They barely scratched the surface right now. And then on top of that, I think the labor piece, I think it's real, right? I think that if you can get labor productivity, you're talking about trillions of dollars from that piece. So I think the agents are going to take over and I think it's going to be a pretty cool time. We'll get into agents. in a bit. But first, I want to talk about sort of almost the last platform shift. Let's do this thought experiment. Let's pretend we're back in 2009, 2010, and you're doing the fun pitch, and you're talking about sort of shift to cloud and SaaS. And then you go in a coma for like 14 years or 13 years.

36:52And you wake up, and there's a version of you that has been around for the whole time and have seen how the enterprise has evolved. And you're explaining to the version of you that has been a coma for 14 years. You wake up and you're like, what happened? What changed the world that changes the landscape of enterprise investing? How would you summarize to the version of yourself that was in a coma, the big shifts to the big evolution so that that person gets up to speed? That's great. It reminds me of the whole Rip Van Winkle thing, but maybe it was 20 years to 25 years. I got to tell you, Eric, the thing that blows my mind right now when people started talking about the cloud was when the banks were allowed, when FINRA said, hey, banks, you can use the cloud.

37:35I thought it would take forever. I thought we'd live in a hybrid world where it would be on-prem only and then maybe pockets the cloud. The fact that the cloud and every enterprise on the planet said cloud first, that to me was absolutely incredible. And that spurred everything else. Because when you have that replatforming and the architectural shift, You got to rethink how you do monitoring. You have to rethink how you ship code because now you can ship code fast. It just goes right into production super fast, right? All of that changes everything on the infrastructure side, which is why you spawned all these great companies.

38:05Also security. Now the boundaries have changed. I mean, Palo Alto did a great job turning from selling firewall boxes to kind of becoming the cloud provider and becoming an absolute giant, right? So all that changes. And so we think about that same analogy. You're moving that into the Gen AI platform shift. How do you think about it? You're looking at a data-first infrastructure, right? The way you build applications is going to change. How you monitor those applications is going to change. How you secure those applications is going to change. And so that's kind of what we're looking at now. The question, though, is I think you have a lot of companies that are a lot smarter around the cloud.

38:34So a lot of them, particularly in the infraside, it's like, great, Datadog can add AI monitoring. That's not that hard. I don't think there should be a standalone company, right? So you've got to be very careful there. And I think more value is going to accrue on the app layer side as you kind of move up the stack because a lot of these info providers that didn't exist before the cloud started, it's easy for them to add some AI piece to it. So I'm watching that. And the second trend, by the way, which is very interesting, I kind of joke about is that everything kind of goes back in cycles. So now if you're a company selling into the enterprise and you have an AI related product or you're leveraging AI, guess what?

39:08Some of the largest firms now say, hey, can you actually deliver it on-prem or can you deliver it into a VPC for me? I literally just had a conversation with the founder today. He's like, yeah, I think I can drop that into VPC without going on-prem. But so now there's partly this trend back because of security, privacy, data sovereignty, all those things. So this is a really interesting aspect, but I would have never predicted how fast enterprises adopted cloud and went all in. I'm talking about the biggest banks in the world saying, we're cloud. That to me is the most mind-blowing thing. The question is, what's the same analogy, Eric, moving forward right now?

39:38And I can tell you this, we're still early right now and have not seen, and by the way, Finran and all these people, the regulators still have not approved a lot of these AI applications that banks want to deploy. But I can tell you this, I've never seen more interest from the C-level around a new technology or platform shift even back in the cloud than with AI today. And so you got to invest in the trend. Disability devil's advocate on that. Isn't that also an argument for why incumbents will accrue a lot of the value or capture a lot of the value that this platform shift to enable? Some people say AI is more of a sustaining technology than perhaps a disruptive technology because anyone with data, aka the incumbents, gets so much stronger.

40:18Do you agree with that characterization? And if so, where's the opportunity for startups? I would say, once again, I don't think in absolutes. I think that there are going to be pockets where incumbents are way better. It's just all about the people, how they perform. And then I think there are opportunities where startups will be better. So for example, if you talk to lots of enterprises, they'll say, hey, platforms, I love platforms, I love consolidation. But I think as you move to the world of agents, you have to ask yourself, is the great unbundling of SaaS happening? Meaning that why should I buy one great product, 11 mediocre products, when an agent or agentic infrastructure can actually pull data out and actually allow the end user not even look at that shitty software screen and provide me with value?

41:01Basically, that SaaS app becomes just another database. And I'm not saying that Salesforce is going to go away. No, I mean, that's going to be a very powerful company that have lots of data, but the very thing that they're selling this whole agentic platform, and obviously it's built on Salesforce only infrastructure, it's not like it's a horizontal place, so you're locked in. But if you brought up Crew AI, for example, which has grown like crazy, I mean, these guys are just crushing it in terms of how many teams and agents are being run across various applications. I do think there's a world that some of these SaaS companies become just intermediated.

41:31So if you're not actually cutting edge, if you're not the CEO coming back and kind of driving kind of change, then I think you're absolutely up for being kind of reinvented, right? And then secondly, I think you have got to ask yourself, how do you reimagine kind of what's already out there in the world? And since this podcast will come out after Guy's announcement around TESOL, I'll give you a great example. I think AI coding assistants are very cool, right? They provide a ton of value. The thing about it, they just turbocharge existing workflows. They're not doing anything different. It's like Grammarly for IDEs, right?

42:02Your code, tap, tap, complete, tap, tap, complete, boom. Very cool, right? It's very productive. It's not hard to see. What if though, if you re-imagined how software was developed, what if you had a prompt? You said, hey, I type into the prompt what I want to build. It knows the business requirements. It actually knows how to build it technically in the backend. We'll call it spectrum and development. And all of a sudden, it assembles kind of the whole diagram of the technical architecture. It puts the software together, not even the code. It It puts it into components or tiles. And then all of a sudden, it's smart enough to auto-maintain itself, smart enough to secure itself, smart enough to patch vulnerabilities on its own.

42:40Well, that's a completely different way of thinking. Will it work? I have no fucking idea. Will a guy actually push this envelope? Yeah, he's built a lot of stuff right now and it's kind of pushed the envelope. But I think there'll be opportunities like that. How do you reimagine what's already done? Just because it works now, the comments are never going to actually take that leap of faith and go forward. One last example would be like a data dog. Datadog's making a fortune right now, but people are actually like, why am I paying so much for lungs? But what if you actually, I've got a company that's sitting in the middle, grabbing AI, grabbing data using AI and ML and saying, yes, you can still use Datadog, but I'm going to reduce your bill by 96 or 98%.

43:16And we're going to put it all into S3 buckets, and then we'll rehydrate that stuff as it's coming together. So there's also business model issues with some of these incumbents as well. So they're not all going to win, but some of them will. So you just have to really be smart about each category and how you navigate where you're going to invest. Yeah, to that end, Chris Pike from Pace Capital had this blog post, The End of Software, where he was making a comparison of what the internet did to media in terms of content explosion. And what happened there was the bar for being able to charge for content became much higher because there was all this amazing free content out there that the same thing is going to happen in software and in SaaS where there's going to be an explosion.

43:57It's going to be so much easier to create this kind of software that the bar to charge, certainly the amounts that you just alluded to, is going to be way higher because there's going to be all of this free software. And that's going to change how we think about SaaS as well. Yeah, I think there is going to be in Bunt. I mean, look, I think it's a very thought-provoking essay. And the way I think about it is what are second-order effects, right? I'll give you second-order effects if more code is being written by AI? And if it's easier to write code, and if guys, Tesla takes off, will that democratize software development?

44:27For sure. But will there always be areas where the enterprises just have regulators and doing all these things around it that need better security and everything else? That's not going to go away, but there will be around the edges and the fringes for sure will go away, right? It will change things. So if you're just a shitty little app that doesn't do much, yeah, you're not going to make a lot of money. But if it's complicated stuff, I think that's the potential for value there. The second thing I think about too, by the way, on second order effects is imagine a world where thousands of agents are working at thousands of companies and they all have rewrite access.

44:59How do you actually think about, how do you manage them like employees? How do you manage your digital employees? How do you make sure you secure each agent accessing different data? How do you provide them with runtime real access? These are all the things that you have to start thinking about. Just let's assume the future. If you don't believe the future is coming, then maybe you should just go buy all the incumbent in stocks right now. But if you believe the future is coming, you have to start thinking to yourself, especially for guys like us, because if we're first at inception, typically it takes two to three years for some of these things to roll out.

45:30And I do like investing in new categories, but if it's already a sexy category, I'm done. It's not like Guy came out with another coding assistant. I don't want to go against 30 coding assistants. He's coming out with something absolutely bonkers, right? So it's like, yeah, let's go for it. I love Guy. I think he's going to go for it, but you've got to come out with some bonkers stuff. When we invested in Protect AI and AI security in early 2022. That was before Chad Chabit even came out. Part of the memo we had was, Ian, do me a favor. I know that you can probably get it to$10 million because he was, by the way, running GoToMarket3AWS's AI and ML group.

46:01Just make sure it lasts three years because I just don't know when the seminal, neither of us know when the seminal AI moment may come where they care about security. But we did know, based on his experience selling into the largest corporations, that they did have a need for what he had. I just didn't know if it would be bigger than$10 million, right? So fast forward now, we're seeing every large enterprise kind of move around. So we do have to make those bets and think about the future way ahead of everyone else, because if we don't, Sequoia and Andreessen and everyone at Lightspeed, they've already thought of it.

46:27And I'm too late. I get paid, meaning that my companies do well and get funding at up rounds when those guys have something, when they see something that didn't exist before and they never got to. And I want to be the first one there. So I don't call that. Some people say that non-consensual consensus, I just call it fucking early. We want to be there so bleeding fucking early that maybe people are like, you guys are crazy. This is nuts. But maybe within three years, it's not as early. And people are like, oh, now that's consensus. But I don't like using those words. I just want to be so early that I think I can see the future.

46:59And I call it intuitive TAM. If this works intuitively, do you see this being big? Yeah, I think so. Almost like if there's a market map, it's too late or something. It is. That's why I say if you're Gardner, we're market mapped, it's already too late. You just nailed it, man. I don't want to be on someone's market map where I don't want to have a company on a market map right now because it's already done. Yeah. So say more how you've thought about investing opportunities in AI and where you've chosen to play and not play. You alluded to earlier that you're not backing model providers. Some of the people who've done that have done very well.

47:32How have you thought about that? I just think about, man, the end result. How much money will a company need to raise over time? And can I stay and maintain that ownership over time? And the model providers, yeah, there are already fabulous fortunes being made. We're just never going to be that fund size. So therefore, it automatically eliminates our ability to do those things. Not to mention, I do think it's a very hard space to invest in. It's like playing Russian roulette. And it makes sense if you're a very large fund, right? Because if you hit it right and you have four others that don't work, that one right will probably make up more than make up for everything else, but that requires a shit ton of capital.

48:13And so for us, have we done some models? Yeah. My partner, Ellen, led an investment, we'll just call it in a, I can't say who or what, but it'll be released soon, but in a foundational kind of model with some folks that came out of Google Teeth Mind for robotic hand dexterity. And they have a proprietary device to gather proprietary data with these 3D printed hands that have cameras on it that gathers data and they have a way to use synthetic data to develop it. And so basically now a robotic hand can have a task in front of it that requires a lot of dexterity without having seen before and actually open up a pill box and dump out pills or pick, pack and ship some small things.

48:51But that was an incredible team kind of behind it who had actually, some of them were on the Google Palm Me paper. One other guy came out of Boston dynamics. And so that was an incredible team. They've done it before and there's a proprietary data mode and we're able to get good enough ownership to start in which if it does kind of move forward, then it's not going to require the same amount of capital that some of the other providers may need. And so maybe gearing towards closing here, what are the biggest questions you have, whether it relates to AI or enterprise investing more broadly about how certain things are going to play out?

49:24Or like if you could ask your future self, we do the coma self, but your if you just said like three years from now, hey, what played out in XYZ? What would those questions be that you're waiting to uncover? I think the biggest question we have to all ask ourselves that are deploying capital is, look, the AI prices right now or anything related to AI are kind of Zerp-like in certain ways, right? And the question we have to ask ourselves is that if you're paying these prices now and multiples have compressed significantly and they're still down, will multiples expand again? Because if they don't, all of our ability to make money now has gone from, here's where it is now, here's where it used to be.

50:02And the question is, is that going to expand? And how does that expand? Well, I think everyone's going to have to show, particularly public companies first, will have to show that this AI stuff is delivering real value. And it shows that we're delivering real value, we just say the industry. And if we actually are aiding into some of that labor, right? Because remember labor, you're talking about$60 trillion of potential productivity gains kind of from labor alone. And so if we even scratch some of that surface, then the TAM multiple is going to expand over time, okay? And plus some dropping interest rates too.

50:33We always need that, right? Then that's something that we can't control. So I believe that we will have some multiple expansion over time. And if we don't, I think all of us are in a world of freaking hurt. We can tell you that, right? I mean, all the models would be both around it and everything else, this whole industry and this whole complex is going to be questioning themselves about what's the right thing to do. Right. And going all the way up to LPs in terms of like, does this asset class make sense basically? Yeah. I'm a believer that I've seen the worst of it, man. I've literally seen the worst.

51:03I lived through the worst of it. It was the most depressing time in my entire life as a venture capitalist during the dot-com death and also during the 2008 financial crisis. And right now, I'm just saying that even if we hit kind of a skid, we'll rebound very quickly. It's a different world. Totally. Well, that note of realism and optimism, Ed, thank you so much for coming to the podcast and discussing firm building, the asset class at large, and AI and enterprise investing. Thanks a lot. I really appreciate it. And keep cranking that podcast, man. It's awesome. Thank you.

51:45Thank you.