Unpolished Signals Matter More Now | with Yohei Nakajima, Untapped Capital

0:00It's that ratatouille, anybody can cook kind of mentality that like true innovation can really come from anywhere. And that philosophy really resonated with me. It's something I still hold on to. I will always be at the forefront of innovating within venture and openly sharing that as a way to increase learning. Human plus AI is going to outperform pure AI. So then the question is, what part of everything do you want to keep human? what you decided not to automate sends a strong signal about what you believe your strengths are and what you believe your job is.

0:33Yohei Nakajima: Yo, hey, thank you very much for joining me. Thanks for having me, Dan. You are the GP of Untappd. It's a small firm, which I would say is kind of on the exciting frontier of like applying AI to venture capital. Where do you feel like your firm fits into the ecosystem? What role do you play? And a bit about like how you got here as well would be interesting. Yeah, I mean, I think how I got here is very relevant to where I see myself sitting in the market. My first role in investing was 10 years ago when I joined Techstars to help spin up the Disney Accelerator. So it was simultaneously working with the earliest stage startups and with the largest companies in the world at the same time, which was just a fascinating experience.

1:16Yohei Nakajima: But Techstars was the first place I worked at. Startup Weekend was the first event I attended. And actually, even before that, I wrote my college thesis on angel investment organizations. So one of my things that I've been fascinated by is innovation and venture capital as a theme. And I think everything from those angel investment organizations to micro funds to accelerators to studios and incubators, tech tooling and now AI. So it's a theme that I've followed for a very long time. With Techstars, you know, one of the things I really loved about it was really their global focus on innovation being able to come from anywhere.

2:07It's that Ratatouille, anybody can cook kind of mentality that like true innovation can really come from anywhere. And that philosophy really resonated. I mean, it's something I still hold on to. So for that reason, I like the earliest stage investment as well. Just inception stage investing, the phase where someone's going through a major life change and sacrificing things to dedicate themselves to a problem or a customer segment. That transition is just so addicting to be around. So I'm going to be an early stage inception stage manager. And that's what I'll always be. and I will always be at the forefront of innovating within venture and openly sharing that as a way to increase learning across venture capital as an industry.

3:01Yohei Nakajima: I think that's a great position to be in. One of the funny things that I've spoken to GPs about a few times in recent episodes is that venture capital is an industry built on innovation obviously but itself has not innovated a lot uh there's a a good line from from one of the partners at sequoia a few years ago who said that venture capital still operates on the equivalent of a floppy disk like not much has changed do you think that's true do you think that's a fair criticism of the industry yes but i think it's largely structural right it's 10-year fund life cycles. Typically, a GP will run multiple fund cycles, which means turnover at the top is every couple of decades.

3:49And then on top of that, you have institutional LPs who think in decades and invest in four to five funds in a row at a time, which means the turnover in those pockets are every, yeah, again, every couple of decades. and I think that there's not much incentive to change because once something starts working, that's what people want to invest in.

4:10Yohei Nakajima: I guess the implicit question there is, is it still working or will it continue to work? And let's just say, for the sake of argument, this is a bit of a tangent, but I'm curious, do you think venture capital would benefit from shorter vehicles, longer vehicles? If there was one structural tweak you could make to maybe correct some of the incentives? Is there something that comes to mind? I think my gut says the long fund cycles are great. I think the idea that founders get capital from investors who are in it to support them for the long run, I think is very healthy. They're going to almost fail many times.

4:53And I think structurally not being able to be an exit at that time is actually a really good thing for the market. That being said, I do feel like the opportunity to invest in startups and venture funds, the laws around that do limit opportunities for people. I think that could benefit from having the opportunity to do so. And the lack of liquidity is a problem, right? I do like the fact that VCs are by default holding on in the long run. But again, there's some people who want to buy assets that can't. and the lack of liquidity where there's demand, I think is something that would solve some of the issues.

5:30Yohei Nakajima: Are you generally supportive of early-stage VCs selling positions at the Series B or somewhere around there? Or do you think that introduces some short-term thinking of its own? Or does that allow companies to stay private longer? What's the right approach, do you think? My personal opinion is that... Well, it's already happening to some extent. My gut opinion is that any change there should not happen too fast because I feel like it is working well as it is. That being said, there is knowledge around, you know, selling a portion of an asset that's growing heavily to start creating DPI. Strip sales are things that people are starting to do and talk about.

6:15I do think there's, you know, even LP secondaries are starting to pick up. I do think those should exist and be available more readily to everybody than I think they are today, both as knowledge and tools. But again, I say that while hoping that it's not something that everybody will start using immediately and aggressively.

6:38Yohei Nakajima: Yeah, I think that's a very fair argument. Switching gears a little bit like now, back off this little divergence. you're also interesting as a VC I think because I would say this is just an observation may or may not be be accurate but you're probably known at least as well if not more well for being kind of like a a hacker and a builder rather than than just a VC you know you did a very interesting TED Talk, you know, you built stuff like Baby AGI, which led to a lot of coverage, a lot of attention. Could you talk a little bit about that process and also how that fits with you as a VC as well?

7:20Yeah, I'll admit that I am probably better known as a builder slash hacker today than I am as a VC. But that's just a matter of time until my returns hit.

7:33It's interesting. Building is something I've always done personally as a hobby since high school. I've never been a builder. I've never held an engineering job, but I've always done it as a hobby. Usually building personal tools, that's helpful to myself, and I've always liked sharing that. It was only after I started running my fund that I realized how strategic it could be for the role I was doing. So initially, sharing what I was building, there was no goal behind it. It was during COVID. I was sitting at home working alone and probably lonely. And I was like, I'm working.

8:05Yohei Nakajima: I want to tell somebody I'm working. But the result of me sharing my no-code tools, what I got a lot of no-code deal flow. And I did something similar in Web3 where I saw NFTs coming up. I did some of my own art and I sold them and I got a lot of Web3 deal flow. And then by the time I started doing it during AI, it became intentional. And I think what's great about this, you know i'm kind of reflecting back on on seeing having having had it happen and the benefit is that it's one that i can continue doing and repeating um as long as i stay on top of of tech trends and which is great because for i think a lot of vcs uh find it difficult to find new ways to market themselves and they end up often marketing uh a history of what they've done But there's only so much time that you can market your history if you're going to be a VC for decades.

8:58And so I think this idea of figuring out a way to continue doing interesting things that resonates with the audience you want to reach as a VC is great. And I'm not the first to do it, right? Doing podcasts and those are known ways to do it. But I think building was an untapped opportunity that I fell into. And I think I've successfully owned for the time being.

9:21Yohei Nakajima: I'd say so, yeah. And it seems to me like a very useful way, as you say, to kind of stay on the frontier, stay on the edge of what's happening, but also to understand the limitations. So I'm going to throw my kind of framework for thinking about AI and venture capital at you. And I'd love to know what you think. And obviously, feel free to disagree with me. My kind of read on today is a lot of venture capitalists are looking at AI and are trying to understand how it fits into their job. And the easy conclusion to reach a lot of the time, or the problem that they kind of are most quick to reach to solve, is the problem of like needing confidence about the investments they pick, you know, how they select how they screen these kind of things.

10:10And my feeling is that's probably one of the worst places you can apply AI in venture capital.

10:18Yohei Nakajima: You should think more about how it can help drive operational efficiencies, everything around the picking. The picking probably should be the human part until we have real AGI. Yeah. I 100 % agree with you. I start from, let's assume that this is going to be entirely automated at some point. Let's start there as a hypothesis. Because I do think we can automate any aspect of venture capital if you wanted to. Compared to pure AI, at least for a while in venture, AI, like human plus AI is going to outperform pure AI. So then the question is, what part of everything do you want to keep human? And I think as a venture firm, what you decided not to automate sends a strong signal about what you believe your strengths are and what you believe your job is.

11:15And for that reason, for us, decision making, scoring is the last thing we're going to automate. And as you said, we started with data entry, right? Operations, note taking, now everyone's doing that. And we're slowly moving into tracking relationships and intro stats. And we're doing all the things we wish we could do if we had unlimited number of very junior people and then slowly moving up the stack as we get comfortable with those aspects.

11:44Yohei Nakajima: Yeah, I think that makes total sense. One of the manifestations of this that troubles me the most is the firms experimenting with applying AI to screening cold inbound. And that's partly because I'm hugely biased about the topic, because I think cold inbound generally is overlooked and undervalued by VCs. But also, it's another one of those areas where if you don't understand the limitations of the models, you very quickly fall into traps you can't see. Like, for example, it's been shown in using AI to screen job applications that a model will favor AI-generated applications, and it will heavily favor applications that were written with the same model.

12:28Yohei Nakajima: So you end up with a problem where you start using AI to screen and then you suddenly start seeing loads more slop pitches, basically. I think it's important, no matter how you use AI, that you own the... You break down the process and that you own how that's broken down and that you apply AI in a very specific way to each of those processes. I mean, in terms of screening, I think there are good and bad ways to do it. automatically identifying things that are not a fit so you don't spend time looking at it i think is a great use of ai um logic gates kind of fuzzily ranking things that all could be a fit for you is where you fall in the trap of everybody looking at the same things i think i think like that's two different things and even when you're asking like is this a good investment like that's that's a horrible question to ask an ai right you should be asking like what are the underlying hypotheses that would make this investment more successful or less successful and give me the existing proof points that help me understand whether whether or not i might decide those hypotheses are the truth or not and it really breaks it's the same question rewritten but it really breaks things down in a way that you as a human can still make the decision yourself versus trusting a black box yeah definitely definitely there's an interesting example of like a say like an older kind of systems thinking approach to originating deals that where ai could be quite useful and it's uh from michael dempsey at compound he's done a few podcasts where he talks about the process of you know i'm sure it's not the only thing they do but it's one of the things they do is they look at research papers in certain fields and they can track something like for example the number of citations a paper has and then if suddenly that number of citations starts to rise you know something interesting is happening in that field and maybe that's a place to look i think that's a yes interesting way you could start to apply agents to look for these things automatically i'm sure he's thinking about this as well but is there any anything in those kind of lines you think about for for origination particularly i often do think about you know i I often dream about really mapping all the market data, right?

14:49Like all the funding data, all the social signals, all the news signals, like the most private and public funding data to see if you can come up with a way, like which of those signals are most correlated with potentially good early stage investment stages. I'm not there yet, but I do often dream about that big data experiment.

15:11Yohei Nakajima: I've done small experiments. One I did that was interesting was I had AI build a big list of all the largest industries. And then within each of those industries, list out all the major jobs, right? If you were to go search for jobs in that industry, what are the biggest jobs? What are the most common jobs that exist in all of those industries? Now, for each of those jobs, list out the major day-to-day things you have to work on. and then uh embedded all of them bucketed them and then said if this was a startup what would it look like and then did the vector search against funding of those of companies that were similar to that idea to see which like common job roles had had few funding so that that on a massive scale with more data would be really fun to do well if if i know anything about you i expect that will be thoroughly researched and eventually published as a paper which i will read the day it comes out.

16:10Yohei Nakajima: Speaking of papers, another one you've written, which I really enjoyed is about the shared discovery paradox. And I'm going to briefly try and break it down. Correct me where I'm wrong, because I'm sure you have a better understanding of this. But essentially, it's the premise is a game. There are 16 boxes. There are eight players. Each of the players has a clue, but it's only right 20 % of the time. And there's one prize in those boxes that they have to find. And I think to like cut to the conclusion a little bit, the like game theory optimal strategy is that they share the clues with each other.

16:47Yohei Nakajima: They pick the eight most likely boxes and each goes to one of those. Yes, that's, that's great. That's correct. Okay, perfect. When you were coming up with that, or when you were writing this, Were you thinking of the parallels like seed investing or was it just a thought experiment? No, it was. It almost seems like a joke, but it genuinely started with me thinking about I should try to figure out a way to justify why I don't listen to VC podcasts. Like I don't I don't like VC content. I just feel like I feel I intuitively felt like there's a there's a huge benefit to not engaging with VCs as often as like with other people, for example, right?

17:29Like limiting VC content. And that led me to like, is there like a mathematical way to show this? And actually, it started with thinking like, but, you know, wisdom of the crowd suggests that, right? Like, the more people's ideas you have, the better estimate you can have and thinking through that, but like that, that's doesn't intuitively seem true for venture. What's different here? And the aha moment was realizing that estimation and discovery are two different types of games. That's the biggest aha moment for me. The wisdom of the crowd is all about estimation. If you're estimating the weight of a cow or how many gumballs are in a gumball machine, the more people's opinions you have, the closer to the average you're going to get because you're collectively trying to pick one answer that's either correct or not correct.

18:17In discovery, like in this case, there's 16 boxes. It's not just about being right, but it's not just about guessing what's likely to be right, but it's also about allocating your guesses, right? If there's eight players, you have eight guesses. And if you don't think of allocation as a key part of the game and you focus on just getting it right, then all eight players would choose the same box. I think the counterintuitive aspect of it is that if nobody shared clues, there's an 83 % chance that one of the people will pick a box and each person has a 20 % chance of getting it right. If everybody shares clues and looks at the clues, there's like a 34 % chance that it's a single box.

18:57And if there's no coordination, everybody picks the same box. So even though everybody's in each individual's likelihood almost doubles, the group's likelihood cuts in half. And that's just like, again, everyone chose the same thing, obviously. And then there's more math around even if there was incentive to choose other ones, information sharing is not bad.

19:23And so, yeah, I think as we think about venture, this is going to completely oversimplify, but I think alpha comes from discovery and beta comes from estimation.

19:35Yohei Nakajima: I think that is absolutely right. And slightly frustratingly, you sent that link to me the same evening. I finished writing an article about why having many small firms in the market is so important, why it's a necessary balance to have with the fewer large firms. There has to be some kind of equilibrium. And I think it's basically an expression of this idea where you need the many small firms for the discovery and to some extent as the companies mature it becomes i think more of a game of estimation for the larger firms perhaps you think that's that's about right yeah i think so right again and also as you get to later stage companies as well you're not really you're almost guessing which one of these three companies are going to grow faster right yeah it's not like will they grow like you're all kind of agreeing like everyone kind of agrees the top three companies are going to keep growing it's just which one's going to grow faster is becomes more of the question which to me is more of an estimation like you're doing math and estimating versus at the earliest stages it's like what have people never tried before that might work is really the question you should be asking yourself which i think is completely different from estimating because there's nothing to even estimate you're starting from a zero and seeing what zero exists that might become more than zero.

20:56Yeah.

20:57Yohei Nakajima: I think the discovery process also kind of relies on the participants to play in the knowledge that they are probably wrong or at least might be wrong, which itself might be why this logic isn't as prevalent in venture because nobody, it seems, likes to admit to the reality of that, certainly not to their LPs, even if it is the truth. Yes. And I think the challenge with venture is the fuzzy line between late stage and early stage. Like if late stage investors are all late stage firms and early stage investors are all early stage firms, the narrative around venture would just be much cleaner. But then we have multi-stage firms who do both.

21:44And so we compare everything together and it's just really hard to talk about.

21:49Yohei Nakajima: yeah yeah that that's that's something i've complained about many times i i i wish industry reports you know the pitch book reports crunch base whoever it may be i really think it's time they start dividing the market you know they they have to express this bifurcation a little bit more clearly because yeah the comparisons don't make any sense it's like seeing like 93 percent of returns come from six percent adventure whatever those crazy numbers bizarre is really frustrating right because of course like that's that's where the money is that's being put in in the first place but again it's also something again like a small divergence bear with me but like the statistical literacy in venture capital is annoying because the the statistic there as published is is exactly what you said it was about the the volume of returns that are attributed to this small percentage of investors, but it has been shared elsewhere as the volume of profits, as if they are also the most profitable firm, which is clearly not the case.

22:52Yohei Nakajima: There we go. Small rant. But yeah, so yes, early stage is discovery. And one of the interesting things about this game is how simple it is and how kind of deterministically you can do math around it. And so I have a whole autonomous research lab that's called Distributed Discovery that's proposing and executing additional research on top. So in this specific first game, the clue has a 20 % chance of pointing at the right box and 80 % chance of pointing at a random box. That's a very specific shape of clue. Now, one of the experiments was, what if the clue was, it's definitely not this box? if that's what the box said.

23:37And it always picked one box that it definitely wasn't. In that case, that type of clue sharing actually does increase group discovery, always when it's shared. And so there's shapes of clues that are more or less beneficial to share with the group. So that's an interesting research, Eric. Separately, you can also calculate, is there an incentive to ignore clues? Let's say the case where everybody shares clues. If you as one person, you also share your clue, but you decide not to look at it if you're the only person it doesn't does it increase or decrease your roi and so these are fascinating questions that you can start asking um again and these are these are bound so it doesn't it's never going to be one-to-one against venture but it creates a lot of interesting like ideas around uh information not being a clear share is better it's fascinating genuinely and the the first example you know i'm i i don't want to like jump too deep into this

24:38Yohei Nakajima: because like you say it's bound it's not a perfect reflection but something i believe is true about venture is the winners are outliers therefore inherently there are not patterns for success but to some extent there are patterns for failure so that kind of clue of like this is something that probably isn't the right one or isn't the right one. It is maybe a better parallel for venture in that case. And it has similar lessons about, you know, you learn what to avoid, but you don't focus on what to pick. So then maybe you're a bit more independent in your thinking, you avoid some of the traps. Something else I wanted to ask you about was, there's a great paper from, I think it's from Harvard Business School that I've referenced a few times, talks about AI as the cybernetic teammate.

25:26Yohei Nakajima: And particularly in the context of an individual knowledge worker, and how it helps them, kind of empowers the individual. But it conflicts a little bit with other research, or well, not necessarily conflicts, but overlaps, that shows AI is potentially more beneficial to the individual. And actually, those benefits fall away when it's a team, particularly with creative work so i wonder do you have any thoughts on the use of ai in like for a solo gp versus a partnership and like different ways to to think about it i mean i think there's a lot of use cases of course it goes back to the what do you use it for right if you're using it for administrative operational work um you don't fall into the the creativity goes away problem.

26:19So I think there's, you know, using it for execution versus using it for kind of ideation are two very separate things. So I think we can kind of talk about those separately. On the side of execution, I mean, it's incredible, right? There's so much back office stuff to do in venture. We don't have the ability to standardize the way LPs engage with us. We don't, unfortunately, have the opportunity to standardize the way founders engage with us. We have so much unstructured data coming in from so many communication channels. And the result is that most firms have access to a ton of data that if you were to analyze could provide a lot of insights, but they never get the chance to.

27:09And I think that's one of the biggest opportunities, right? beyond just making sure you don't drop the ball but if you if you do data entry if you can do enrichment you can get that right you have so much insight that's unique to you through the private conversations you have with companies that no one even know exists i think that's a huge huge

27:26Yohei Nakajima: opportunity yeah for sure that's something i was talking to to jenny from leona's capital at in the previous episode to this one was the the kind of legacy baggage of like venture capital and how it's operated as an industry, which means people haven't thought about questions like, what is the data we're collecting? What's the best way to store it? How might it be used? Now suddenly they have all these tools where it could be hugely valuable, or it is now hugely valuable. But if they haven't been collecting it and maintaining it and structuring it nicely, and even enriching it, perhaps, then it's hard to get to that value.

28:04Yohei Nakajima: Is there a like a kind of first principles way you think a new manager should think about like building for this kind of reality to oversimplify it i think there's two major ways a company can get good at ai one is have an ai czar to lead and get everybody on ai and that's i think the most common one and we see it not just in venture firms but outside of you have somebody taking the lead exploring what's possible, proposing something, maybe someone above them like checks it off, but they're in charge of pushing it through. And maybe if it's a big enough company that becomes a team, but it's a group of people.

28:42The other opportunity, which I think is more powerful, which we're aiming for is like build a company where every single person is an AI superstar. And like, we love investing in startups that feel like that. Where like every single person is just like a 10X, whatever they are. And I think that's the bigger opportunity if you're a brand new manager, because it's the harder one to do, but it's easier to do from ground up. Where every single person you bring in is AI native, and in that case, you're thinking less of AI as, at least in our case, we did have a couple AI tools that felt like teammates, but today, I feel more like me and my ops person are both leveling up in what we are capable of because we have these amazing AI systems around us, and we're constantly collaborating on how to make our AI systems improve.

29:29At this point, that they share repos and communicate with each other as well um so those are two kind of different first principle approaches that i think you can take and then they're both good and the first one

29:39Yohei Nakajima: is i think easier to some extent yeah yeah that makes sense you've kind of joked in the past about you know being replaced by ai and we've talked about it a little bit already i think the kind of inferred conclusion is for early stage investors they're probably safe like at least for this generation of AI. Is there a cutoff point, do you think? Is there, you know, beyond series de-investing, it's now seriously threatened by AI or pre-IPO? Or is it like strictly a public markets thing at the moment? Is there a point where that balance shifts? So I do think AI investing will happen. It will grow.

30:17I don't think it threatens anything, right? I think if you look at the public market, I think a third are machine transactions already. But it doesn't threaten humans' ability to transact in the market either. right the market does balance itself out somehow i think i mean i think something similar a little bit maybe yeah something i think something similar happened to private as well i mean i think in the secondary market um you know there's still rofer and stuff but it's becoming more liquid but if equities end and those guys assuming they have apis even if they don't with browser use i think somebody today could build a successful fully automated vc that focuses on trading secondaries because it feels liquid enough.

30:59I'm going to go ahead and just say that I think it's possible.

31:01Yohei Nakajima: Interesting, yeah. Yeah, I think that's a good market for it. And if that's possible, then it's really just a matter of like how liquid it is, how easy is it to access. So again, if you look at SPVs, for example, if you can automate your engagement on AngelList, then yes, you can build a venture fund that invests. So then the question just becomes, what are the transaction types that AI can't do? And today the more challenging ones are signing a safe note because you have to meet the founder and do emails and whatnot. But I know some angel investors who are very busy with their day job that they say, I never jump on calls with founders.

31:36Now, if those angel investors wanted to fully automate that email address, they could do that, which means they could also angel investors in the earliest stages and you can automate it. The perfect person to do that is like a successful CEO where everyone's like, okay, they're too busy. I understand they don't want to jump on a call. And they're like, here's the email address I use for my angel investments and they just automate that thing. But I think the pieces are there, gets harder to do the price rounds. If the lead investor is like, why should we let you in? Like you have to justify yourself.

32:03But as we get more and more digital, more and more people work async and the better AI gets, the more access it's going to have.

32:14Yohei Nakajima: I do love the idea of something like the US VC vehicle, particularly as they kind of grow that more into backing emerging managers and backing earlier stage companies, like testing that with like an algorithmic approach is a fascinating topic to me. I think that's like a super powerful potentially. But the question of relationships, this is also something you've talked about in the past. And, you know, obviously, relationships are fundamental to venture capital, meeting founders, getting to know them, getting a real feel for them in the flesh, potentially a lot of the time. Have you changed how you think about that?

32:54Yohei Nakajima: Does that become more important in a world with AI or less? I think both. On one hand, everybody's pitch is better. It's way more polished. It seems more thoughtful. But again, a lot of it's AI. So we have to look more at the founders and the way they communicate. that's not polished, right? Like it's more important to see things that are not prepared for us. At the same time, there's increasing amount of information about founders. I can find out more about a founder today than I could 10 years ago because everybody on average is more active on social, more public and more information is just available about what they did.

33:43And people are sharing information about their kids, right? So like 20 years from now, like I'm gonna be able to find their baby pictures online, right? And also AI can do the search for me. So not only is there more information, I don't even have to be the one to search it. I can just ask my AI to do it. So there's actually more unprepared signal also available, which is what I meant by like, it's just as much, it's more important to look for like signal that's unpolished, but at the same time, there is more of that that's not direct one-to-one. So I do, I read their Twitter, founders Twitter.

34:18i read i go to the reply tab see how they reply to people i go to their twitter posts and see the ones where they've had engagement and see how people reply to how people reply to their posts and how they reply to them because those are kind of the raw interactions where you can start getting

34:34Yohei Nakajima: some like personality signals out of it interesting no i saw a post today about i forget who it was um i think a gp we probably both follow or both know but they were talking about it was the first time that they got a pitch that was a an ai generated video like a very very polished uh like animation you know infographic visual animated experience and they said they loved it it was like a great way to like to to be shown all the information to understand what the company he does and so on i have like a instinctual doubt about this like my feeling which maybe is old fashioned i would rather get like an essay written by the founder about why he's doing it and why it's important have it straight from from their hands which whether where on the two of those do So I haven't seen this one or know what it is, but my guess is that that person wanted to showcase how AI native they were.

35:45It's a unique idea. They were probably very thoughtful about it. I would guess that they were, I would guess, because I do weird things like that, that they were worried that it would come off wrong. so they would be they were double thoughtful about it and as a result i would hope that it turned into actually a better visual showcase a better intuitive way to explain what they do and if that's the case that video might have been really good the problem is when that works and they talk about it online and everybody copies them now those i don't want to look at yeah there's

36:12Yohei Nakajima: a big difference between those two yeah yeah that's very true that's very true i had a conversation with another GP earlier today or yesterday about this idea that inevitability is a great green flag in a founder. You want them to project the idea that they know their company is going to succeed, that it's going to be a part of the world in 10 years. They have zero doubt about it. The problem with that is as soon as you broadcast that, it's quite an easy thing to fake. And it's not easy to test if you're not a domain expert so yeah there's there's a bit of a risk of putting too much emphasis on these things yeah yeah i do like i do like people who are creative and sometimes some building creating a video like that can signal creativity again well executed thoughtful but but yes um the problem is yeah once that when something like that works you always end up getting copycats and the quality just starts going down have you seen any trends in in the pitches that you've been receiving any themes about the way founders are pitching or what they're pitching well there's always trends in what people are pitching i mean i think the the complaint now is everyone seems to be pitching something similar and i think that that the trend does feel i think um true um maybe it's the same like if a whole bunch of vcs are asking ai what they should invest in and maybe a whole bunch of founders are also asking ai what they should build and that would also be a problem as well on the way people pitch decks are much more polished in general which is good but also you can see there's a difference you can easily tell a deck where thought has been put into or at least i think there's a difference between decks where thought has been put into versus one that any i put together but maybe someone has really good ai and i'm actually being fooled so technically i don't know but i think there's a difference uh even that's a signal though i suppose yeah yeah yeah that is right um one thing i noticed is i and i don't know if it's brand new but i do get like memos i like like the two to three page legible memo again and maybe it's because it's like so easy to turn it into a 10 page deck that i would rather get a concise one to two page memo uh as like an initial pitch than like flipping through a highly designed deck all right we have three three questions we round up with it's not not quite a quick fire but quick ish fire i suppose okay the first one is very standard but often quite useful what is the best piece of advice you've been given as a as a manager as a vc that other emerging managers might benefit from hearing well that's a great question i think this one i'm gonna say because i think it's just so obvious but no one keep everyone can hear it often enough but just the importance of follow-through uh like everybody's so busy you know whether it's founders but especially LPs and fundraise and even other fellow VCs that just being thoughtful about following up with people being responsive that really goes a long way and like I'm still constantly reminding myself of that and I feel like I hear it often but I don't think anybody can hear it enough yeah I think that's great advice definitely and very founder friendly which we approve of of course um the second question is it's you can you can either be specific to venture capital with how you answer this or you can go broader if you want i don't mind either but it's essentially the the idea that you as a vc have some vision of what the world will look like in 10 years which which kind of guides your investing a little bit where we're heading what the opportunities are what the what the risks are certainly as well so assume you're right assume your your your assumptions are correct what does the world look like in 10 years and that can be as i said that can be venture capital specific or general obviously things will have have shifted a lot moved a lot and we're seeing this right now the world will be better connected technology will be crazy there'll be tech shifts that that we can't even imagine in the course of 10 year robots will be around and all that but but i but with that question what pops to mind is like what won't change right is we're still going to want human to human connections family is going to be important education is going to be important we're going to need we want to take care of that you know the elderly and make sure that they're not they don't get isolated find if you know we want to make sure that that there's you know that people are healthy the economy is healthy plants animals are healthy um those things aren't going to change so a lot's going to change but but there's going to be a lot of things that aren't going to change would be would be my answer final question easy one if someone wants to connect with you or reach out to you what is the best place for them to find you the easiest place to find me is on x so my dms are open i don't i can't respond to everything but i try to be pretty fairly responsive on uh especially on two replies but also to dms perfect all right well yo hey it's been a pleasure uh i'm a huge fan of what you do i will be following eagerly the you know the research that you'll be coming out with and you know maybe we'll do this again at some point as well thank you i'm a big fan of yours as well thanks for having me and happy to anytime