Don't Watch This If You Want to Sleep at Night

Intro

Kevin Hi everyone, welcome at Wavect, today with Joseph Miller, a serial founder and scientist. And as you can see, Ernesto is already super excited for this episode, because we were talking about all things, you know, from robotics to his entrepreneurial journey to data privacy, ownership, to all sorts of crazy stuff. And if you watch this episode, you will see that I just didn't know what to say sometimes. So it's really an exciting episode and super authentic. I love Joe, he's just a great guy. And yeah, enjoy the episode.

Kevin Hi man, thanks for taking the time. It's really, really a pleasure talking with you again.

What would he tell his younger self starting out

Kevin I will just throw the first question at you. As most of the viewers already can see in the description, you have founded many companies in your life, and the first thing that I'm definitely sure most of us would be curious about is: what would you tell your younger self when you were just starting out with your entrepreneurial journey?

Joe My instinct was to answer "don't", but that would be a total lie. I've had a great time, it's been a lot of fun. I mean, entrepreneurship is a really interesting, unique life trajectory, right? It's very unique in a sense, it's very non-linear. So if you're the type of person that really enjoys that sort of non-linearity to your life, like I do, then it's really the only thing you could do, right? Diving into a new domain is basically impossible in a normal career trajectory, whereas in entrepreneurship, that's sort of what you do. You move from one business to another. I've had a board game company and an enterprise software company and a Web3 company, a hedge fund, all these random things that are seemingly random, but the backbone of that is a curiosity, a passion for solving problems, just trying to do something new, right? Trying to do something new and meaningful and exciting. And I like to say: live as many different lives as you can. So I would tell my younger self that you get all of that, that indeed is actually true, you do actually get all of that.

Joe And then I would probably give myself the warning that while all of that is the upside, the downsides are very feast and famine. When things are working, it's sort of a [unclear] type of thing, and that's for the good and the bad. And the crises come from anywhere, so there's really no security in that sort of life trajectory. There's just no security. If you're not building upon a stable sort of thing, you're always trying to do something new, it's just a fragile way of existing, really, which makes it exciting, but also [unclear]. So I'd probably tell myself to have better risk management, better decision making when I was younger. I think I could have capitalized on a lot more of the opportunities that I had. But, you know, live and learn that way.

Kevin Well, sounds to me like there were of course a lot of ups, but also a lot of downs. Was there a moment where you thought, ah, well, I'm not sure if this is going to work out?

Joe I mean, in every single company there are multiple moments like that. And then it's very bipolar, right? One moment you're manic and everything is just exciting and an awful lot of opportunity, and it's all working, it's humming along, and then literally just out of nowhere, something happens in the global economy that has nothing to do with you, and it can change everything. And so, you know, there's a great study, I forget who did it, I think it was an HBS study that I read a long time ago, that basically followed a bunch of entrepreneurs for ten years or whatever, and it basically found that the number one predicting factor of success for these companies were the things that didn't happen to them. So in some sense, it's like being the last soldier on the battlefield, that you just didn't get shot. But that has almost nothing to do with you. So most of the danger is things you don't really actually have a ton of control over.

Joe Now, having said that, if you succeed by sort of surviving in that sense, then the name of the game is to make decisions that allow you to keep playing, right? It's like playing poker or something like that. You just got to keep being in the game, because you're waiting for opportunities, you're looking for a little break in the window to be able to capitalize on. And if you're not in the game, you won't be able to play, right? So that's what I say about this risk management: it's really understanding that that is a big, big part of just trying to build anything, trying to do something difficult. People make a big part of the grit, totally necessary, but the grit won't matter if you can't play, right? So you gotta survive, you gotta look for the opportunities to hit, and hopefully you get lucky and you can avoid some of these things that have nothing to do with you.

Kevin Yeah, I've seen, not sure if it was a study, but I think a couple of entrepreneurs in Silicon Valley or something like that, and if they survived over ten years, they all have in some way been successful. And resilience, and keeping going, and survivorship bias, yeah.

Joe Yeah, of course. I mean, I think there's not a single company that has just worked without those existential moments, right? I mean, Vivun is, in all regards, an unbelievably successful company. We raised our first round in 2019, great time. The next few years was just amazing, but then you have 2022 and this global macro space. It's a SaaS company, very sensitive to the global macro environment, and SaaS as a whole sector was really brutalized, right? So you go from these really, really highs and feeling like, oh, we've really got something, to a hard wake up call. There's really no silver bullets to anything. You've got to block and tackle, you've got to execute well, you got to manage your cash flow well, you got to be prepared for these unforeseen events or changes in trends and in the markets that you're really sensitive to. And when everything is working, it's really hard to know that, right?

Joe The best place that that happens is in the finance space, in the hedge fund, when you're trading and you're just crushing it and it's really working well. You think you have something, right? You think, I am a genius, I have figured out the secrets to the market. And that's the thing about statistics, man: if you play enough hands, you'll have lots of strings where it just looks like you're a god, like you're untouched. Similarly, though, the opposite side of that, where you're like, man, I am having an unbelievable run of bad luck. And you need the balance between being able to say, when it's going great, you're not as good as you think you are, but when it's bad, you're not as bad as you think you are. And luck plays a big part of this, and you've got to be able to reason about it statistically and say, okay, should I keep going, or do I have something systematically breaking here?

Kevin That's a great point. Someone told me, I think a few years ago: the ups are getting higher and the lows are getting lower.

Joe Yeah, some volatility there for you.

One of his most crushing experiences as an entrepreneur

Kevin But looking at your companies, what was one of the most crushing experiences for you so far? If you want to share that, of course.

Joe Oh man, the most crushing experience. There's so many.

Kevin We have time.

Joe I mean, I think it's actually not true, like there's not that many for me. A lot of these companies have not worked or have not succeeded the way I hoped they would, but, I mean, it's cliche, but those failures are like trophies to me, right? When you're doing something truly new, when you're really trying to, even if it's just new to you, right, that doesn't mean new to the whole world, it's just new to you, it's really hard to fail in some sense. You can run out of money, the company has to go bankrupt or whatever, but the experience of the whole process is sort of just like a river, it just kind of flows, right? So there's not really a good or bad type of aspect to the quote-unquote failure or success of things.

Joe I would say the hardest thing, the darkest I have felt about the thing, is in finance. Because Battery is a systematic FX trading fund, and it truly is soup-to-nuts autonomous. That was sort of the thing: I had a bunch of holy grail properties I wanted to build into a hedge fund. One was that it has to have short-term trading, but not high frequency, not intra-second trading. It's not trying to skim things or anything. It has to have a view, but it has to only hold the trade for a day. It can't trade over the weekend, so you can't exploit interest rate changes or whatever. It is truly a noise trading system. It should be, in theory, impossible if the market was efficient, but clearly it's not efficient, otherwise this would be impossible.

Joe But there are many, many times when I thought, you know, I've been building this for four years, and for the first two years especially, there were many, many times where I thought I got it, and then realized that there's foreknowledge here, or a data leak here, or whatever. And the whipsaw effect of that is brutal, especially when you have this haunting feeling that maybe the markets are efficient. And it's hard to hold conviction on this kind of roller coaster swing, and I've been riding that roller coaster for years. And that is really, really painful. That can be very painful at sometimes, because you feel really stupid a lot of times.

Joe And especially, this is another thing I would say to my younger self: when you're launching companies, if you're an entrepreneur, you're sort of arrogantly optimistic. It's almost like a necessary but not sufficient condition to succeed as an entrepreneur, that you kind of gotta think you're really special in the face of extremely daunting probabilities that you're gonna succeed, right? And so for anybody to go do it, you sort of have to think, I'm that good, which is very, very arrogant to say. But also you got to be able to sell, right? And so being enthusiastic about what you're trying to do, and excited about the way you see the world and the way you think the world should change, or whatever it is you're trying to do, is a big part of it too. So every Thanksgiving you go to your in-laws or whatever, and they ask you what you're doing, and you come with the same fervor and excitement about your latest ideas. That's amazing, yeah, that's amazing, I'm so proud of you.

Joe And then it doesn't work, right? And then you're, oh, I got it, and then it doesn't work. And for everybody that's not an entrepreneur, that doesn't viscerally understand why somebody would stick through that, you look super stupid, and also possibly unstable. You kind of look like somebody that is like, oh, do I really want my daughter with this person, or whatever, is this guy gonna make it? You know, fortunately I've had a lot of support from everybody around me, but that is a feeling that is real. You look delusional until something works, right? So it can be very isolating, can be very lonely that way. And then it's just a grind, man. It's tiring to continually fool yourself, right? There's a great quote, I forget who it is, I think it was a physicist, it might have been [unclear] or somebody, but it was like: the first rule is don't fool yourself, and the second rule is that you're the easiest to fool. So that is a very good principle to live by when thinking about that.

Did Joe lose his superiority complex

Kevin That gets me curious. You have to be optimistic, of course. You have to be, as you termed it, flipped it, coined it, arrogant in some way, right? To think that you are superior over all others to make this work. Which, of course, there's this thing, superiority complex, basically, that most entrepreneurs have in common, from what I've seen so far. And I'm just wondering, when you have multiple companies one after another, don't you lose that at some point? Because you get, I wouldn't say more... no, more in terms of: you understand the likelihood of achievements much better, you know? That's just my thought a little bit, that you get more, yeah.

Joe Yeah, I think that, like, anybody that knows me would not ascribe me as a humble person. It's not one of my strengths, right? But there is a fraction of humility that I think has been pounded into me, but mostly just because you fail so much that, if you sampled my startup history, the distribution looks like the normal distribution. You have mostly failures, but you have some successes, and that's what the world looks like, right? So you sort of recognize that when you're trying to do hard things, you're gonna fail. If you had a perfect success record, you're probably not doing hard enough things. If things are just slam dunks all the time, it's not very hard. So I think that putting all of that into perspective just gives you a little bit more of a comfortability with failure, a comfortability with the probabilities of success. And your mind starts to go less from, oh, I'm the thing that makes it special, to more like, I have a process that allows me to circumnavigate myself, right? I can get around my weaknesses, and I know that I have these weaknesses, and these things have led me to increase my probability of failure in the past, so now I can get around them.

Joe Like, I used to have a very strong disposition to want to do everything myself: the design and the engineering and the leadership and everything, right? And it's just a stupid way of behaving. One, [unclear], not the best at all of those things. I'm not the best at really any of those things. And two, you just can't do as much. And so it literally just doesn't make sense to live that way. The only reason you would do that is for an ego, but even that, it doesn't serve your ego, because you can't do as much. So it's a little bit silly. But going back to this thing of saying, the humility comes in the realization that you have to survive, and a lot of it is just not up to you. And so you can be the greatest guy in the world, but if you're on the front lines charging the beaches, stray bullets have nothing to do with your talent or skill or anything, right? So a lot of it is just luck, and the realization that you win by playing a lot of games, a lot of hands, right? So you gotta just keep playing the game. As soon as you stop playing the game, then you've lost. But as long as you're playing, as long as you're at the table, you have a chance.

Kevin That's a great mission. If I could ask you for one single favor, it would be that you hit subscribe on this channel, because it helps our channel more than you can imagine. And, you know, the bigger the channel gets, the bigger the guests get. So thank you very much for watching, and let's continue with the episode. That's a great way to put it, yeah. I think there's also this saying: if you look back one year and you don't think, oh, I'm embarrassed what I said, how I expressed myself, you know, all these kinds of things, or what I knew, or how I viewed the world, then you're not really growing, right?

Key learnings from his last venture (Quivr)

Kevin And just let me go one step towards, when you spoke about your companies: the latest venture of yours was Quivr, basically a social validation protocol, identity, a Web3 project to some extent as well, or huge extent. And I'm just wondering, since it's the last venture that you actively engaged in so far, what were some of your key learnings you've got out from that?

Joe Yeah, that's a good one. So, I love Quivr. There's a lot of things I really love about this company. Specifically, I have to give kudos to the co-founders, Ray and Vinny. We've built a lot of things together, and it's really fun to build things with people that you've done this before with. And the reason is because the processes you guys want to go through are similar, and so you agree upon how you should chart uncharted territory, right? And that's a really valuable aspect with your co-founders: the agreement, not just on, you don't have to agree on where you are trying to get to, but you should agree on how you are going to go about getting places. And that's really, really important. And so Quivr is an interesting company, because we sort of raised the money, the seed round, in the hype of the whole bull market. And we wanted to do something really big and really difficult. And I would say that is the biggest thing I've learned: that is actually, it was very big and very difficult.

Joe Quivr's about trying to create a home for people's real identity, right? So I used to say, when you go to a bar or you go to a book club, or whatever you might go to, and you meet new people, how do you engage with them as humans? You ask them, you know, what's their name and whatever, but then you ask them: what do they do? What motivates them? What hobbies? I want to know what this person's like. And I rarely ask any of the things that are on a typical social media profile, right? What is on your Instagram is almost, I mean, antithetical to what I'm trying to say. That is not at all who people are, and that is probably part of the problem: most social media sites, they're just not authentic in any way. It doesn't represent anybody at all. I can see a whole Instagram and be like, I have no idea what this person's like, right? And it's very different when you're in person. You meet people, and then it's authentic, and you're like, oh, it's real. It's not so showy, it's not doctored up and all of this stuff. And so we're trying to create a place where that identity could exist digitally.

Joe And so what we wanted to do is be able to allow people to get these badges and then authenticate themselves and their passions in just a non-filtery, non "it's on your friend's feed" kind of way. Just sort of like, hey, I am a fan of Peloton, or a fan of equestrian horse riding. And here's videos that are just real-life camera roll videos to support that evidence and authenticate those passions. And it becomes, there's this kind of dream of all the things that you can get out of that, right? It's like, this is what your profile is going to be in the future, right? It's your real identity. You marry that up with your professional resume, and it's like, I have a really good idea of what this person's like. I know what this person's about. And that was the dream. That was sort of what we wanted to do: create a human place in the digital world, right? A place to represent your humanity in a digital world in an authentic, validated way. That is incredibly big, incredibly audacious, and was incredibly hard.

Joe Obviously, it's difficult for lots of reasons. One is that, you know, in the Web3 hype, one of the big narratives was that people want to own their data, they want to own their identity. And that may be true or not, but the adoption rates of, not just at Quivr, right, which is a startup that's difficult for anybody to know about, but at any of these types of things, it's just not that great. So it suggests that people actually don't care that much about their data or their privacy. And indeed, if you looked at behaviors across social media, that might not be that surprising, actually. It's a thing that we say, but then we don't do, right? And that may or may not be the case. That's just my experience of looking at the data and trying to understand the space. It seems like that may be the case. And then secondly, nobody really wants another social media type of site. That makes total sense. So trying to get people to adopt both a Web3 thing and a new identity solution was maybe a bit of a bridge too far. I think that the product required a lot of fundamental change to the way people operate on the internet. And that, I think, going back, thinking about it, I would be like, well, that is going to take a really long time. Even if you get it right, that's just going to take a long time to effect that kind of change societally.

Joe So that's okay. I'll just say this one last thing. The way we went about iterating through the product, though, was glorious. And I love, we started out with lots of different ideas of how to achieve this, and we sort of iterated through it, rolled it out, saw what people thought about it. What did you think about this? We pressure tested it. We fell in love with our ideas, and then spent a week doing something else, and then come back and, you know, you shit all over your own ideas, and you change it. I felt like we really pressure tested the way we built the product, and I really love that process. It takes a long time to build product that way, but every single decision I felt was validated and driven by the data and driven by our iterations through the product. I was really proud of the way that we went about trying to build the product, even if it is going to take a long time to get adopted mass market.

Kevin Well, you definitely can be, right? Anyone that I've shown Quivr to, you know, especially the front end, if you open the application, it's like, wow. It's different, you know? It has that wow effect. And what I really liked is, when you were mentioning before that adoption rate is one of the major challenges, that just reminds me of everything, or basically the perception of risk, let's put it that way.

Kevin When you look at security, or even your health and all these kind of things, you know, or you say you don't want specific things to happen, but you still act against it, you know? Just for the simple reason you don't really see the risk, you know, the thing happening in the future. And it's the same about data ownership, right? You want your privacy, because you know that some companies or people might do something bad with it, or sell it, or spam you all the way through your life and all that kind of stuff, right? And maybe it's really, as you said, a mindset change that needs to happen, or it's just ingrained in our human behavior. I don't know.

About data privacy

Joe Yeah, I mean, you're exactly right. Like, data privacy is a big deal, but it's a little bit like eating your vegetables. You know that you got to do that, or you're going to have a lot of problems later on, but people don't want to do it, right? Because it's not sexy, it doesn't taste good. It's a thing you got to do. And I think that the reason is because people just really haven't felt any pain from the violations of data privacy that they are already undergoing. People's data is constantly shuffling around. I remember, you know, I'm 41, I remember when AOL, the first commercial internets, were coming out, and the first credit card payments were, like, [unclear]. People were losing their minds. Nobody would put their real name on the internet. Nobody was going to put their credit card on the internet. That was crazy. Because people were concerned about their identities being stolen and their credit cards being stolen and everything like that. And now you've sort of worked out so many of those kinks that it's all seamless. It just is so convenient and so seamless and so normalized that people just don't even realize what they're doing. They don't realize what data is moving around, from their location to their cell phone numbers and all the stuff, right? All of these identifiable things.

Joe And also, you know, it's a little bit of my job to build inference type systems. And people don't even realize, just because certain things about you are passed over explicitly in a JSON package, you can infer all of the things. You can link these things together and infer everything about people. Buyer profiles and all of this stuff is very mature, advanced technology. And the issue is that it's been used in such a subtle way that I think people don't realize that their identity is already gone. Their digital identity, everything is already out there. The LLMs are already training and using people's addresses and phone numbers as part of all of this type of stuff. And it just hasn't hit anybody so painfully directly that they have to change their behavior. But I think that if people see direct manipulation against them using their own data, that could spur different behavior. But it has to be almost so brazen that it's hard for me to imagine that actually happening. Like, it's already happening. 23andMe literally sells your DNA. So, I mean, if that's not enough for you to change your behavior, then what is it going to take?

Kevin Well, that's a great question, right? Just think about these phone calls that some people are getting from scammers already, you know, when your daughter or sister or whoever is calling you, screaming for help, send me 5K or whatever, and with their voice. And people know that. I think they possibly then still think, well, the likelihood that this happens to me, it's, again, far away. And that's probably the issue with our human brains, maybe, right? It's the same about everything else. If a war happens on the other end of the world, honestly, you care much less than if it happens in the country next door. Of course. That's just in our nature. And possibly that's, I don't know, just an overall issue we have to tackle.

Joe Yeah, I mean, in economics, we would call this a coordination failure, right? You have basically a group of people that are all failing to protect each other. The group would be better off if we all did a thing, but because individually we don't feel the pain, we don't do the thing, and now the group's worse off, right? So these are the spaces where, like in Europe, you have all these regulations to try to solve that coordination failure problem. And it's a tough balance, because obviously that crushes a lot of innovation there, too. So it's a mixed bag, man. But, yeah, I think that in general, people will not change on this topic until they feel direct pain.

Joe And I don't know. Like, to me, you look at Instagram. Instagram is a great example, because the company is just, I think it's a diabolical thing. And that feed, the order in which the feed comes in, is psychologically damaging. The mental health of children growing up on that is absolutely awful and atrocious. And it's down to a science. And I know the science of it. So it is literally down, that process, the removal of chronology of the feed, was all about this. It was all about being able to push images in a particular order, because you can prime people, and you can do things. You can induce feelings for the ad that is eventually coming, or for the opportunity that's eventually coming.

Joe So it is not just, hey, I want to get an ad in front of a lot of eyeballs that are interested in, you know, going on vacation. It's that I want to make sure this person feels FOMO of all their friends that have been on vacation or other things, and then deliver the ad, right? That is manipulative. And it's dangerous. It's really dangerous. And that is a form of people using your data against you. That is a form of data privacy violations, I think. It is not legally a form of that, but that is a real way in which people are being manipulated, and people's mental health is suffering from it. And still, nobody cares, right? They would rather doom scroll all day long. So I think it's a really dangerous trend that we have to deal with.

Where social media & AI lead and the consequences

Kevin I didn't know that they actually also adapt the post around it. That sounds just crazy to me. But looking at that, where do you think all that might lead us to? Like, what the consequence could be?

Joe Well, I think that you'll probably end up getting some sort of regulation against social media and things like this, because of the mental health impacts that it's having. And these findings are pretty conclusive as they're coming out now. But that will grow to a fever pitch. You're looking at a generation that has grown up awful, right? And high school, I don't think, was easy for anybody. But it is, you know, when you grow up under that sort of scrutiny and that sort of sensitivity and that sort of triggering feelings, right? The internet is, you go on a comment feed, it's awful. People are awful.

Joe And it's crazy, because you go to a bar, or you go to a restaurant, or you go to whatever, you go out to social places, and people are not that way. Or there's a selection bias to it or whatever, but people are mean on the internet because they're anonymous. And in real life, people are more restrained than that, right? People are nicer than this in general. You know, not everybody. But I think you put all of that on teenagers growing up and, you know, trying to sort out who they are. There's already a period of time in every adolescent where we're sensitive to feedback during that time, social feedback. That's part of our evolution. And I think it's a very damaging manipulation of that process. And so I do think that you'll end up seeing an eventual boiling over of that topic.

Joe But I think in general, where AI is going and all of this stuff, I think it's away from this sort of social media feed screen feeding. Anyways, at some point, this is not that exciting, and people do wake up to the idea that this is not what real life is. These people are not that happy. Life is not, I am not uniquely unfortunate or something, right? People will find balance that way. And I think they're not as interesting at that point. And then also, I think that there will be different ways to target and advertise to people. And then you got to think, ultimately, I just gave a lecture about AI and labor up at Yale, actually, in the economics program. And people are debating a lot about how AI and robotics is going to be changing the labor force. But it's certain that it is going to be disrupted. The question is, okay, hey, we've had a lot of technological disruptions over the time. We don't need to be Luddites about this, whatever. But you have to look across and say, well, what is different about this time that looks more like times in the past? People use the example of the spinning jenny, or the textile industry, and then those people go get other jobs. And you're like, yeah, they went and got other jobs. Or the steel manufacturing, or steam engines and all of that, and how that was a big boom for the economy and such, because it created more jobs.

Joe What do you think is going to happen here, when labor is the thing, actually? I think it's going to be a lot of contention. Not a particular job where you can have flow into other places. All jobs are reducing. There is nowhere else to go. Where are you going to retrain them into? The place that you would retrain them is also using robotics, right? So, all at once, this is, I think, going to happen. I'm not going to be a doomsayer about it. I think it's a very positive thing, actually. I think it will move us, advance us into the future. But the population as is, it will be shocking for a general population. I think it will be shocking for a generation or two.

Joe And so then you think, well, what is the world like in a world where labor is manufactured? Labor. Not a product that a robot does. Labor itself is manufactured. People are going to have a lot more time on their hands, and I think that's overall a good thing. You can talk about how that can be sustainable or whatever, but I think that when people aren't having to spend half their life working all the time, or so many hours or whatever, what are they going to do? I think going into digital worlds, games, entertainment, travel, I think these things become more and more important. And then I think also marrying that digital world with the physical world, so that it's not just like, hey, I can't actually go out in the world now. I have time, I have space, I have capital to be able to go out and experience things, and touch things, and feel things, and smell things. I think people will do a lot more of that kind of stuff. And so I think social media and this kind of stuff is way less interesting in that sort of world.

Kevin Oh, Ernesto, what do we have here? As you know, Ernesto is our chief meme officer at our company, Wavect. And, yeah, we just launched a Web3 card game for kids. Two to five players, plus ten years old. And it's all about Web3: the risks, the opportunities, the challenges. And basically, you play this card game and have to collect Ernesto, like, how he's eating french fries, popcorn, sitting in a rocket, these kind of things. So it's really just a super playful way of engaging with the space and actually understanding how it works. So I would really appreciate if you check it out. Of course, we earn some money with that, but every cent, every dollar goes directly towards the free Web3 workshops we do for children. We just had one in Bangkok in November. So I hope you like it, and if you have some feedback, please leave it below. Well, sounds like a great world to live in, if it works out from a political standpoint, right? But, yeah.

Joe Certainly, I think that the world will get better. Like, if it got worse, I mean, there are times in history where life got worse, right? People kind of take it for granted that life always gets better, and you're like, no, it doesn't. There's plenty of times in history where the quality of life got worse for long periods of time. And that may be the case. But then eventually it does get better, because when things get worse, humans don't want things to be worse. They want things to be better. So they do things that make life better overall. And that's why you have progress in the long run. So I think you have to have a longer-term perspective about all of this stuff. In a thousand years, life will be much better. It will probably be completely different, in terms of, you know, AI, robotics and all of this stuff will be integrated into us. We'll probably be another species type of thing, that kind of stuff. But it will be better. Progress will move forward, if we don't nuke each other. Or, you know, Eliezer is not right and the robots don't just kill us all.

Kevin Yeah. There's that too.

What working with Ray Dalio is like

Kevin Well, speaking of data, let's go a little bit in the past. We talked a lot about the future now, you know, I just want to avoid that people get like, oh, oh my gosh. But you worked with Ray Dalio in the past, right? At Bridgewater. And many people know his public persona. And, of course, you can't go into details, totally, totally acknowledge that. But I would love to understand a little bit, like, how was it working there? What were your experience? And, yeah, what did you learn?

Joe Yeah. I mean, so overall, you know, there's a book out now, so a lot of this is public stuff. And I think the book is, like all good pieces of entertainment, embellished a lot, right? But I think that my experience at Bridgewater was great. I learned in some sense how to structure my thinking. And it's kind of a weird thing to say, but, you know, I came to Bridgewater, I got recruited, while I was in my PhD program. And I did a lot of stuff in my PhD program. I had a weird dissertation. It was two parts: one part was on nuclear fusion, another part was on cancer research. I invented all kinds of weird stuff. I had an NSF grant, and so I had a lot of freedom to do what I wanted to do there. I worked in neurosurgery for a while. I did a bunch of random stuff. And I was always able to go into a field and innovate.

Joe But I think, prior to Bridgewater, it was like a cloud. My thinking was more cloud-shaped, if you will. If I tried to draw the connections between things, it was very abstract, like a colorful cloud. And after Bridgewater, it's sort of like I was able to clear fog, and I learned how to break down my thinking and structure it, so that I could synthesize my thoughts better and evaluate risk better, frankly. Because when it's a cloud, it's a lot of vibe, right? It's a lot of, oh, I think this is innovative, right? But then you can't really think about, one, how to execute it. It's kind of a random walk if you don't have a good structure. Two, you can't really evaluate the risk. And the most important thing, which I learned from Ray specifically, is that I can't evaluate why I got it wrong. The most important thing about systematic thinking is that when you build a system that is, I am going to do X, Y, Z, it's like the scientific method. That is why we do it. So that I have a hard structure that has lots of control variables and everything, and I have a hard process in which I'm going through, so that when I see the outcome, I can compare it to my expectation, and the difference becomes quantifiable against the process that I just ran. And if you do not have systematic thinking, what the hell are you doing? You're just doing random things.

Joe And you can't learn as fast, because you won't be able to identify what would you change. Why did you get it wrong? I have no idea, I just did a bunch of stuff. You're like, yeah, it's not good enough. So what I learned from Bridgewater was, one, how to hold a high bar on rigorous thinking, and two, was how to be a rigorous thinker myself, right? And working with Ray was, I mean, a real joy and pleasure. It was like that with everybody there. There's a definite type that comes out of Bridgewater, and it's a type that I appreciate enormously. It is a hard place to work. It is challenging. It's not for everybody, as they say, blah, blah, blah. But for people that want to be the best that they can actually be, it's like an iron sharpens iron thing. You've got to cut out the fluff. It's a hard place. If I do something wrong, you just need somebody to say explicitly: you got this wrong, this is what's wrong. And if you could just get past that emotional reaction that you would have when people tell you that your thinking is bad, you can get on with the work of making your thinking better. And it's just a really rich place to work. I will forever be grateful about learning that skill. It has helped me enormously.

Joe And then Ray specifically, I mean, I did get the pleasure of being able to work directly with Ray. And I love Ray, because Ray is, his public persona is actually his persona. That is who he is. He's not two characters, which I think, for better or worse, is an extremely admirable attribute about somebody. He is incredibly systematic. And so he is actually that way. And it can come across, I think, kind of robotic and a little bit cold. But it's not. It's disciplined, right? It's rigorous. It's transparent. Because that's the other thing, too, about systematic thinking: if you claim that you want to be transparent, you have to be a systematic thinker. Because otherwise, what are you communicating? How do you communicate a nonsystematic idea of what your intentions are, if you can't be clear about what are the variables that you're considering, how are you thinking about them, et cetera, et cetera? I think people think that they are being transparent by telling people their vibes and shit. But that's not transparency. Transparency is about telling me how you came to the decisions that you came to. What are the things you consider? How do you think it's all connected? What are your motivations? All of that stuff has to be part of transparency. And it requires rigorous thinking. So, yeah, I really enjoyed it. I had a great time there.

Joe I also got to work with Dave Ferrucci there, who had a big impact on the way I think about AI and the way I think about knowledge representation. And sort of the old semantic reasoning techniques that I think fell out of vogue have suddenly become so important, as people realize that you can't just throw more data at the problem and get better results. It's actually less data, more semantic reasoning gives you better results, right? And so I'm excited about the period that we're in with AI for exactly that reason. All of those techniques are coming back up, married with the new embedding models and this kind of neural networks and this kind of deep learning stuff. And you're bringing these two together, and you're like, this is actually probably going to blow our minds, and there's going to be a lot of exciting stuff. And I got to get some foundational learning from both: not on AI from Ray, but just thinking and epistemology and that kind of stuff from Ray, and AI from Ferrucci. I was pretty blessed.

General purpose robots are coming soon

Kevin That's awesome. Is there something you are most excited about? Like, something that you would like to see?

Joe I am most excited about AI robotics. I know I just painted a doom story about it. But the thing that people, I think, do not write about or talk about is, if you think about, last year, you know, GPT came out, and it kind of blew away people that were like, oh, crap, if you just stack more transformers, you get this magic to come out. That was pretty surprising to me. I didn't think that that was actually going to happen. And so that was pretty shocking. But when you really tear it back, you're like, what is this thing good for? This thing is really good at translating. So, you know, English to Spanish, pretty good, right? But also English to code. English to mathematics. English to Python, right? And then, I think more importantly, you're getting to a space where you're like, English to computational language, right? This is the Wolfram Alpha type of stuff, even Wolfram type stuff. But then you go from computational language to robotic language. And there's a ton of great work being done at Berkeley and MIT and Carnegie Mellon and other places around how to teach robots.

Joe Actually, Elon's robot is based off these principles as well. But how to teach robots not from codifying knowledge inside of them, or procedure, but from giving them an ability to see or hear other things, and then build the robotics that can do those things. And you have this unbelievable translation engine that can do that. And then if you separately think about, okay, well, what have robotics been up to? Boston Dynamics has basically been doing the same sort of versatility improvements to robots. So they have now degrees of freedom of motion in physical space. You have degrees of freedom of thinking in embedding space, or latent space. You're going to bring these two things together, and I think we are probably five years or so away from general purpose robots that change the economy dramatically. Right now you need a robot to wash the dishes that can only wash dishes. It can't go change your tire. That makes robots very expensive. Just economically, that's not a practical way to build robotics, unless you're a giant manufacturing company. But if you're a general purpose robot, the same robot that washes your dishes can go change your tire, can go mow the lawn, can go fold the laundry. This is wild. That will drive the price of robots, general purpose robots, to the floor.

Joe And it's scary only in the sense that what you're doing there is manufacturing not a product. Currently, everything in the world that is manufactured is a product. You manufacture a solution or a service or whatever. This is the manufacture of labor. So if you think about the Cobb-Douglas, I'm trying to make production, and I have capital and I have labor to substitute. This is, I only need labor now. Or rather, I only need capital, and it removes my labor. I don't need labor, because my capital can do my labor. And so it's a complete hard substitution. And that is a wild, wild world that we're about to go into. And so I think, you know, the writing's on the wall. You look at the labor market, it is unbelievably tight. You can't get fast food workers. You can't get construction workers. You can't get transport drivers. Also, most of those blue collar type of jobs are worked in the United States anyways by the boomer population that's about to retire. Who's going to drive the trucks? Who's going to go do the welding? It's a perfect storm. You have no labor supply, and you have an opportunity to build cheap general purpose robots that won't unionize on you. That's pretty good. You're about to have a convolution over the next five to ten years that is going to fundamentally change labor, and I think permanently.

Kevin Crazy. Because people are saying this about robots for quite some time, right? And it absolutely makes sense, right? Everyone has seen those crazy robots from Boston Dynamics and so on. But five years?

Joe Yeah, five to ten years, I think. Because the main problem is the programming of the robot's procedures. And I'm saying the GPT advancements are allowing you, at the rate of improvements of these types of systems, you are able to translate natural language into computational language that powers the robot. And I think in five to ten years, you will probably have a general purpose robot that you could just speak to and show it how to wash the dishes, and now it doesn't need to be programmed to wash the dishes. You just taught it like you would teach a child, right? Alan Turing, I know we got to go, but Alan Turing had a great quote. He only wrote a few papers, actually. One of the papers, and this was at a time when expert systems in the, you know, 80s were super popular, 70s, stuff like that. But the concept of an expert system goes way back, way back, right? And so Alan Turing had this whole concept, or whole discussion, on expert systems. And he's like, you know, instead of building an AI system, or you could call it that, but instead of building a thinking machine that is trained like an expert, maybe you should build a thinking machine that's trained like a child. And then you raise the child to be an expert. That is more true than any other, like, prophecy in AI right there. That's what's going to happen. You don't train expert systems. You train children. And then you teach them to learn, and then they become experts.

Kevin Well, thinking about this, I can see a whole industry emerging just of secondhand robots that have been taught certain kind of things. And if that happens, that's really wild.

Joe Yeah, there's a lot of stuff. There's a company or a research project happening, I think, also at Berkeley, that is training robots on how to solve real-life problems. Actually, NVIDIA is doing it. It's training robots how to solve real-life problems from simulations. They're not even having to do the problems in real life. They solve the problem in simulation world, and then they have the ability to solve it here. That is full-blown Matrix style, right? So it's very exciting. And I'm saying, this is happening all within the last year. So the thing about, you know, exponential curves is that humans think linearly. So by the time you realize you're on an exponential curve, it's over. So things you think are five years away are probably here in two years. So I'm probably underestimating my time, if anything, saying five to ten years. But I think that, just because manufacturing plants of robots take time to stand up, probably roughly about five years, that's basically what I think is happening. And indeed, if you look at manufacturing trends, China has invested more into general purpose robots over the last couple years, into manufacturing plants. The United States, Japan, and in Europe as well. But all of that is happening. All of a sudden, there's a bunch of plants spinning up to build general purpose robots. And it's probably because they know the software, the tech, is coming by the time these manufacturing plants get stood up. So, pretty good signal, I think.

Kevin Wow. Okay. Yeah, I'm a little bit stunned, so to speak. You know, I always thought robots are far away. Like, I see all these videos, but I just deleted it from my radar a little bit, to be honest, right? Because it was just too obvious for me that we're far away from everything that you see in the movies, an actual general purpose robot.

3 things to remember

Kevin But I have one last question for you, Joe, if that's okay.

Joe Sure.

Kevin We have talked about a lot of super interesting topics today. But are there three things that you would like the audience to remember?

Joe So, I don't know if I can come up with three things. But if you're an entrepreneur, I would say the most important principle I am myself feeling and meditating on is the grit concept. It's just: keep going. There's a great quote, I don't know who it was, another physicist, I think it might have been Einstein, but it was like: it's not that I see further, it's just that I stick with the problems longer, right? And that's a superpower. You don't need to be a genius. You need to stick with the problem. By the time you solve it, right, it takes a long time to solve any sufficiently difficult problem, then you look like a genius. But it is all retroactive, or retrospective, that way. It is grit that makes geniuses, right? So just sticking with problems. Believing in it. Just keeping it going. That is by far and away the most important aspect to me right now.

Joe Two would be, I would say, my Quivr reflections would be: things take a long time. Everybody knows that anyways. Things just tend to take much, much longer than you think that they will anyways. And even, I think, a great idea, like, I think Quivr is a really great idea, the adoption rate is going to be slow, because the change that you're requiring of people is large. And so you got to understand: are you working on a problem like that? If you're working on a problem like that, it's going to take a long time, and you have to consider that as a business standpoint. Do you have the capital for it? What are the milestones you could break down, that you could then fund it that way? Especially going into the next five years, where I think venture capital is going to be writing much smaller checks in a higher interest rate environment. The model just doesn't really work as well for venture. Venture, actually, in general, is not that great as an asset class, from a risk-adjusted perspective. But it's going to be much harder. So the checks are going to be smaller. So how are you going to do projects that take longer? You're going to need to get to profitability faster. You're going to need to show progress faster. That means smaller milestones, more deterministic development paths, et cetera. So I think that that's a really important thing over the next five years, too.

Joe And then third, I'd just say, you know, I think building companies is extremely difficult. Everybody knows that. It is also a ginormous privilege. It is an incredible privilege to be able to work in tech and entrepreneurship and be able to chase kind of crazy ideas. And so I think just remaining grateful, even though we went through, you know, the last two years have been pretty tough, pretty brutal for a lot of entrepreneurs. But it is really, I think, just reminding ourselves to be grateful that this is really cool work. We get to work on really neat things. We get to effect change in a really meaningful way. And, yeah, that will keep you motivated.

Kevin Cool stuff. Hey, thanks, mate, really, for your time. It means a lot.

Joe Of course. Good talking.

Kevin Likewise, likewise.

Joe All right, buddy. I'll see you soon.

Kevin I hope you liked the episode. I did, for sure. I think my shocked face sometimes was really difficult to not see. And, yeah, please leave your feedback in the video below, basically in the comment section. Hit subscribe. Engage with the video. Anything helps. Thank you very much. And, yeah, I'm still baffled, if that's the right word. But everybody left a lot of comments, so I'm very, very [unclear] here. I'll see you soon.