Transcribe your podcast

This episode is brought to you by Cofan, one of the fastest growing fintech startups, I discovered Koyczan earlier this year when I asked Twitter for the best Bloomberg alternative. And the overwhelming winner was an intriguing new product called Coifed. Coifed is a Web based platform that lets you analyze stocks, ETFs, mutual funds and other assets all in one place. I now use it daily to track what's going on in the market and I think if you try it, you will.


To Coifed has tons of high quality data, powerful functionality and a nice clean interface. If you're an individual investor research analyst, portfolio manager or financial advisor, you should definitely check them out. Sign up for free at coifed dotcom. That's Korowai Fien Dotcom.


This week's episode is brought to you by the MIT Investment Management Company, also known as Temko, the endowment office of MIT. New and Small Investment Funds. Listen up. But TIMCO is looking to find investors starting funds today. But Temko is partnership driven, long term focused and has an extensive history of backing investors early in their careers. These partners are key to delivering the outstanding investment returns required to support MIT's pursuit of world class education, cutting edge research and groundbreaking innovation.


But TIMCO is focused on finding and partnering with the best investors across the globe. No matter the market environment, no firm is too small, too young or too noninstitutional. If you or someone you know is currently in the process of starting a fund or recently launched, please email partner at Temko Doug again, that's partner at MIT EMCO Dorjee or discover more on their website. Temko dug.


Some of MIT's best partnerships have been initiated during challenging market environments, but Temko looks forward to hearing from you.


Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best. This show is an open ended exploration of markets, ideas, methods, stories and of strategies that will help you better invest both your time and your money. You can learn more and stay up to date. An investor field guide, dotcom.


Patrick O'Shaughnessy is the CEO of O'Shannassy Asset Management, all opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of O'Shannassy asset management. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of O'Shannassy Asset Management may maintain positions in the securities discussed in this podcast.


My guest today is Michael Dempsey, general partner at Compound Ventures. Michael invests in a broad range of areas, but has a unique talent for combining brand building and direct consumer relationships with technically demanding sectors. Our conversation covers the rise of virtual influencers, robotics and how to best identify key inflection points in the evolution of new technologies. I hope you enjoy my conversation with Michael Dempsey. So, Mike, I'd love to begin with a simple thumbnail sketch of your career to this point.


Describe for the audience how you got into the business you're in now and what the major stages of your career have been to this point.


I'd start with in college, I spent a lot of time thinking about finance and trading a little bit and really became obsessive with markets broadly after school went to work at a hedge fund, doing everything from long short on the public side to some derivatives trading to eventually some cross-border private equity as well from Asia to the US, and eventually started to do a little bit of seed investing there, which was a little off for 2011 hedge funds doing seed stuff. There weren't a ton of people doing it at that point, really, like fell in love with the seed investing stuff.


I think that was both intellectually stimulating to me and still had the level of paranoia that public markets probably, which is there's always someone working harder to destroy you on the other end of a trade.


I wanted to understand the startup ecosystem a little bit more.


And so I joined a company called CB Insights pretty early on, a company's most well known for the newsletter. Me and this other guy, Matt Wong, would spend all day looking at all the data coming in across private market financings, fundraising, app store ratings, etc., and basically figured out what should we write about today? And we would write 12 to 15 blog posts each week on data driven venture capital. Private market insights did that for a couple of years and really saw the team scale.


And in that time, really two things informed my view a lot about my career. The first, as I got to sit in this really interesting seat, I was talking to a bunch of founders. I can do a bunch of investors, a lot of corporates, but we're never really, really deeply digging into areas. And I think the first thing I came to notice was obvious example of everyone's kind of standing around a pool waiting to see who jumps in first and venture.


And then once one person does, everyone does. And we saw that both on the data side, but also just on the topic side. Secondarily, I became really obsessed with a lot of these more kind of out there, deeply technical type areas. And those things were at the time like robotics, machine learning, computational biology, VR space. And I realized there were a lot of people focusing on that and there weren't a lot of people really going deep on those areas.


And the idea of people trying to pattern match lessons from the Internet or mobile into those categories just felt really broken to me from a lot of the hedge fund background. I kind of thought a lot about taking this very research first approach really just started to write a ton about that and started to really focus on that and make that better for my career of the things that I really cared about in the future I believed in. And that eventually led me into a venture where I am now at compound and relates a lot to how we think about investing.


There's so many different areas we could dive into today. I went back and read your whole archive of stuff you've written. It's cool that you started your career with the 12 to 15 posts a week. It's pretty intense output. So obviously you developed that muscle early on as I was. Think about how to structure this.


I thought that beginning with your thinking on inflection points as a source of opportunity and therefore something that people should pay very close attention to would be the right groundwork for all the different other topics that will cover. Maybe you could begin by just introducing the concept, the who, what, where, when, why of inflection points, why they matter so much. And then once you do that, I'll ask for some examples just to ground the audience in the sorts of things you mean.


Inflection points are basically the moments that things are changing, and that's something that I think a lot about and we think a lot about as a firm when it comes to investing. And I'd say specifically when covid started to happen in March and April, a lot of different people were kind of saying the same thing of great companies are built at every time and downturns don't matter for technology. And they talk about the Uber story of Uber being founded in 2008 and Airbnb and Dropbox.


And there's prior parables from the twins. And it's a really great thing to say when the world is melting down around you. But as someone who's a little bit obsessed about the fallacy of Alnwick data, as we call it, it's insights I just started thinking about more of like, OK, what actually happens during these times or actually the drivers that changed the trajectories of these businesses to be able to be built during this time. And I think that the economy and markets are one thing.


What is actually happening in technology is another. And so really just went down a rabbit hole of reading, kind of both news articles from back then a bunch of research papers on the different ways in which businesses have emerged over time, which industries have emerged over time and started to dig into what actually happens when things change. And what I saw was across twenty years, like, yes, there were companies that survived at each of these massive downturns, each of these scorched earth times or capital markets on the private side.


But I think what really happened was they were pretty strong catalysts that drove a lot of the growth of these companies. And you could look at it from your either the Internet usage continuing to proliferate and continuing to mature within developed nations to it becoming more global as well as corporate structure. Like each of us launching in 2006, our high speed Internet or launch of the App Store, what it kind of looked at was the different types of inflection points over time from infrastructure level to distribution level and then eventually to technical level, as we call them.


It kind of led me down a rabbit hole, as many things do, of what are the takeaways that we should have when we want to look and understand what is a real inflection point versus not. There's a lot different avenues to dive into there.


Maybe we can talk about what is and what isn't the fool's gold. Type of inflection point that looks like one, but isn't something like Krypto would be an interesting example of that. It looked like in twenty seventeen we were going to have this unbelievable explosion of development and creativity because of a new technology that may still happen. Of course, maybe this happening, but certainly a couple of years afterwards, it seemed like we'd gotten overexcited. I think about inflection points as like unlocks.


They make a lot of other things previously impossible, possible. How do you think about the difference between a real and a fake inflection point or what looks like an inflection point where it doesn't turn out to be one?


Yeah, I think a lot of it is something a science project or a venture scale business. Is there real scalability and production level processes that can be created from this change?


And sometimes there are companies that are creating their own inflection points. I'd say SpaceX is a great example of that. They spent a lot of money in R&D over time. They eventually are going to be able to be the company that unlocks the private space economy due to their ability to drastically lower the cost of launch. And I think that's going to pay off heavily for them. And they're specifically a really interesting company because all of the R&D they did along the way, they actually had a business use case for.


And granted, there is a bunch of dynamics there on the regulatory side that smarter people than I have litigated around why that type of exact business within aerospace was able to be created. But I think SpaceX specifically, they were doing things where they were losing rockets, heavily expensive R&D, but they were still delivering a product during that time, which is a very rare point. There are other times where you see a false signal and you say, OK, this must be the moment in time, right?


You can say that's the rise of crypto in twenty seventeen. With all of these all coins, everyone was like, OK, this is it. But then if you really just stopped and litigated and said, OK, well, what can we actually do? It still takes a really long time to clear a transaction. I remember sending Bitcoin took thirty something minutes and I was like, this can't be the future at that time. And granted you could pay more to push it forward.


But that doesn't necessarily make sense to me. And there are all these other alt coin use cases as well. Taking a step back. The large tech companies have become so acquisitive and so heavily focused on R&D and so heavily focused on continual expansion and new categories and understanding the existential threats that they are faced by new interests, that they're very good at creating these moments themselves. The examples I'd point to would be Oculus related to Facebook and cruise related to GM, where Oculus gets taken out for North a billion dollars by Facebook.


Zuckerberg later in his memo that he wrote about unity. It's very clear he has a view of how this plays out and he has a view of multi decades. Right at the time, though, doesn't have that context. And they do the demo and it's kind of mind blowing. And they think this is a really interesting, really compelling new technology is going to change how computing works forever, I think too obsessed with the novelty of it and the idea that there was this massive outcome that they didn't really actually litigate.


OK, what are the processes and compute requirements? What are the cost requirements to have this be large scale? What is software engineering that goes into this to make the experience as good? If you were in VR early on, Hallmark was how many times you got sick each time you tried a new demo. That happened a lot. And it's because there's a hard engineering problem. And the venture community and I think the broader tech community became really obsessive with VR for a few years and then realized what Zuck had been saying all along, which is this is a multi decade bet.


This is not I'm taking a company out and we're scaling it right now. I would say that that specifically is something that is very unique to this set of incumbents won and also speaks a lot to how you have to really understand what some of the existential threats to larger players are to understand what is actually happening in those markets. I'd say Cruz and GM would be the other one where that acquisition, again, north of a billion dollars kind of out of nowhere, was also something that a lot of people thought about and said, oh, autonomous vehicles must be ready.


We have wammo and what the five, ten systems team have been doing there and we got crews. They're going to solve this. And what you learn as you talk to people in that industry is, one, the amount of diligence that gets done on some of those types of deals is actually quite low because there's so much famo being driven by other players. And again, think about existential threats. You look at OEMs, you have to have a core belief when it comes to autonomy that there's a going to be some amount of time between when the first person solves autonomy and when everyone else does around it.


Commodity's, if at all. How long you believe that amount of time is, is how existential it is for literally the entire industry. The counterpoint of everyone who talks about Elon and Tesla being valued more than the entire automotive industry, you could say is, well, if he solves autonomy, then that's a pretty long existential threat that you can destroy a lot of enterprise value and captured have that value occur to yourself very quickly. In that case, again, people kind of overestimated and over bought into this narrative that these technologies were ready without actually thinking first principles.


OK, if we take this to a different city or a different path or if we need to throw a ton of compute at it or one hundred twenty million dollars a year, engineer in some cases at Google are actually building something that is scalable.


And I think those examples show no, the trying to find these things in terms of what they unlock is another interesting way to approach this. And maybe this is an excuse to get into what you call computational creativity, because I'm looking at that cool chart from one of your posts on inflection points where it shows the price history of the Nasdaq, the tech index, and it charts major inflection points along the way, such as Internet penetration in the early 2000s or the launch of four or five.


The launch of Bitcoin RWC, like these are all things CRISPR is another one that I want to talk to you about today and the sequencing of the genome and how much that cost. You mentioned space. All of these things are sort of platform launches that enable creativity on the other side of it that didn't exist before. So maybe begin by talking to us a bit about this idea of creativity being a key component of why these inflection points matter and what that means today.


Because I think we're faced with certainly as I watch an explosion, potential explosion of creativity, as some of the tools that were only available to very expensive companies now might be democratized access to just about everybody.


One of the really interesting examples is even something like Planet Labs, which is a nano satellite company that images space multiple times a day. One of their most creative things was being able to look at all of the R&D that was going into the mobile phone wars and say, OK, what are the ways in which this technology is applicable to doing something else? And this is called a kind of like a domain shift inflection points. What they basically did is they figured out they could build significantly more cost effective satellites thanks to the R&D that companies like Apple, Samsung, HTC, Nokia, cetera had been doing for years by throwing a bunch of these cell phone sensors onto a satellite and figuring out also that the ability to launch these things into low earth orbit was going to continually fall down and cost.


That's a really creative idea. You have to have, again, a different way of thinking, OK, there is something that is being unlocked here and there's a moment in time that we need to understand what are the tertiary or secondary effects of what all these other people are doing in mobile phones to try and make consumers happy and how that then can relate to entirely different industries. I think with computational creativity broadly is something we think about these intersections of these industries.


And luckily, machine learning is somewhat of a platform technology, we could call it, or more horizontal focused technology, despite it being viewed often as a vertical. The path to getting there for us and for me really was something that's really interesting, where early on had started to hear a lot about how a lot of these teams post crews were starting to spend more time using deep learning to understand perception. How do you see the world? How do you basically figure out what is the car seeing?


One of the things I started to do is dive deeper and deeper into just various deep learning papers, as well as the bleeding edge of that technology. And as you learn and spend more and more time, you start to see, OK, there's all these actually adjacent types of machine learning. And one of them was this guy, Ian Goodfellow, who was a researcher. He's now at Apple. Previously I was at Google and opening I who wrote this paper called Generative Adversarial Networks.


What again is is it's a Duling machine learning model where you have two different sides, which in the simple sense, one tries to fool the model into thinking something is real, the other tries to figure out what's real or fake, and they go back and forth and do until they reach parity. What they're usually early on used for was generating all sorts of images. The most famous example now is these faces. Is this person real or not? It's improved immensely over the past few years, just from a resolution perspective, from all these different use cases, et cetera.


One of the interesting things, though, is that in the University of Tokyo, a lot of researchers were doing it to do anime face generation. And they were kind of saying, we have this data set of all these different types of anime characters. They have their own facial properties that are actually different than humans. But like we think this is kind of a fun, interesting research project. And there's a lot of different industries that could emerge from that.


In the gane rabbit hole. You just started seeing more and more people generating like these animation centric tools or animation centric use cases. It kind of got me thinking a lot about if machine learning is ultimately used as a tool and if there is a tool that can create lots of scalability and also enhance creativity in some way, that's really, really compelling. And so what I did was spend way too much time doing a deep dive into understanding the history of animation and understanding what are the bottlenecks of these different studios have.


And we helped start a company in the space that was kind of a machine learning first animation studio. And the company has since raised more money and is still very early, but has been a really interesting example of just seeing that end end process of going from a Tokyo research paper to a fully formed, venture backed company and in the process really starting to understand the historical views of how technology is and wrapped into a single industry that is highly, highly creative.


And what I learned there was that the film industry broadly, all the technical innovations have basically come from this idea of how do you build something that is servicing clients and how do you spin out that technology to create an independent technology company or fully vertically integrated approaches like a Pixar. They have RenderMan. They also have some internal tooling Pixar films. You can tell it look, different things they do on the technology side.


On the research side, it just leaps and bounds above what other people do. And so that was kind of a close pocket into the animation thing. What doing that then Sol was OK. There's a lot of other use cases that all sorts of creators are trying to solve for, and they can be fashion people, they could be music people, they can be visual artists or they can be designers. And it's this idea like if we have this continual forward looking idea of these individual creators, how do we give them scale?


How do we drive the cost down? Nobody build better tooling. A lot of the thinking that we had was that machine learning is going to continue to influence the creative process across these different types of arts. Essentially, Adobe is doing a lot of really interesting work, but. It's likely they are going to need to have some new primitives, new ideas of what you look like and and even new innovations on how much do you build internally versus let the community decide.


Because the thing that machine learning has proven is very good at is researchers continuing to break the state of the art every two to three months. In some cases, it doesn't make sense to try and build the state of the art yourself internally.


We could just zoom in on an animation as a special topic because everyone will be familiar with it. I've seen a Pixar movie, everyone seen a Disney movie. I think fundamentally the technologies behind the evolution of animation are a great way of telling the story of technology itself. So maybe walk us through that history and the most interesting points from your perspective through to today. The first form that we saw of animation was this idea of the super laborious process. You have to draw every single frame you stitched together and it was incredibly expensive and no one could really figure it out.


The first big technological breakthrough there was in what's called cell animation, which is like, OK, we have this background. How do we now overlay a single moving image that we have to continually redraw? And can much more scalability kind of create multiple frames for longer periods of time? And that really was the big thing that broke through into creating Snow-White among a bunch of other Disney centric films. And again, if you look at what was happening there, a lot of these things began to happen within Disney.


You had the multiplying camera, which allowed depth of field in some ways, and not having this very flat feel, we then started to see skipping forward is moving this process into digital workflows. And that, again, was an internal project that Disney had built for computer animation. It drastically increase the scale, the pace and also the ability to collaborate on different types of work. And that took over 60 years from beginning to end of that even happening.


And I think where we are today is, again, this shift of now figuring out how do you make it so that there isn't a team that has to do 10, 20, 30 people. That's to work on a given piece of content in order to get something that feels very premium and feels very complete. And I think that's where things like Unreal engine and blender and other pieces of software that have really figured out, like all of the processes that a single artist needs to push their view of creativity into their workflow.


I think that is a future that we are now living in, and that's with things like real time rendering. If you look at animation broadly, you see this arc of the scope of which people wanted to create beautiful art to them, the scope of people wanting to scale that art to now the scope of people wanting to do beautiful things that can scale and doing them with fewer and fewer people. The key point of the post that I wrote in animation was that animation was eating the world as that was the Tongue-In-Cheek title.


And I think we've seen just like an insatiable demand for this type of content, especially as creative people can figure out how to break it out of a kid's medium. It's been something that is now for adults more and more, I think that expands the tan material and all of the different buyers of this Netflix, Amazon, Hulu, etc. if you look at their slate, are just pouring money into animation because of the retention dynamics within families, because of, again, the ability to take that IP and spin it out into other adjacencies that are also digital.


You can think of gaming as an example and specifically with this year the ability to create animation during covid you don't need to be on set around a bunch of humans has proven a massive inflection point, if you'd say so, within that industry. So the lessons really are compelling as how creatives think about it. And I will say the last thing, though, is one of the lessons of building a technology centric animation studio early on was how do you just recruit the right talent that actually wants to do this?


You're doing something that is augmenting both their workflow, but also their idea of what is perfection, what is the pursuit of greatness in their craft. And not a ton of people are super thrilled with you skipping their intricate steps of creation, using technology. Some of the things we would see early on as we built a neural network that could do automated thinking of our basically taking a paper sketch and turning it into something that had harder lines. It was easier to then be colored and actually used in animation.


And a lot of the artists would say, you know, if this isn't one hundred percent, exactly how I would do it, I'm not going to use this despite it cutting off three to five hours of their time, perhaps because it is such a romantic craft. So I do think one of the things that you do have to think about when you're working in these spaces that have a pre-existing human component and you're trying to bring technology into them, is how do you make sure that you understand all the stakeholders and their emotions?


And there's this saying you can't bring facts to a feelings fight.


So I think that is the big lesson that I learned, that at least the title, obviously damage to the software eating the world concept, which I don't think anyone would argue with now, probably people would say, wait a minute, that seems like an unfair analogy, like the way animation is as big as software. So I'd love to challenge that idea and describe what would have to happen for it to be something on the scale of how much software has impacted the world, what sorts of things might animation, quote unquote, eat in the coming decade that might surprise people that are thinking about this for the first time?


The most interesting thing about animation is like you can build a relationship with an animated character in a significantly faster time frame. That allows a lot of really interesting things because you're. Ending both judgment on Workingmen's, we judge each other and when we see each other and we interact with each other, we have a baseline of what we like and we don't like with animated characters. Often you don't have that same kind of baggage. And so even in World War two, Disney and Mickey Mouse was used as helping to sell bond stamps at the time.


That's a very influential type of piece of IP. And so I think if you think about just the idea that IP is going to continually be incredibly valuable, this idea of this connections you can build with these different animated properties, I think that really you kind of can go two ways. You can say, one, the technology is going to continue to increase where we can replace humans across everything that humans do from a artistic sense, acting, influencing the machine learning, side voice acting, anything really on the more pure play animation side, you could say both the tool sets between gaming and animation and pretty much converged are beginning to converge much more materially in most people's workflows.


You can make the argument that any type of entertainment influence or visual art will be ultimately done through these pipes. That speaks to a lot of why some of the value that is occurring within companies like epic games that own unreal is really material.


So I think it's, I would say, a more long term bullish on the enterprise value of software combined than animation combined. But I do think people don't necessarily understand some of the dynamics that will relate to both an insatiable desire just on the content side of people to buy these properties, as well as on all the adjacencies that exist within them.


You mentioned the sort of what I'll call flat examples of the fake faces that have been generated, which are really stunning. It's creepy to go through like a carousel. All of them or something just looks so incredibly real and they don't exist. I love the Twitter Callick. This person does not exist. That's a great one to follow if people haven't and want to see these. But more interesting is what I call the flat but the alive version of these things.


I'd love you to give an example to really hit the point home of a digital celebrity. You can pick whatever you want, maybe tell the story of how it was created, why it's interesting, and maybe the scale of the impact that one of these examples has, because my guess is people won't appreciate it. Sounds silly to say like we're going to replace all the regular influencers and celebrities and actors and voices and stuff with digital variants. But I think there's some interesting existing evidence that this might be possible.


There's two approaches and two examples.


The first, the most famous one is Hatsune Miku, which is almost UGC centric digital celebrity. And so she's a performer that has kind of an anime look, is very massive in Asia and is starting to actually become quite big here. The interesting component of how the tsunami was created was it was based on this idea of something called vocalised software, where they wanted to try and figure out how do you create music or other types of songs and sounds for this character and get people to use the software more and more and what the approach that the creators took for her was, OK, we're going to basically make it so that in this celebrity, the audience sees a part of themselves in it.


It's almost taking the concept of crowdfunding where you invest in something because you feel like you're part of the story on steroids, like to one hundred X where people will go in concert and they can see theoretically the clothes they design for it's anime. Cue the songs that they made, the dances that they made, and it'll be a holographic performance. And she sold over a billion dollars worth of merchandise and is massive. It's really this kind of like core like network effect.


Of the more people that you get to care about the celebrity, more that will create for them, the more dynamic they will be and also the scale that they can have. Theoretically, pretty incredible because nothing to stop you from seven PM performing in Tokyo and 8pm performing in L.A. at nine p.m. performing in Boston or all those places at once. And also the economics of that business are materially better than I would say, like humans are so beautifully volatile.


And like I think that is something that people start to realize. And I guess that leads to the second other, for example, that both a company that we helped start and we had an influence here called Astro and then on a larger company, another company called Brad has an influencer named Kayla and thinks that Kayla is a really interesting example where it's almost like the vertically integrated approach, we'll call it, where Trevor McFadden is.


And Sarah both had a very core view of, again, working within the artist management perspective. And humans are really volatile. And I wish we could cut out some of this volatility that makes it difficult to scale them. And it's difficult to scale them in two ways. One, if you spend any time in L.A. broadly, there's a lot of people who want to continue crossover into different industries, influencers who want to become YouTube ers or the others want to become actors or different people trying to cross over.


And what you really learn to realize and what people in town and your business will say is like these people just don't have all of those talents. They have one. And so what McKayla has done is it's a digital celebrity and lack of a better word. Synthetic media, 3D rendered the process. At least last I knew about it was they actually take photos of humans. Then they wrap them with a digital skin that makes her look somewhat fake. Also, anonymize whoever the humans are taking pictures of are she is an influencer and she can do anything again.


She can be in New York, she can be in L.A. She has music. She can be a runway model. She can change over time. They don't age. Well, you can age with your demographic. You can continue to capture more demographics and also we were originally doing on some of the studio side stuff is you can create this flywheel where if you can build an audience with one given influencer, you can then spin out adjacent influencers for different niches and also different types of skill sets pretty quickly.


They always could be controlled by like your internal creative powerhouse. And what it enables you to do is build a portfolio of IP and also tell a story across that IP if you want. Broad has done that with the Mikhaila property, that idea of your being able to fragment talent, being able to age up and down, being able to capture all sorts of different types of markets, and then you either even go a layer further, which is you could use things like machine learning to scale.


The creative process for these types of IP means you're going to just like materially better economics and also theoretically, if you do it right, materially, better ability to connect with a larger audience. I would say the difficulty in that, though, is that you do have the uncanny valley effect where McKayla is.


She looks like a human. And so it doesn't look exactly like a human, though. And that kind of erodes a lot of trust. And you also have some social dynamics where early on that company kind of lied about whether or not she was real. And when she was buying into the story a lot and especially depressed, they've done a few different things publicly that has annoyed some people. And I think you do have a lot of interesting social dynamics as this becomes more widely accepted or known that you have to be very mindful of.


As we know more than ever in twenty twenty one, misstep can really destroy your IP or your career. All of those things are really compelling, but do take a layer of thought in the creative process.


Say a bit about what you think the most talented digital celebrity creators will share in common based on what you've done so far. I think it's just like a understanding of what markets are are drastically underserved by this kind of aesthetic inflation we've had on all sorts of social networks. I think that that is one of the things that matters a lot, which is despite the Internet being a place where anyone can get online at any time and create content, there are still a both aesthetic inflation that happens towards a lowest common denominator and just entire parts of audiences that are left out.


One of the things that we notice internally at châteaux the company was there wasn't a lot of content for women between the ages of 13 to 17 are pretty like untapped area because you had these 18 plus and you had these 12 and under pieces of IP. And there wasn't anyone speaking directly to those people as much. Rozenes high volume. And I think we've seen similar dynamics around different types of ethnic backgrounds as well. And so I think like that understanding is really interesting.


And I would say to you, credit for McKeyla, she's kind of a racially ambiguous character and that can work in both their advantage and their disadvantage. Right. You could say that they're playing it safe, but you can also say, well, they're not just trying to be someone that looks like a Kardashian or someone that looks like this prototypical view of what Hollywood thinks. Again, I think there's very fair criticisms on both sides. But I do think that core idea to then figure out, OK, how do I speak to this audience and how do I realize that this audience is starved for content that is high velocity and big, is really valuable creative process that people should go through more when they're thinking about creating IP?


I love how the Internet has been in many ways all about the service of niches. It enables serving small audiences or market sizes. And that may not make sense for real humans. But you can just spin up a celebrity. You can serve a niche more deeply. I love that concept. So, I mean, I'm sure we'll see it in all different ways. How do you think about consistency at a point time over five years? Mikaela's personality consistency is probably reliable, but since there's people behind her, how do you thing about the longevity of these potential celebrities, or do you think it even matters because some of the best IP is IP that is valuable forever.


And I guess Disney and Pixar are great examples of that. But there's others too. But real celebrities, they have to be consistent because they're a person. How do you think about the persistence of these things through time and how that will shake out?


I think if you look at any of the, like, historical animation studios or I think a lot of IP houses in general, did you have this creative genius concept? And so I do think you bring in key man risk a lot. I think the thing Pixar is at least been good at continually talking about is their brain trust and expanding that brain trust over time and trying to more recently bring more diverse thought to it. And obviously, John Lasseter, he was kind of the core of that early on and since left the company for obvious reasons.


But I think that idea is one of the core risk. But I think it's the risk. And, yeah, as you mentioned, like all creative pursuits, is if you can build something that is this kind of unstoppable force where you go from, OK, we have established what the brand guidelines are, what the books are, what this person's profile is, which is ultimately what you are doing when you're creating a piece of IP, is you're creating someone that has core principles, core beliefs.


One of the things we think about is like the three goals, the goals that they have individually, the goals they have for the world and the goals they have for their future and their loved ones. You kind of have these north stars that you create when you create a piece of IP that said, as with anything like when people start to fall out of favor, if you start to slow growth or whatever, all hell can break loose and people can start breaking those rules pretty quickly.


Where in humans, like you said, they're humans, they are who they are.


The people want one more.


What would be one company that you would have them go investigate, read about, learn about honestly, maybe Riet, what they're doing with League of Legends and the idea of expanding IP outside of that, I think is just different. It's really interesting. And the other, I guess would still be epic games and unreal. Just because of that, I think is a very core piece of infrastructure.


You give me my perfect transition. So. Well, McKeyla and these other examples are broadcast examples, celebrities with a one way relationship. Obviously, they don't personally know anybody. Gaming is sort of this other interesting application of rendering animation and visual arts and visual technology where it's more one to one. Right. A lot of interaction and games these days. How much do you think about gaming and the future of gaming as something both interesting for the world and something that's investable at the stage that you invest?


There's two views.


The view we've taken is we think there's interesting infrastructure of all things. There's another view, which is the studio side, and there's larger funds who are deploying tons of money into similarly how people bet on pieces of IP writers betting on studios over and over again as a small fine. It's not something that we're interested in where we've come across and we've looked at game engine architecture. And the infrastructure broadly is there are two massively entrenched players and immunity and unreal.


And both of them have pros and cons on the creation process. And they also have really unique moats. Unreal, specifically with the idea that four nine is such a cash cow that it can then fund a lot of the other parts of the business over time. And they can also basically bring the cost of distribution down with the epic game store to whatever they want to basically just continue to eat market share from other players. I think it's a really interesting.


I used to have, but I do think that both of these companies still are like very stretched in terms of what they're trying to accomplish today without being the focus too too heavily on what the future of gaming looks like. And so where we've been often looking, where we've been thinking more about is this fragmentation of these engines, where right now they're like these very bundled things. And basically you have this bundled rendering engine and a few other adjacent use cases that help people create games within unreal or unity.


But then you have all of this custom engineering work that goes in over and over again for every single game studio.


We've looked at a bunch of different companies that are thinking about, OK, if we could build the ultimate store, essentially back end the ultimate in-game economy, infrastructure, all these different things, matchmaking, etc., does it start to look more like software development where you're piecing together elite parts of the stack and creating the best experience while still having the creative part of the game studio being what does that front at the front of the game? So that's really where we've seen opportunities.


And I think that there are other investors who like the game studio side and think this is a macro wave. And so you should continue to invest in games that can become platforms and roblox. We're seeing just the demand in public markets and how these things have risen. Both roblox unity and I think on the private side, even in the past six months, is almost doubled in valuation.


I think all of those things is still very good from an investment lens, but from us of we're thinking about futures we believe in and six to 10 years from now, not three to five. That's what we think about within gaming.


I'd love to pivot hard away from Bitz towards atoms and talk about robotics. You mentioned earlier your investigation of Abey's. And I think robotics is an area that my guess is people don't think about all that much. I think there's been a lot of false starts in that space and progress. Maybe it hasn't been as explosive as other software or Internet related stuff. So it's not in people's minds as much what is exciting and interesting in twenty twenty in the world of robotics.


So I think you're right. I think robotics companies generally what it looked like was maybe in twenty sixteen, maybe even before then, what we saw was everyone saying, oh, robots are here, automation is coming, it's going to work, we're going to able to do this across a bunch of different industries. These dull, dirty, dangerous jobs are going to be no longer great. And again, back to the example of these components being subsidized by mobile phone wars.


That was a key driver for a lot of these robotics companies finally being able to build robots that are one hundred thousand dollars, not a million dollars and different types of autonomy coming on that allowed them to truly operate in more dynamic environments. Understand what they're doing. I think what we saw was a bunch of companies go out and say, we're going to build robots that do X robot as a service, as the model, essentially, where you build a robot, you lease it to someone and they say, we'll pay you two thousand dollars a month to have this awesome apple picking robot or this awesome robot that sorts packages or whatever.


Think the reality those kind of back to that point of false inflection points and understanding when something is production level versus research level engineering is a lot of these customers weren't really willing to shift and make a meaningful bet on robotics within their workflows. And that could be two fold. One, it could be massive capex expenditure on changing their factory or on changing their store or something like that. The other could be the uptime or the percentage automation that actually happens.


And there's a saying within robotics, robots are great, but five figure robots are often the cheapest option, which is humans. The thing that we saw was a bunch of companies go out, raise money with this promise, build pretty interesting technology that was maybe 90 percent reliable. Companies would pilot. And then when it came time to renew and expand these pilots, a lot of them, the companies would say, yeah, I just don't really know if I want to spend seven figures a year and really start to meaningfully integrate this in my process and be reliant on automation, as you call it.


I think that we saw a bunch of companies continue to just tread water and die because of that dynamic of you have to convince these often entrenched players to adopt this leading edge technology. How we then think about that space is we like doing the hard things. We prefer to have a full stack robotics company because we are going to drive efficiency. We're going to have dominant economics relative to our competitors. And more importantly, back to the SpaceX example. We can perform R&D over time to get from 60 to 80 to one hundred percent or ninety nine point nine nine percent automation.


But we're going to capture value as we continue to increase that. We invest in a company called Onofrio, which is at its most basic sense. Early on was a robotic food truck. They built a highly fault tolerant, highly cost effective robot, could shove it in a sprinter van. We could bounce around, it could be in a van and still to deliver automated food consistently over time.


And they were both the brand and the robotics company. And it was a very unique set of founders, again, at this intersection of two creative and deeply technical areas. I think that core idea is what we think is more and more compelling because you can short circuit this candidly, just annoyance of dealing with these go to markets. And secondarily, the other thing, too, is infrastructure, things that need to happen to at least allow people to feel more comfortable to adopt those things.


One is on the safety side. How do we make sure robots can actually interact with humans or be around humans? How to make sure of it? Any time human trips are laser, we want to stop everything and that can cost companies hundreds of thousands of dollars per minute. We invest in a company called Fort Robotics, which basically is safety rated infrastructure for robots, both on the telepresence side as well as managing different robots across the worksite. And then we're also investors in a stealth company that tries to do the secondary thing, which is how do companies that really want automation or how to automate companies that are building automation centric tools get to that 100 percent without spending 20, 30 million dollars over the life span of a company just to get to market.


In that case, we think that human in the Loop is a pretty interesting and compelling way in which you can do that to continue to drive efficiencies for your customers or expand into other adjacencies. And so I think the robotics industry right now has been at this tough moment where VCs have this belief because we all have science fiction, we're nervous about all that stuff, but have been disappointed repeatedly. And some of them have lost a lot of money. I'd say now we at least are starting to see business models that make sense or maybe are more ambitious.


But also secondarily, I would say with covid, we have seen a pretty massive interest in automation broadly because we've seen just how fragile our companies are to their labor forces. And I think both from a the fragility component that we saw in twenty twenty as well as from a PR component, companies now are going to be able to make that shift and say, hey, we're going to introduce automation that might replace some humans, we're going to reskill those U.N. workers without taking a ton of flak for it, either because they've already laid off a bunch of people and they've dealt with that blow or because people just realize this isn't scalable necessarily.


And some of these more hybrid situations have a ton of humans.


It's kind of an interesting parallel to the scalability of a celebrity, right? A lot of the same theme. Right. Which is like how do we rely less on humans? That begs all sorts of questions that we will go into around UBI and employment and inequality that I think are really important, but probably beyond our pay grade. I'd love to talk about staying in the world of atoms space. You mentioned earlier Planet Labs, which is leveraging other technologies, blanketing lower that orbit with the ability to map the Earth.


And obviously everyone knows space X, what's a click or two deeper than those high level which become cheaper to launch stuff into space? That seems good. But why? Why is lowering the cost of launch and increasing the amount of technology off our surface of the Earth? What might that lead to that matters to people over the next five, 10, 15 years? Jimmy Crawford was the CEO and founder of a company called Orbital Insight. He has this concept called the macro scope, where we seem to clearly understand the value of having a microscope and being able to zoom in more and more and more and more.


But we don't really seem to appreciate as a society the value of being able to zoom out more and more and more and see what is happening across a broader scale and being able to do it in a repeatable sense. I think a lot of things related, just like how do we observe the Earth? And the honest answer is a lot of these have been humanitarian and government related thus far. There's another type of technology called synthetic aperture radar, which basically solves the question of if satellites like planet labs have cameras they can't see through clouds.


Often a large portion of the world is covered by clouds at a given time. Synthetic aperture radar can see through clouds, can see easily at night. And so you can actually see a lot more in this idea of having a revisit rate of the earth. There's also the stuff on understanding weather patterns better, which is for a bunch of reasons, going to become more important to the world over time as we've seen over the past few years. I'd say the thing that is most interesting to me as someone who thinks launch is a race to the bottom and thinks there's some really interesting CALM's infrastructure that Space X is clearly working on.


And talk to this guy who at space industry veteran, and he once told me about something he calls the Las Vegas principle, which is once we get privatized space stations, we're going to have this concept of what happens in space stays in space. And it's this idea of that is the unlocking factor of truly understanding what is like the space economy. And so you see on the manufacturing side, there's a bunch of things, whether it's pharmaceuticals or fiber optic cables, whether it's a bunch of different things you can test to understand that building them in a low gravity environment is materially more cost effective than dealing with gravity on Earth.


And then there's the farther out things that people who care about, things like asteroid mining or space civilizations and all of those types of dynamics. But I think the core thing, the space has always been how do you continue to finance it over time? And I think that idea of like sequencing has been what Elon is as an operator across both of these companies. And I think that that will be the continual question is how do you figure out how do you get in this example of this space veteran told me with a privatized space station, you might need to subsidize it by sending four of them up with a couple of millionaires who want to pay to go to space and to scientists who will be able to be the subsidized version of that trip.


Then people in the space station for a week having to pay more money and getting those people to be able to do things in space over time. And I also think that robotics, robotics will play a pretty large role in space as well until we reach the point where people want to spend material time up there. Being able to remotely control or automate a lot of the processes that we have up there I think will be pretty important. So from an investment lens, a lot of how I've been thinking about it is waiting for that moment in that inflection point to see.


I think Space X is captured a lot of the value today, as well as some of these other companies you mentioned of immediate opportunities of Earth and space. I love it.


I love the macro scope. I mean, that's just such an obvious idea once you say it out loud. That it's going to be valuable to have a persistent, fine grained view of what's going on on the Earth. And it just seems like really valuable.


I never thought about it that way, especially the cloud thinks only the next area to explore is timely during 2020. Everyone's focused on thinking about health. We talked about machine learning a little bit earlier as it relates to some animation and creative tools. We haven't talked about it as it relates to health, biology, et cetera. One of the inflection points you point to on your interesting chart is the thousand dollar genome sequencing, which happened several years ago. And there's sort of a race to the bottom on that, I'm sure to that.


And who knows how long it'll be a dollar or something ridiculously cheap. The probability of that unlocking interesting technologies and research is high. What have you learned about computational biology? Is this something extremely early, meaning we have not really realized any of the benefits yet of this technology in terms of treatments or medicines or whatever it might be? What is your research taught you here?


I think that is right. I think that if I had to make like a single bet on technology broadly, is this idea of how little we understood about living organisms until a decade ago is just mind blowing, how quickly we are understanding them now? The bet I would make is that is probably one of the biggest, if not the biggest opportunity over the next decade or two would be massive value created and I think it will drastically improve quality of life.


Rallying cry. How is the point of science to outpace the problems that we create? Essentially, we creating is obviously a jaded view on it, but it is this continual thing that I think is very real and I think we saw it most of this year and I'm not a scientist. So there's people who know this far better than I. How do we continually experiment and figure out with a bunch of experiments what the right path to go down is?


And we're trying to solve a given problem in health care biology.


I think once you come to that realization, it kind of scares you a little bit because you realize all those times you go to the doctor and they're asking all these questions like they don't know the answer. They're just, again, probing towards this experimental questions of getting to a hopeful answer. And I think what we've been able to do recently has been able to automate a lot of experimenting and scale a lot of that experimentation with things like high throughput screening, which are kind of enabled by things like lab automation, again, robotics and machine learning.


So it allows us to run significantly more experiments over time. The next kind of frontier is actually truly understanding these biological components. Again, most recently, mind solved a lot of the protein structure, understanding protein structures. And so using machine learning in the way they did it was basically had three different skunkworks teams working on different approaches and eventually realized that what they needed to do is not use any traditional approach or machine learning, but actually one that took into account how scientists would think about this problem.


So, again, I think it's a good idea and a good framing of why a lot of these hard problems in these industries aren't just going to be a bunch of machine learning people sitting in a room but are going to be an intersection of these core competencies that really matter. Yes, broadly, a lot of things are going to change. I think as an investor, how I think a lot about it is that idea of what are the areas that we can see breadcrumbs through our investing process in and what are the areas in which we feel uniquely suited to invest in.


And so sometimes it's investors in a nanomedicine robot company that you inject it minimally, invasively point of care. You can steer it throughout the body with magnets and you can deploy drugs very, very precisely. In that case, we really understood the robotic component. There's a specific design. We do a lot of diligence there. And we partnered with another firm that really understands, like the full FDA process for us. It's making sure that we have the depth to understand these areas that are very complex.


And I think not trying to say what I think some of these are doing now, which is these companies are essentially tech companies. And so thank you, biotech investors, for the past 30 plus years of innovation and funding and spending these things out and understanding this business incredibly well.


We're going to take it from here and we're going to take it from here and pay twice the valuation because we're tech investors, we have different ownership dynamics and different views on upside. I think that is a dangerous game to play.


What else in terms of you think twenty, twenty? Is the genie out of the bottle year for a lot of different types of technology and behavior, like the way we work, the way we travel, the way we do lots of things, we would conduct digital health, telehealth, all these sorts of things. What other things have radically changed this year? Meaning twenty, twenty and how it impacts your investing views, the opportunities that you see. What are other genie out of the bottle observations that you have this year on the regulatory side?


Just pulling forward the future of telemedicine. And that's obviously incredible for a bunch of reasons which we all kind of know about, but we think is very valuable and has drastically changed trajectories of a bunch of different companies. Let's say second is also a macro view, and I don't actually have passed exactly how this manifests itself, but this idea that if you have an efficient and the incentives aligned enough around the capital component of a given scientific problem, turns out you can solve that problem quite quickly and you can do it at a scale that we didn't think was possible before.


You think about it in March and everyone's saying, well, the fastest vaccine ever has been in five years and it's just really hard to get these things to market and this doomsday scenario where it's going to be at least till twenty, twenty two, you better strap in was. Quickly solved and was like, no, we're just going to throw a ton of money at this problem and the entire world is pretty closely aligned to figuring that out. And turns out we figure it out pretty quickly.


And so I think that is something that should be more top of mind where we now have this approach of how do we make sure that we're continuing to think about what are the problems that are maybe in the back of our mind today that we should be trying to knock off from a scientific sense over time?


And I think that there are more and more people talking about that and they're talking about it both because now they have a prior to base that on. Right. We've done this vaccine thing.


Fingers crossed, we're really putting the cart before the horse here.


But hopefully secondarily is we have enough people that like I think now this will be more of a calling for them because they've realized just how much our world can quickly come to a halt. And the third thing is digital living. And I think that it's really been a function of two things.


I think, one, it's been a function of time expansion for things like gaming and collaborative software and things we're doing here on Zoome, where the age range of people that are now doing these types of things because they've been forced to move their entire workflow to digital, whether they like it or not, presents a pretty compelling opportunity for the TAM expansion perspective. Let's say the secondary thing, which kind of goes back to the inflection points. Marc Andreessen has said software is in the world, but everyone is kind of looked at it and said, OK, this is a bubble.


You guys are just pouring money into more and more software companies. And what we very quickly saw was actually the opportunity that a lot of these people saw that were smarter than me was real. It was a future they believed in, but it was pulled forward in a matter of months instead of a matter of years. And the upside of these businesses as you shift everything to digital workflows is just materially more. And we saw that very quickly become appreciated by public markets.


And I think when public markets appreciated that we then saw was a secondary flow back into private markets of March. And April is a pretty scary time in venture. And then maybe everyone was kind of like, oh, wow. It turns out all of these tech companies that we've been talking about that are great are actually going to be way bigger than we anticipated because all of these futures that we believed in were pulled forward maybe a decade and maybe they were permanently changed.


Maybe this was a something that wouldn't have ever happened before. Those things is what we saw kind of this year. And there's a bunch of others in other categories, but we're pretty narrow in our scope of how we think about things.


What technology trends have we not talked about yet that you think are really important or interesting? The biggest thing I've been thinking a lot about over the past year and has been routinely repeated has been more of a macro thesis. And I think that that macro thesis is something that I tongue in cheek call cyberpunk is now what I say when I say that. And a lot of my friends are kind of like every time something crazy in the world happens, I say that and they say, All right, great, shut up.


Is this idea of these themes that we've always had and a lot of these scientific and dystopian type novels are starting to show little interesting breadcrumbs. And whether it's the scale of the largest companies in the world and watching the CEOs of four large tech companies go in front of the government and kind of just say, we're here, we're dealing with this, but we're not really listening is like a really compelling, like first wave. A week ago, we had a cyber attack from another government that has been in our country for months now.


It's really compelling the way in which Trump won the election in twenty sixteen is very interesting, the way in which we're thinking about China as this superpower moving forward. All of these things relate to then like these narrow thesis areas. And so some of them are privacy, preserving, machine learning and how we think about continually guarding our data while using all the benefits of machine learning. Some of them are adversarial attacks on machine learning models. If we continue to bring automation into the world, people are going to try and attack those systems in certain ways.


And so how do we defend against those?


So it's a bunch of different little breadcrumbs. And I'd say generally how I think about investing is either going into a category and saying, OK, what are the types of businesses that should exist and will go out and find those partner founders, let them shatter our theses as well, highly prescriptive or saying what are the futures that we believe in and what is happening in the world and what does that mean for venture investable businesses? And so I think that is one thing that has been on my mind and I have a very long piece written about it.


But every month something new happens that changes what that piece would look like. So it's kind of now just turned into this private log.


I like the idea that using sci fi novels as research for what might be going on in the digital world. Right. Well, this has been so incredibly interesting. I highly recommend everybody that's found this fun or educational. Go check out your writing, because I think I talked to Sam Hinkie in recent episode about this idea of putting breadcrumbs out in the world. And you've done a great job of putting breadcrumbs out for people to get to know you and your ideas.


So I highly recommend people go avail themselves of that. My traditional closing question is the same for everybody, which is to ask you what the kindest thing that anyone's ever done for you is.


One of the things I've learned is caring about the step function events is really important. But noticing these in the moment, really individual type of things that are meaningful is equally as important. So I guess I'll say you, after a really hardest year of my life, the thing that I'm continually thinking about more not to steal your thunder is what is just the most recent thing someone did. For me, that is something that we don't think enough about. We think three years later, wow.


It was amazing that my boss gave me the chance to do this. We don't think yesterday one of my. Friends woke up at seven a.m. and had a call with me to talk through a bunch of stuff that I wanted to talk through over the past couple of days. I've had a bunch of different calls with friends early in the morning. And in my life at that times were in the moment I've said, hey, I need to chat and you talk and have dropped everything and gone.


And those are the kind of things people have done, things have done for me. And I would just say think about it every day, not just in these reflection times or maybe it's too late to talk to someone or life is passing by and it'll feel weird to reach out. I love that. What a wonderful challenge to the way of thinking about it. A great answer. Unique to seems to be a theme with you. Michael, thanks so much for doing this with me today.


I learned a lot as I knew. I would appreciate your time.


Really appreciate it. If you enjoyed this episode, you can sign up for a new email newsletter sent out each week called Inside the Episode.


Each week I condensed that week's episode to my favorite big ideas, quotations and more. I've been recommending books to members of this. Email us for years and we'll keep doing so.


In this weekly email, you can sign up at Investor Field Guide dot com forward slash book club.