The Exchanges

Q And just as a quick detour, do you want to talk about that, um, that, that, that post that, like, great innovation will look like a toy? I think that for anybody that's an entrepreneur, I think it's a super interesting concept.

A Sure, yeah, it's, um, it, it, I mean, it's really, it's not, I mean, I guess that was my blog post. The idea really comes from Clay Christensen Um, and, uh, the idea is, uh, is essentially that, um, like, if you go back, that essentially, um, a lot of technologies start off kind of half-baked, but they get better at a, at a rate that's sort of faster than kind of people need the technology to work. So human demand is sort of a straight line, and the technology is kind of coming up on a curve. And so, like, I remember Skype as an example, um, you know, when it started off, I actually worked at a VC firm. I was a junior person at a VC firm that invested in it. And in our memo, like, the key risk was there weren't microphones and computers at the time, really, literally, that was, like, the key risk, because that, you know, it wasn't clear they were going to be. Also, it dropped calls a lot. The quality wasn't that great. You know, if you looked at the time, all the business world was talking about VoIP and like these other kinds of things, and these sort of high-end experiences, but what obviously happened is it got better and better, and computers got mics, and bandwidth got better, and all these other things, and then, you know, and then eventually you had smartphones, and of course now, you know, that's sort of the way things work, so, um, there's just sort of, you know, tons …

AI assessment note: “a lot of technologies start off kind of half-baked, but they get better”

Q like this amazing, um, you know, overview of all the different, um, areas, and I think you talked about self-driving vehicles, AR, VR, uh, clean energy, all those things. Uh, so it's been a couple of years now, and, uh, I think we've all witnessed some ups and downs in the, in the hype cycle around all of those things. I guess, are you as optimistic about this emerging world?

A I am, I am. I think, you know, these things take time, right? So, like, I remember with, Smartphones. Hearing people talk about it for, you know, a decade before the iPhone came out, and I don't know if people are old enough to remember, like, I had, like, a sidekick, and, you know, Blackberry, and Trio, and all these other things, and they were, like, they were awesome, but they were just, like, not quite there yet to be, to really kind of, you know, explode and go and, and become, you know, where we are today with whatever it is, three and a half billion smartphones, so, um, I think we're in, like, when you take areas like self-driving cars, and VR, and AR, and all these other kinds of things, I think we're in that kind of period. I don't know when Ah, these things will hit, but they will hit, and they will hit. You know that because they're just getting better, you know, kind of Moore's Law writ large, which I think of as not just, you know, ah, packing more semiconductors on the, on the chips, but just all the things that happen when there's a lot of investment in an area. So, ah, you know, if you just look at VR as an example, ah, the, the price of a high-end VR headset has gone down, I think, like, 80% over the last three years, um, and you're gonna see more and more devices come out, and You start to see this whole kind of ecosystem effect where, like, Qualcomm is buildi…

AI assessment note: “I am, I am. I think, you know, these things take time, right?”

Q so let's talk about the AR and VR that, that, uh, you mentioned. So you, you were an early investor in Oculus, uh, right, which, uh, turned out to be pretty good as an investment. Um, I, I, uh, A couple of things, but first, what, um, what do you think is, needs to happen for this to get exciting again, I guess, because it really went up and down?

A Yeah, I think, I think the devices aren't there yet. I think you need a 200 dollar, some sort of consumer price point device, um, which has sort of a high-end feature, so like, and that means like six degrees of freedom, head tracking, I don't know if people follow this stuff, hand tracking, Um, you have to get rid of the cord. You know, the, the new Oculus thing that's going to come out next year, I think they've just, they've just, they've said about, talked about it, Santa Cruz looks promising, but I don't know if it'll hit that price point, but that's generally how it works in consumer electronics, is you need, like, you know, you need that sort of iPhone-like kind of experience, um, and it's not quite there. I mean, it's very, there's a lot of stuff going on. Apple's doing stuff, you know, um, Google and Facebook, and so I think it's going to get there.

AI assessment note: “you need a 200 dollar, some sort of consumer price point device”

Q Alright, so maybe one last question from me, and then I'll open it up to, uh, people. So maybe, like, tying it all together, uh, so we talked about AI, IoT, and crypto. Do, do all the things ultimately work together?

A I think so. If you look at mobile social cloud last 10 years, right, they all ended up intersecting. So, um, you know, mobile, uh, reinforced, you know, Facebook is as big as it is because of mobile, right? Um, because that dramatically increased the, the number of internet users, the time they spent online, everything else, right? In turn, social helped mobile, and that social, like, Twitter, Facebook, et cetera, is sort of a, you know, is an app that people like, they, they, they love, and they use, and drove demand for mobile phones. And then cloud was an architecture that, that allowed, you know, those companies to support, uh, billions of users. Um, and then in turn, then offer those cloud services out to enterprises we've seen. Um, so they all kind of intersected and, and, uh, and you know, and like the most successful startups, for example, like the, the, the new social companies in the, in this post mobile were kind of mobile first, right? So I think, I think it's very likely in the future you'll see a similar thing where like these kind of mega trends will all intersect. Um, how exactly it happens is very hard to say, but.

AI assessment note: “I think so. If you look at mobile social cloud last 10 years”

Q Um, so you're going to be in an 80% or 90% world. Can you, can you walk us through?

A Yeah, there's like an old joke, and it's not really funny, but it's a, it's an old joke in machine learning that it's, machine learning is really good at partially solving any problems, so like, you can basically, the idea, yeah, that's not funny, so, uh, but, uh, that's, I guess, big passes for computer science jokes, so, um, but, uh, the idea is, though, that you can sort of, like, you know, you can download TensorFlow, download some data sets, and over the weekend, probably come up with, you know, if you're a good programmer, come up with something that can do, like, 80% accuracy of whatever, let's say OCR or something, right? Um, but then the last 20% takes you, you know, a decade if not a lifetime or something, right? Um, and, and, ah, and so a lot of the strategy around building an AI startup is figuring out how to get that 20%. And I think of it as, like, there's different ways to do it. Like, one, one is just to sort of punt on it. In some ways, you can think of Google results. They give you 10 results, right? They're sort of, you know, like in an ideal world, they give you one result, but they give you 10 results, because they let the, the human figure out the last 20%, right? So that's sort of having a fault-tolerant UI, I'd call it, or something, right? That's one way to do it. Um, uh, you know.

AI assessment note: “a lot of the strategy around building an AI startup is figuring out how to get that 20%”

Q very promising company that had an early exit, um, and I think at the time you had mentioned that, uh, this was a concern about, from an investor standpoint, about investing in the space, like all those companies got snapped up so quickly. Um, it seems like we are past What's that phase now? Do you think it's a better time to invest in AI, where there's too much noise,

A or Yeah, it's a really good question. It's an interesting, I mean, so, you know, it's not a secret that AI is a big deal, right, and it's, and it's the, it's basically the main R&D line item at Facebook and Google and other places. They spend massive amounts of money. They pay huge salaries, um, and so, and, and, and moreover, they've, uh, they've commoditized a lot of the infrastructure. So, you know, the, first of all, all of the, Algorithms are published. The, ah, you know, some of the best frameworks are free, like TensorFlow from Google. So, you know, historically in the venture world, like a new trend happens, maybe you invest in the infrastructure or something, like, that's not really a thing you can do here. Plus, On the service side, they have their cloud platforms. They're sort of baking us all into the cloud platforms as, you know, frankly, kind of a, you know, just an adjacency, almost loss leader for the real business, which is selling storage and compute and everything else. And they're building their own chips and like the TPUs and everything else. So like, it's really capital intensive to like compete against that. So that makes, that kind of rules out like a whole set of investments that you might've imagined making. Um, so, you know, for the most part, I think, I think the, the way that I think you invest in it is sort of in, well, there's a couple things. Lik…

AI assessment note: “the way that I think you invest in it is sort of in”

Q All super interesting companies in the space. So, you know, yet again, the, sort of the same question, but I'm, I'm curious if there's any difference, right? It's like another one where, like, everybody's super excited two or three years ago, and that seems to, uh, Yeah, yeah. I think, um, Is that regulation?

A Is that, uh, A lot of it's regulation in the U.S. It's, uh, there's, like, beyond line of sight drones, for example, are not legal. Um, which is a big issue for a lot of applications. I think a lot of the interesting applications are commercial applications, and it's just, the adoption's just starting. It's, it's pretty, drones have been pretty successful in mining, um, so people will fly over mining sites, and the reason they're, and, and they survey them, and then they create three images, and they can check for safety, and other kinds of things. Um, and that's actually a, a bunch of companies are having success there. I think they're successful there because those people are used to heavy equipment. They have hard hats. Like, it's just like in a, um, the, I think the drones aren't quite there yet. I think they need autonomy. That company, Scadio, you mentioned is a company that builds autonomous drones. Um, so they don't, they, they, they don't have collisions right now.

AI assessment note: “A lot of it's regulation in the U.S.”

Q Okay, so that's why decentralization should win. I guess the next question is, how does it, how does it win?

A I think the way, so ultimately the way you win in technology is you build, and specifically software, in services, you build better software and services. Um, how that happens here is they attract the best developers. That's how Linux won. So why did Linux beat, you know, everything else? Because it attracted the best developers, right? Because it was open, and they could, they could build on it. It has better security. Like, you could, the fact that anyone can audit the code gives it better security. The fact that anyone can go and contribute to it makes it integrate with other devices better, right? Like, it's this huge consortium, effectively consortium. You know, very interestingly, like, the top Contributor to, uh, uh, Linux is, is, uh, employees at Intel, right? So it's not as nearly as sort of, you know, this kind of commune kind of idea that people think it is. I mean, there are these enthusiasts, but a lot of it is our, you know, corporate interest. It's almost like a consortium. Like, Intel people contribute to it because it's very important to them that they have an adjacent operating system to their chipset. That is open, right? Um, and so, you, you got this, so you basically, because it was open, and because everyone kind of worked together, you got better software, and the same idea happens with these crypto networks, where, you know, you have all these people bui…

AI assessment note: “how that happens here is they attract the best developers”

Q The idea maze of data. Can you talk about this? Basically something to help people think about where the opportunities are and how to navigate them.

A Yeah, that's not, so it's not my, so it was my friend, just my friend Balogy who came up with this kind of concept of idea, or, you know, maybe an older concept too, but he wrote about it. Um, but the idea is just that, that, that the journey of an entrepreneur is not, it's not, it's not like you just have an idea and then you go, yeah, implement the idea. Instead, The, the analogy he uses, it's more like a maze where you're kind of exploring the maze, running experiments, running into, you know, monsters or treasures or whatever, pick your metaphor or whatever, um, uh, and it's just sort of this ongoing process, right, and so, like, you hear this, you hear this sometimes in the, in Silicon Valley, people say ideas don't matter, but of course ideas do matter, but it's really that they matter as dynamic processes, like ideas are really, like, startup ideas are really sort of dynamic processes that you iterate over and, And great entrepreneurs are very good at, they, they, first of all, they enter, they enter with, ah, you know, intelligently, they go study the history, they study adjacent markets, they talk to customers, they do a whole bunch of things, and then they, as they go, they iterate, and they, and I'm sure, you know, um, You know, you talk to most entrepreneurs, and the original idea evolved over time, um, and so, it's just a nice, it's a nice metaphor, and I think of …

AI assessment note: “the analogy he uses, it's more like a maze where you're kind of exploring”

Q Um, um, and, uh, content matters a lot as well, right?

A So that's the other thing is, yeah, like, so, like, I have, you know, Oculus here, and I use it, but, like, it's basically, like, demo kind of level software right now, and you don't have, like, AAA games being made and things like this. And if you talk to the game makers, they'll say it's just the user, but you know, there's like, you know, a couple million, I think, high-end devices now out there. There's tens of million low-end devices, but you really, I think, need the high-end experience, and so the, you know, like Fortnite doesn't have a, doesn't have a VR version right now, and so, um, so you need, so it's sort of this chicken and egg problem where you need a certain number of devices out there before, but, you know, but these, these things always, it always ends up happening, you know, if you just look at the history of computing, and so, Um, it's just a matter of getting the right, the right device, and then the right kind of content partners, and then eventually, and then once you get that going, you know, I think it's going to be huge. Excuse me, I'm, I'm, I'm still very bullish on, I know it's sort of fashionable to be like AR, not VR. I think AR is very cool, but I think VR.

AI assessment note: “it's basically, like, demo kind of level software right now, and you don't have”

Q Um, what do you make of the, you know, amount of data question, so that's become the, you know, inevitable comment on any kind of, uh, AI-related panel by VCs that say, no, no, no, you can't, like, you know, Google has all the data in the world. Um, do you, do you worry about that?

A Uh, I think for some applications, like speech, you know, uh, speech is going to be, I think, very tough for non, you know, I mean, like Google and Amazon with their, you know, with their, and Apple with their phones and the, and the, you know, the, the, uh, Alexa's and things like that. Are going to have a massive amount of data. I mean, there's a lot of, sort of, technical questions as to, like, what, how much data do you need to be good enough? Um, you know, does it matter if you have ten billion, you know, data points versus a billion or whatever, you know, and how, like, at what point do those curves kind of asymptote, um.

AI assessment note: “how much data do you need to be good enough”

Q Yep. I guess some question for the Internet of Things, right, that was all the rage, uh, you know, two or three years ago. You still bullish on that?

A I am, I am. It's just like, like, I mean, I think everyone here has probably used these devices and like the Apple Watch and everything else. They, they just don't feel like they're, They're quite there yet, right? They're not, uh, you know, and like, I love the Alexa and things, but what do you use it for? You probably use it for music and weather and like one or two other things. It's not quite, you know, it's not, Um, it just, it just doesn't seem like it's quite there. I mean, they subsidize it heavily, and it's very inexpensive, and so it's gotten adoption, but I don't know how well that would do if they didn't have the business model to kind of subsidize it, but, but, um, but I don't know, I just think the devices aren't quite there, you know, you look at, but you just look at the trajectory, like the Apple Watch now has, you know, you know, it's gotten better and better, and has the, you know, uh, is getting to the point where it's going to be kind of a standalone device.

AI assessment note: “I am, I am.”

Q Any, any preference if you, if you could influence the regulators, as I'm sure you can, uh, any preference about where that should go, or?

A Yeah. Um, well, I think, I mean, like, I think, I think the key, the key concept as is I understand that, you know, I don't know, I, I, I, I'm not a lawyer, so I don't want to get too, uh, into the details there, but, but, um, I think the key, the key concept is you have enough room for entrepreneurs to build, you know, like, these digital services done in the proper way, um, but also, you know, protections to make sure that, um, Information is symmetric, that someone, someone who's not in Silicon Valley can make an informed, if they decide to buy something for investment purposes, can make, ah, you know, an informed decision, so that's where, that's where a lot of the regulatory framework, so for example, there's an important distinction between securities and commodities, um, and, ah, you know, regulators have said that Bitcoin is a commodity, which essentially, essentially the idea there is, There's no information that like I have about Bitcoin that would give me some advantage in buying Bitcoin over anyone else. Anything I know is from the internet. It doesn't really, it's a fully decentralized protocol. Everything's open. Um, and so I think that some framework where we sort of understand, kind of balance, um, you know, the, the, the, the protections for kind of mainstream investors with enough room for people to innovate.

AI assessment note: “balance, um, you know, the, the, the, the protections for kind of mainstream investors”

Q So what, what does that look like long term, something like Crypto, CryptoKitties? Go into crypto dogs.

A I think, for example, like, um, well, I won't speak specifically about them, but I'll speak about why I think the area is interesting. Um, I, I, I think that, uh, something I believe is that, um, uh, the internet should be, uh, better for creative people, uh, as a business, uh, for, for the business models of creative people than it currently is. That there are, that are, there are undiscovered business models, for example, for musicians. Um, and so I think one very interesting application of, of what you're going to see, I think, in the near future is, uh, um, so if you look at musicians today, they make, they make very little money on streaming. They make most of their money on offline merchandise and shows, and that's because that's where the scarcity is. And so I think you're going to see musicians who now make money by selling digital virtual goods as an example, which is not a model today, and that's like a very interesting new model as an example. And so, you know, and so you can say like, okay, it's just, is it just silly in games, you know. Um, on the flip side, look, if it's, if it's something that can help fund creative people, like musicians, and writers, and, you know, and game makers, and a whole bunch of other things, that's awesome, right? Yep. Um, and if you look, uh, I have another blog post about this if you're interested. It didn't get that widely read, but …

AI assessment note: “I won't speak specifically about them, but I'll speak about why I think the area is interesting.”