The Exchanges

Q Devastated. But people look to the likes of your GoPros and your Fitbits. But as we mentioned there, there are many more skeletons with, you said, Pearl, Juicero. Recently, obviously, both Struggling, um, to say the least. Um, so first, talk to me. Are the two above that we mentioned, GoPro and Fitbit, purely outliers, and is consumers still really viable today, then, for startups, considering the huge capital requirements?

A I definitely think consumer electronics as a product category is still viable, and you're going to see some really interesting stuff going on, and certainly you've had things like VR, and they've had acquisitions that are starting to, you know, that's growing a little bit slower than I think everyone would like, but also, let's be super clear, you know, DGI is just a consumer electronics company. It's a ten billion dollar company. And they're showing no signs of stopping. So I think that the, for Fitbit and GoPro in particular, people kind of like fell out of love with them. But at the end of the day, there's still public companies that you can buy into or sell out of. And there's 200 private unicorns today that wish they had the problem of being a public company that fell out of love.

AI assessment note: “I definitely think consumer electronics as a product category is still viable”

Q Okay. Talk to me. What are the catalysts that make that no longer true?

A So there's a couple of different things, you know, first and foremost, there are like incredible design tools. So, you know, my first job in aerospace was in 2002 at Northrop Grumman. And at that time, you know, CAD machines are still 15 to 20,000 dollars per computer. And the software is 15 or 20,000 dollars on top of that. And it still sucked, you know, kind of fast forward to 20 10, 20 11. And today, any 500 dollar computer can run incredibly sophisticated CAD. And on top of that, the software is Free. And then you kind of add in rapid prototyping like three D printing, and you're kind of timing for your first unit, plus things like Arduino, make it so it's orders of magnitude cheaper to get up and running.

AI assessment note: “first and foremost, there are like incredible design tools”

Q Okay. Talk to me. What are the catalysts that make that no longer true?

A So there's a couple of different things, you know, first and foremost, there are like incredible design tools. So, you know, my first job in aerospace was in 2002 at Northrop Grumman. And at that time, you know, CAD machines are still 15 to 20,000 dollars per computer. And the software is 15 or 20,000 dollars on top of that. And it still sucked, you know, kind of fast forward to 20 10, 20 11. And today, any 500 dollar computer can run incredibly sophisticated CAD. And on top of that, the software is Free. And then you kind of add in rapid prototyping like three D printing, and you're kind of timing for your first unit, plus things like Arduino, make it so it's orders of magnitude cheaper to get up and running.

AI assessment note: “incredible design tools... rapid prototyping like three D printing, plus things like Arduino”

Q there in conversation, so I'll pat myself on the back. But I think, I think, well, I mean, A, actually, let's draw an alignment. They both have extremely large capital requirements. So there we go. So we've drawn the alignment, but we both have some pretty strong feelings on space. A lot of excitement around space. You've said before that you're bearish on it. Why are you bearish on space?

A So there's two main reasons that I'm bearish on most space companies. So, you know, to be clear, we're investors in Spire, which does non-imaging based remote sensing. I'm incredibly bullish on Spire. I think Peter has built an incredible company with an incredible value proposition, and I think that they're in a really good position. But there's a couple of different things, and I've got like a whole map of this in my notes, but take remote sensing and Is it a high frequency need? Does someone wake up every day and say, I need to have this right now? And so, if you're the US military, images is incredibly important because they can do a lot of intel on that. But in general, there's this mythical hedge funds that want to buy in pictures of, you know, aluminum piles in, you know, China. And really, the reality is that the hedge fund industry is not doing that well, and the amount of money they spend on data just really isn't there. And you're also comping that picture with 10 other ways they could come Possibly get that information. The reason that I like non-imaging and what Spire is doing is because if you're a supply chain manager for a fortune, 500 company, you need to know where every boat in the world is. And if that boat goes off course, that might like mean everything to your business. And so you need to know every minute of every day where it is. And then if you look at…

AI assessment note: “there's two main reasons that I'm bearish on most space companies.”

Q 24 to 48 hours, as I, as I like to do, uh, and one thing that's blatantly apparent is your love of robotics. So talk to me. You've said before that robotics are coming fast, but it has nothing to do with AI or ML, so if we unpack that a little. What are the mega trends that make you so confident that robotics are potentially finally coming of age?

A Yeah, so there's five major kind of mega trends that, you know, are the reason robotics is happening. And, you know, I've got a good friend who says he likes to invest in companies that surf Moore's law. And what that really means is that, you know, and this is kind of a thesis from the, you know, late nineties and early 2000 of, you know, today, maybe it's really expensive to do what you're doing. 18 months from now, Moore's law says, okay, now it's pretty easy. And then You get to ride the tens of billions of dollars of investments in kind of semiconductor. And so today what we're seeing in robotics is things like connectivity, cheap sensors, advances in computer vision, storage costs, and just overall computation. All of those things conspire to make the cost of starting a robotics company a fraction of what it used to be.

AI assessment note: “things like connectivity, cheap sensors, advances in computer vision, storage costs, and just overall computation”

Q You have said before, though, that robots require hardware, software, and the cloud. Now, as an investor, often I get quite nervous when I see multiple different components and layers all within the same stack. How do you think about this multiple layer process, and does it concern you as an investor?

A Well, and I think that in general, great investments are hard, but not impossible. You know, if you want to go make a new jet engine, I don't think that's very good as a startup, because at the end, it requires too many different expertise, people with different types of expertise, on top of just too much capital on day one. Whereas now, you know, the cloud has gotten so cheap and so easy to use. Like, overall software has gotten much better and also easier to use. And hardware, the stack for a hardware company is shorter than it's ever been. And so that's why I think it's a great time to start a robotics company, because it's hard enough that it's defensible, but easy enough for a startup to do.

AI assessment note: “it's hard enough that it's defensible, but easy enough for a startup to do.”

Q You have said before, though, that robots require hardware, software, and the cloud. Now, as an investor, often I get quite nervous when I see multiple different components and layers all within the same stack. How do you think about this multiple layer process, and does it concern you as an investor?

A Well, and I think that in general, great investments are hard, but not impossible. You know, if you want to go make a new jet engine, I don't think that's very good as a startup, because at the end, it requires too many different expertise, people with different types of expertise, on top of just too much capital on day one. Whereas now, you know, the cloud has gotten so cheap and so easy to use. Like, overall software has gotten much better and also easier to use. And hardware, the stack for a hardware company is shorter than it's ever been. And so that's why I think it's a great time to start a robotics company, because it's hard enough that it's defensible, but easy enough for a startup to do.

AI assessment note: “it's hard enough that it's defensible, but easy enough for a startup to do.”

Q Can I be very controversial? Does all that, uh, incredible technology produce the computer vision, the storage, uh, the computation power? That All fantastic, but does it not require AI and ML to really translate that into a core usable asset?

A So I really don't think it does, and I would also argue that what people consider a core AI ML today is a bunch of nifty tricks, and so you see lots of nifty demos, but in terms of actual production AI or ML actually out there, a lot of it's a little bit mechanical turk. If you actually go back and look and say, well, you know, let's take a self-driving car company, for instance, they all have teams of people that are still tagging data manually, and if you go and look at Apple Application on application, a lot of it is large group of humans tagging data manually, and then sure, you use a little bit of ML to make that easier and more scalable, but every single use case you have to pick off individually.

AI assessment note: “So I really don't think it does”

Q terms of your vision for Lemnos, with the kind of capital requirements needed, but also aligned to your incredible skill set in these spaces, and with the team that you've built, is the mission with the fund to have a scaling fund that can support the companies throughout the life cycle, or do you believe this is always the optimal place to be investing in these really hardcore hardware investments?

A You know, we, we really like being an early stage fund for a variety of reasons. You know, I think that we really focus on helping our companies find product market And then I think when you look at a classic series A or series B fund, the reality is that they invest in lots of different sectors, but a big part of their skills is helping companies scale from 20 people to 2000 people. And so for us, you know, we think that being, you know, the best at early stage in hardware is really where our focus is. You know, we don't really have ambitions to go raise five hundred million or a billion or five billion dollars like NEA does, just because that doesn't fit our model very well.

AI assessment note: “we think that being, you know, the best at early stage in hardware is really where our focus is”

Q Can I ask them with those difficulties and with the kind of incredible technological advances that we've seen, and talking about the multiple layer stack there, what are the commonalities you've seen then in the hardware companies and the robotics companies that really manage this value chain successfully and well?

A I think it's about understanding what application you're going for. Now is not the time to go build what I call the final robot, and you'll see this occasionally that companies are pursuing, and it's like a biped and C-IIIpo kind of stuff that can You can just drop in and train to do anything. I think the best robotics companies started today will have like a singular use case where the ROI is obvious and apparent. So we're investors in a company called Marble, which does last mile delivery. And so, especially as e-commerce has grown, as on-demand has grown, we think there's going to be this growing and growing demand for on-demand delivery stuff. And the ability to have a robot to actually go and do those deliveries is a key critical component of that.

AI assessment note: “I think it's about understanding what application you're going for.”

Q 24 to 48 hours, as I, as I like to do, uh, and one thing that's blatantly apparent is your love of robotics. So talk to me. You've said before that robotics are coming fast, but it has nothing to do with AI or ML, so if we unpack that a little. What are the mega trends that make you so confident that robotics are potentially finally coming of age?

A Yeah, so there's five major kind of mega trends that, you know, are the reason robotics is happening. And, you know, I've got a good friend who says he likes to invest in companies that surf Moore's law. And what that really means is that, you know, and this is kind of a thesis from the, you know, late nineties and early 2000 of, you know, today, maybe it's really expensive to do what you're doing. 18 months from now, Moore's law says, okay, now it's pretty easy. And then You get to ride the tens of billions of dollars of investments in kind of semiconductor. And so today what we're seeing in robotics is things like connectivity, cheap sensors, advances in computer vision, storage costs, and just overall computation. All of those things conspire to make the cost of starting a robotics company a fraction of what it used to be.

AI assessment note: “things like connectivity, cheap sensors, advances in computer vision, storage costs, and just overall computation.”

Q Now, naturally, robotic last mile deliveries does make Think of the hub of innovation that is Silicon Valley, and you said before that Silicon Valley will dominate the future of robotics. Why do you think it is this concentrated thesis around the Bay Area?

A You know, I think that there's a lot of different reasons for that, but honestly, I think that the dirty secret of the Valley is that you always know someone more successful than you. So the Kauffman Institute has done a bunch of research, and there's this term called the Midwest Curse, and it's like, well, why do, like, companies in the Midwest typically exit early, one to two, three hundred million dollars? And it's because I think that I spent five years living in Albuquerque. If you sold a company for two hundred million dollars in Albuquerque, it's insane. You'd be the most successful entrepreneur that Albuquerque's ever produced. Whereas if you sell a company for two hundred million dollars in the valley, you're one of a dozen people at a party that might have done that as well. And I think that there is this kind of framing where, like, it's certainly got its downsides as well, but I think that the ability to see that people have gone on to start a hundred billion dollar companies every six to eight years in this ecosystem is I think kind of gives people a different perspective.

AI assessment note: “dirty secret of the Valley is that you always know someone more successful than you.”

Q Can I be very controversial? Does all that, uh, incredible technology produce the computer vision, the storage, uh, the computation power? That All fantastic, but does it not require AI and ML to really translate that into a core usable asset?

A So I really don't think it does, and I would also argue that what people consider a core AI ML today is a bunch of nifty tricks, and so you see lots of nifty demos, but in terms of actual production AI or ML actually out there, a lot of it's a little bit mechanical turk. If you actually go back and look and say, well, you know, let's take a self-driving car company, for instance, they all have teams of people that are still tagging data manually, and if you go and look at Apple Application on application, a lot of it is large group of humans tagging data manually, and then sure, you use a little bit of ML to make that easier and more scalable, but every single use case you have to pick off individually.

AI assessment note: “So I really don't think it does, and I would also argue”

Q Can I ask them with those difficulties and with the kind of incredible technological advances that we've seen, and talking about the multiple layer stack there, what are the commonalities you've seen then in the hardware companies and the robotics companies that really manage this value chain successfully and well?

A I think it's about understanding what application you're going for. Now is not the time to go build what I call the final robot, and you'll see this occasionally that companies are pursuing, and it's like a biped and C-IIIpo kind of stuff that can You can just drop in and train to do anything. I think the best robotics companies started today will have like a singular use case where the ROI is obvious and apparent. So we're investors in a company called Marble, which does last mile delivery. And so, especially as e-commerce has grown, as on-demand has grown, we think there's going to be this growing and growing demand for on-demand delivery stuff. And the ability to have a robot to actually go and do those deliveries is a key critical component of that.

AI assessment note: “the best robotics companies started today will have like a singular use case”

Q Now, naturally, robotic last mile deliveries does make Think of the hub of innovation that is Silicon Valley, and you said before that Silicon Valley will dominate the future of robotics. Why do you think it is this concentrated thesis around the Bay Area?

A You know, I think that there's a lot of different reasons for that, but honestly, I think that the dirty secret of the Valley is that you always know someone more successful than you. So the Kauffman Institute has done a bunch of research, and there's this term called the Midwest Curse, and it's like, well, why do, like, companies in the Midwest typically exit early, one to two, three hundred million dollars? And it's because I think that I spent five years living in Albuquerque. If you sold a company for two hundred million dollars in Albuquerque, it's insane. You'd be the most successful entrepreneur that Albuquerque's ever produced. Whereas if you sell a company for two hundred million dollars in the valley, you're one of a dozen people at a party that might have done that as well. And I think that there is this kind of framing where, like, it's certainly got its downsides as well, but I think that the ability to see that people have gone on to start a hundred billion dollar companies every six to eight years in this ecosystem is I think kind of gives people a different perspective.

AI assessment note: “the dirty secret of the Valley is that you always know someone more successful”

Q Speaking of kind of getting to the revenue targets as quickly as possible, it makes me instantly think about Reid Hoffman's quote about, if you're not embarrassed about your first product shipment, then it's too late. What do you think about that with regards to hardware, where it's not such an upgradable, changeable asset that can be redefined with time?

A I think it's about making sure that you really narrow your scope for your first product, because if 90% of your product works in its hardware, you have nothing. And so, unfortunately, a lot of times you'll have founders that, you know, by definition are very ambitious, and they'll include too much in that first rep. And then when there's inevitably some problems or integration issues, they can't just then cancel that second wheel or that whatever it is that's giving them a problem, because that's all part of the design. And so, there is this little bit of complexity Systems engineering, and so you have to have a good vision from day one, and so I think that you will have these, if you look at, you know, like any piece of hardware, the first one is generally kind of crappy, and then over time it gets better, but that one, that first one that's kind of crappy has to do something that gets high value for their users.

AI assessment note: “making sure that you really narrow your scope for your first product”

Q Devastated. But people look to the likes of your GoPros and your Fitbits. But as we mentioned there, there are many more skeletons with, you said, Pearl, Juicero. Recently, obviously, both Struggling, um, to say the least. Um, so first, talk to me. Are the two above that we mentioned, GoPro and Fitbit, purely outliers, and is consumers still really viable today, then, for startups, considering the huge capital requirements?

A I definitely think consumer electronics as a product category is still viable, and you're going to see some really interesting stuff going on, and certainly you've had things like VR, and they've had acquisitions that are starting to, you know, that's growing a little bit slower than I think everyone would like, but also, let's be super clear, you know, DGI is just a consumer electronics company. It's a ten billion dollar company. And they're showing no signs of stopping. So I think that the, for Fitbit and GoPro in particular, people kind of like fell out of love with them. But at the end of the day, there's still public companies that you can buy into or sell out of. And there's 200 private unicorns today that wish they had the problem of being a public company that fell out of love.

AI assessment note: “I definitely think consumer electronics as a product category is still viable”

Q Absolutely. Where would you say the ecosystem is now, both in terms of hype cycle around investing in hardware, and in terms of, kind of, the metric requirements? Although everyone wants a great business, obviously, and they're, kind of, uh, saleable metrics in terms of, you know, production line, traction-wise, data produced.

A You know, we're really focused right now on, when we work with our companies, to make sure they understand how to get to revenue as You know, I think that two years ago, the ecosystem was in a place where you could go raise a big A or small B on like kind of hype and, you know, good team and like that, the idea that there was going to be a market there. But right now, you know, we're seeing a very much a tightening in the market of when you get to series B, you better have revenue or like be about to sign those contracts or this money will undisputedly unlock that. And so that's a little bit of a change, but I think that that's true. And just adventure comes in waves and cycles across the entire ecosystem. And you think that, you know, you're seeing some stuff right now where Some companies are struggling to get to that next metric, and there's been a couple of companies that have shut down recently, and so for us, we're very focused on, like, making sure our companies can get to those revenue targets that we think they need to hit.

AI assessment note: “when you get to series B, you better have revenue or like be about to”

Q Speaking of kind of getting to the revenue targets as quickly as possible, it makes me instantly think about Reid Hoffman's quote about, if you're not embarrassed about your first product shipment, then it's too late. What do you think about that with regards to hardware, where it's not such an upgradable, changeable asset that can be redefined with time?

A I think it's about making sure that you really narrow your scope for your first product, because if 90% of your product works in its hardware, you have nothing. And so, unfortunately, a lot of times you'll have founders that, you know, by definition are very ambitious, and they'll include too much in that first rep. And then when there's inevitably some problems or integration issues, they can't just then cancel that second wheel or that whatever it is that's giving them a problem, because that's all part of the design. And so, there is this little bit of complexity Systems engineering, and so you have to have a good vision from day one, and so I think that you will have these, if you look at, you know, like any piece of hardware, the first one is generally kind of crappy, and then over time it gets better, but that one, that first one that's kind of crappy has to do something that gets high value for their users.

AI assessment note: “the first one is generally kind of crappy, and then over time it gets better”

Q remaining there, it is also obviously the hub of all things VC, which is why I spend most of my life there. Um, and why most of the time I'm generally a very happy young chap. But, um, Chris Duvos, one of my favorites, asks, how do you analyze the financing challenges for hardware companies, with frequent A and B round investors often finding it hard to really back them?

A So, what we found is That are not a good fit for venture capital. And this is true on software. This is true on hardware. And I think that for a lot of entrepreneurs, they don't always fully understand like kind of what the milestones you have to get to are. So whenever we make an investment, one thing we really focus on is what is the right amount of money for you to raise to get to your next round? And since we've, you know, we've got 42 investments, we've got to see a lot of rounds go down. We have a really good understanding for kind of where the ecosystem is and where those series A and series B investors are. And we'll go talk and we'll say, what would you need to see to make this investment? Now, The reality is, is that they just want to invest in great businesses, and if you can convince them that this is a scalable, venture-backed business, they'll look at anything.

AI assessment note: “we'll go talk and we'll say, what would you need to see”

Q Absolutely. Where would you say the ecosystem is now, both in terms of hype cycle around investing in hardware, and in terms of, kind of, the metric requirements? Although everyone wants a great business, obviously, and they're, kind of, uh, saleable metrics in terms of, you know, production line, traction-wise, data produced.

A You know, we're really focused right now on, when we work with our companies, to make sure they understand how to get to revenue as You know, I think that two years ago, the ecosystem was in a place where you could go raise a big A or small B on like kind of hype and, you know, good team and like that, the idea that there was going to be a market there. But right now, you know, we're seeing a very much a tightening in the market of when you get to series B, you better have revenue or like be about to sign those contracts or this money will undisputedly unlock that. And so that's a little bit of a change, but I think that that's true. And just adventure comes in waves and cycles across the entire ecosystem. And you think that, you know, you're seeing some stuff right now where Some companies are struggling to get to that next metric, and there's been a couple of companies that have shut down recently, and so for us, we're very focused on, like, making sure our companies can get to those revenue targets that we think they need to hit.

AI assessment note: “when you get to series B, you better have revenue”

Q remaining there, it is also obviously the hub of all things VC, which is why I spend most of my life there. Um, and why most of the time I'm generally a very happy young chap. But, um, Chris Duvos, one of my favorites, asks, how do you analyze the financing challenges for hardware companies, with frequent A and B round investors often finding it hard to really back them?

A So, what we found is That are not a good fit for venture capital. And this is true on software. This is true on hardware. And I think that for a lot of entrepreneurs, they don't always fully understand like kind of what the milestones you have to get to are. So whenever we make an investment, one thing we really focus on is what is the right amount of money for you to raise to get to your next round? And since we've, you know, we've got 42 investments, we've got to see a lot of rounds go down. We have a really good understanding for kind of where the ecosystem is and where those series A and series B investors are. And we'll go talk and we'll say, what would you need to see to make this investment? Now, The reality is, is that they just want to invest in great businesses, and if you can convince them that this is a scalable, venture-backed business, they'll look at anything.

AI assessment note: “This is true on software. This is true on hardware.”

Q Can I ask, in terms of kind of the value as a hub for hardware, are there technical and technological advantages for robotics and hardware companies of being in Silicon Valley?

A Absolutely. You know, I think that there's a little bit of an analogy to LA. It's like, I grew up in LA, which is one of the centers of aerospace in the world. But whenever you tell people you're from LA, the first question is, do you know any celebrities? And it's because, you know, for Hollywood to be the global phenomena that it is, everyone has to know it exists. When you get on a plane or use A satellite with communications. You don't need to know where that satellite came from, and I think that hardware in the valley is very similar. For Literally every person on the planet has to know that they're there. Whereas the reality is Apple's largest company in the world does more revenue than either of them. And like, for some reason, there's just never that association of like, oh, Silicon Valley, that's where Apple is. For some reason, the default is Google and Facebook. And there's dozens of incredible hardware companies here, but they just don't have kind of the mindshare that I think some of the social media companies do.

AI assessment note: “they just don't have kind of the mindshare that I think some of the social media companies do”

Q Can I ask, in terms of kind of the value as a hub for hardware, are there technical and technological advantages for robotics and hardware companies of being in Silicon Valley?

A Absolutely. You know, I think that there's a little bit of an analogy to LA. It's like, I grew up in LA, which is one of the centers of aerospace in the world. But whenever you tell people you're from LA, the first question is, do you know any celebrities? And it's because, you know, for Hollywood to be the global phenomena that it is, everyone has to know it exists. When you get on a plane or use A satellite with communications. You don't need to know where that satellite came from, and I think that hardware in the valley is very similar. For Literally every person on the planet has to know that they're there. Whereas the reality is Apple's largest company in the world does more revenue than either of them. And like, for some reason, there's just never that association of like, oh, Silicon Valley, that's where Apple is. For some reason, the default is Google and Facebook. And there's dozens of incredible hardware companies here, but they just don't have kind of the mindshare that I think some of the social media companies do.

AI assessment note: “they just don't have kind of the mindshare that I think some of the social media”