The Exchanges

Q Yeah, like, your take on it, like, is it a feasible avenue for investment when it's not such a scalable path?

A Yeah, I mean, it's, it's hard to, it's hard for me to see, to see that. Basically, When it's also hard for me to, I guess the reason I was asking that question, it's hard for me to speak for the people who are backing these companies, right? So maybe they see something that, that we're not looking for, but the, the, the fundamental focus of most VCs is on, especially in the tech space is, is differentiated technology. And when I think of a company like the Honest company, first of all, I love Jessica Alba. I love, I love the premise and the mission of that company as well, bringing clean, green, Basic necessities to consumers. So this is nothing about the value proposition not working. But when I think about differentiation in the honest company, it seems largely around the brand, right? And so I do think there's a role for VCs who specialize in backing branded companies because brand actually is very sustainably differentiating when it's built correctly and exists. You know, some of the enormously large businesses where the brand itself holds value. I mean, I think just as a quick Recent example, FAO Schwartz, and I think recently the sharper image, like the brand alone's have been acquired for a hundred million dollars plus, right? And so those hold real value and real differentiation. And so if you're the type of VC who can understand that potential arbitrage and opportunity…

AI assessment note: “I do think there's a role for VCs who specialize in backing branded companies”

Q to go from, from one, uh, currently unsexy topic of On Demand to the incredibly, uh, sexy, almost VC orgasmic topic of AI, of And machine learning, which, I mean, is in every deck I've seen over the past month, which is incredible how quickly everyone adopts it. So talk to me, Alex. Is this a bubble? Is this the new reality future? Talk to me. Where are we at?

A Sure. Yeah, I think we're at, uh, we're starting to get pretty warm here in terms of, um, the, the pendulum swingings that I was referencing earlier. It affects AI, machine learning, deep learning as well. We, we, I believe there's a term AI winter, for example, to describe the, uh, The times in history where people have been really negative on AI and where funding has dried up, and that happened in the seventies. I think it happened again in the eighties or early nineties. We're now in this, this golden era of deep learning and machine learning. There's, there are some real issues here though, and I'll get to like the good news, but the bad news is that I believe we're, we're in a point in time where, and this is, you know, I think I mentioned that this is like blasphemy for me to say as a VC, given, you know, AI is sort of hot right now, but We're in an era where what is being promised by a lot of entrepreneurs is just fundamentally impossible from the, the actual fundamentals of, of the current state of the art. And so I think people need to be aware of that and calibrate expectations appropriately. It's not to say there's not 10 X opportunities out there. Plus that are enabled by advancements. Certainly like things are different. Now the infrastructure is different. You have better processing power, GPUs, et cetera. You have much more data and ability to get data and so for…

AI assessment note: “we're starting to get pretty warm here in terms of”

Q to go from, from one, uh, currently unsexy topic of On Demand to the incredibly, uh, sexy, almost VC orgasmic topic of AI, of And machine learning, which, I mean, is in every deck I've seen over the past month, which is incredible how quickly everyone adopts it. So talk to me, Alex. Is this a bubble? Is this the new reality future? Talk to me. Where are we at?

A Sure. Yeah, I think we're at, uh, we're starting to get pretty warm here in terms of, um, the, the pendulum swingings that I was referencing earlier. It affects AI, machine learning, deep learning as well. We, we, I believe there's a term AI winter, for example, to describe the, uh, The times in history where people have been really negative on AI and where funding has dried up, and that happened in the seventies. I think it happened again in the eighties or early nineties. We're now in this, this golden era of deep learning and machine learning. There's, there are some real issues here though, and I'll get to like the good news, but the bad news is that I believe we're, we're in a point in time where, and this is, you know, I think I mentioned that this is like blasphemy for me to say as a VC, given, you know, AI is sort of hot right now, but We're in an era where what is being promised by a lot of entrepreneurs is just fundamentally impossible from the, the actual fundamentals of, of the current state of the art. And so I think people need to be aware of that and calibrate expectations appropriately. It's not to say there's not 10 X opportunities out there. Plus that are enabled by advancements. Certainly like things are different. Now the infrastructure is different. You have better processing power, GPUs, et cetera. You have much more data and ability to get data and so for…

AI assessment note: “what is being promised by a lot of entrepreneurs is just fundamentally impossible”

Q Yeah, like, your take on it, like, is it a feasible avenue for investment when it's not such a scalable path?

A Yeah, I mean, it's, it's hard to, it's hard for me to see, to see that. Basically, When it's also hard for me to, I guess the reason I was asking that question, it's hard for me to speak for the people who are backing these companies, right? So maybe they see something that, that we're not looking for, but the, the, the fundamental focus of most VCs is on, especially in the tech space is, is differentiated technology. And when I think of a company like the Honest company, first of all, I love Jessica Alba. I love, I love the premise and the mission of that company as well, bringing clean, green, Basic necessities to consumers. So this is nothing about the value proposition not working. But when I think about differentiation in the honest company, it seems largely around the brand, right? And so I do think there's a role for VCs who specialize in backing branded companies because brand actually is very sustainably differentiating when it's built correctly and exists. You know, some of the enormously large businesses where the brand itself holds value. I mean, I think just as a quick Recent example, FAO Schwartz, and I think recently the sharper image, like the brand alone's have been acquired for a hundred million dollars plus, right? And so those hold real value and real differentiation. And so if you're the type of VC who can understand that potential arbitrage and opportunity…

AI assessment note: “if you're the type of VC who can understand that potential arbitrage and opportunity, great”

Q to do. So as every consultant hates me, uh, but moving, moving from the topic of ML, but kind of staying on the deep tech layer, we've also seen a number of hard science startups receive pretty significant backing from software VCs who are expecting software outcomes. So what's your take on this? And do you think fundamentally those outcomes that are expected are possible in this pretty unexpected space?

A Sure. This comes from the point of view of having founded a material science-based company, spent out of my engineering thesis work in nanotechnology and polymers, and having lived through that, and then on the, on the invest, that was as an entrepreneur, and then on the investing side, having backed about a couple dozen either hardware or fundamental sciences-based companies have been able to see this play out again and again. We hear as a society, as entrepreneurs, that, you know, the cost of building software has come down a couple orders of magnitude. Also like the cost of starting a hardware business or a life sciences or fundamental science, deep tech based businesses is, is disrupting as well. The problem is the analogy isn't like one-to-one like, yes, it's become, it's become significantly cheaper and faster to start a hard sciences company, but everyone, founders, investors, everyone needs to recognize that we're not yet at the speed of software in the hard sciences. And so nor are we at the cost. And, and so I, The danger here is we have a lot of new money and coming in, seeing, hearing the meme that it gets become cheaper and faster to build hardware businesses and hard science businesses without recognizing the nuance of things. And when I say the nuance of things, like, there are businesses that do exhibit software company-like efficiencies, capital efficiencies, a…

AI assessment note: “we're not yet at the speed of software in the hard sciences”

Q to do. So as every consultant hates me, uh, but moving, moving from the topic of ML, but kind of staying on the deep tech layer, we've also seen a number of hard science startups receive pretty significant backing from software VCs who are expecting software outcomes. So what's your take on this? And do you think fundamentally those outcomes that are expected are possible in this pretty unexpected space?

A Sure. This comes from the point of view of having founded a material science-based company, spent out of my engineering thesis work in nanotechnology and polymers, and having lived through that, and then on the, on the invest, that was as an entrepreneur, and then on the investing side, having backed about a couple dozen either hardware or fundamental sciences-based companies have been able to see this play out again and again. We hear as a society, as entrepreneurs, that, you know, the cost of building software has come down a couple orders of magnitude. Also like the cost of starting a hardware business or a life sciences or fundamental science, deep tech based businesses is, is disrupting as well. The problem is the analogy isn't like one-to-one like, yes, it's become, it's become significantly cheaper and faster to start a hard sciences company, but everyone, founders, investors, everyone needs to recognize that we're not yet at the speed of software in the hard sciences. And so nor are we at the cost. And, and so I, The danger here is we have a lot of new money and coming in, seeing, hearing the meme that it gets become cheaper and faster to build hardware businesses and hard science businesses without recognizing the nuance of things. And when I say the nuance of things, like, there are businesses that do exhibit software company-like efficiencies, capital efficiencies, a…

AI assessment note: “we're not yet at the speed of software in the hard sciences”

Q Yeah, absolutely. I am intrigued on one other aspect, and it's how do you overcome investment biases when you do have Cases where it doesn't work out in a particular space, say the on-demand space and investment doesn't go as planned. How do you kind of have clarity mentally when making another one? How do you get over that biases?

A It's incredibly hard. So just admitting that it's difficult, right? And, and, uh, and, and likely I and all people are, are subject to biases even when we're aware of them. But that is the start. So the start is being aware that all humans have these biases and just Google them. Cognitive biases. There's amazing articles out there about them, and it's worth reading up and understanding, being self-aware about the fact that we are ourselves and to some extent wired like an algorithm, and there are certain things that can train our algorithm to react in a certain way, and just being aware of that and realizing that, hey, we're being exposed to very limited data and sort of going in one, one direction just because this one thing happened may not be the appropriate response. It could be, right? So it's worth It's worth being skeptical and, and, uh, not just believing everything you hear and see, but being open-minded is very important for a VC. Like, believing the impossible could be possible. I guess, you know, we call it suspending disbelief, right? That's really important to do.

AI assessment note: “the start is being aware that all humans have these biases”

Q Yeah, absolutely. I am intrigued on one other aspect, and it's how do you overcome investment biases when you do have Cases where it doesn't work out in a particular space, say the on-demand space and investment doesn't go as planned. How do you kind of have clarity mentally when making another one? How do you get over that biases?

A It's incredibly hard. So just admitting that it's difficult, right? And, and, uh, and, and likely I and all people are, are subject to biases even when we're aware of them. But that is the start. So the start is being aware that all humans have these biases and just Google them. Cognitive biases. There's amazing articles out there about them, and it's worth reading up and understanding, being self-aware about the fact that we are ourselves and to some extent wired like an algorithm, and there are certain things that can train our algorithm to react in a certain way, and just being aware of that and realizing that, hey, we're being exposed to very limited data and sort of going in one, one direction just because this one thing happened may not be the appropriate response. It could be, right? So it's worth It's worth being skeptical and, and, uh, not just believing everything you hear and see, but being open-minded is very important for a VC. Like, believing the impossible could be possible. I guess, you know, we call it suspending disbelief, right? That's really important to do.

AI assessment note: “the start is being aware that all humans have these biases”