233. Investment Decision in < 3 weeks; the Challenges and Opportunities with 3D Printing; and the Future for Industrial Automation (Kane Hsieh)

233. Investment Decision in < 3 weeks; the Challenges and Opportunities with 3D Printing; and the Future for Industrial Automation (Kane Hsieh)
Download_v2
Nick Moran Angel List

Kane Hsieh of Root Ventures joins Nick on a special Crisis Coverage installment to discuss Investment Decision in < 3 weeks; the Challenges and Opportunities with 3D Printing; and the Future for Industrial Automation. In this episode, we cover:

  1. Backstory/Path to Venture
  2. Tell us about the the thesis at Root and your focus areas.
  3. Any unique outcomes or behaviors you are observing as a result of the pandemic?
  4. What’s your investment position on 3D printing… bullish or bearish?
  5. I used to work in the industrial automation area… developed two different products there and saw limitless opportunities but also many challenges.  From your standpoint, what have been the biggest challenges to progress in industrial automation?
  6. What excites you most about industrial automation going forward?
  7. How will jobs in the industrial space, or any space for that matter, change with the immersion of automation and robotics? Is this a net job eliminator?
  8. Startups w/ hardware as a part of their solution (whether off the shelf or developed internally) are often are more capital intensive than their pure software counterparts.  How do you balance progress and capital efficiency for  these businesses?
  9. Do you think the expectations of progress/traction, at different stages, is different for hardware as a service vs. software as a service?
  10. One piece of advice for founders of industrial automation startups… what would it be?
  11. What are some of the deep tech or capital intensive investment areas that you’re avoiding?
  12. Heard a rumor that you’re starting podcast?
  13. I came across a side project, “Transformer Poetry,โ€ a book of famous poetry reimagined by OpenAIโ€™s GPT-2 language model. I don’t even know what to ask here but would love to hear what the inspiration for it was.

Guest Links:

Transcribed with AI

Intro 0:03
welcome to the podcast about investing in startups, where existing investors can learn how to get the best deal possible. And those that have never before invested in startups can learn the keys to success from the venture experts. Your host is Nick Moran and this is the fall ratchet.

Nick Moran 0:23
Kane Hsieh joins us today from San Francisco cane is a partner at route ventures. Rho Ventures is a Silicon Valley based venture firm with investments in Esper particle wild type Nautilus labs, among others. Kane is also founder of brilliant bicycle company, a direct to consumer brand based in Santa Monica. Prior to Reed ventures, Kane was an early stage investor at our E ventures where he sourced investments such as airwave Florida, and crypt Gnostic. He has a background in software engineering and project management at both remote IV and Microsoft cane. Welcome to the program.

Kane Hsieh 0:54
Thanks for having me.

Nick Moran 0:55
Yeah. So tell us more about your background. I kind of went through quick there.

Kane Hsieh 0:59
Yeah. So you know, you’ve covered my professional background pretty comprehensively. But But I think the part that’s important here that informs a lot about why I do what I do, and why I invest in what I invest in is I grew up near Niagara falls outside of Buffalo, New York. And it’s this, frankly, it’s kind of sad, post industrial, burned out factories in the mid Atlantic. And you know, as a kid, we would literally wander around like the old Nabisco warehouse, which was abandoned or like Union Carbide facilities, etc. And when I started a career in software engineering, the gap in what different industries considered state of the art was huge. And in pure software, there’s a lot of really smart people competing, frankly, like, I thought this would be a comparative advantage to look at large industries, which are still lagging in a lot of technologies that reinstall for take for granted. And that’s why I do what I do now, route ventures,

Nick Moran 1:53
what did you study in school?

Kane Hsieh 1:56
I studied computer science and design. And frankly, I wasn’t the best CS student.

Nick Moran 2:04
Now there was, well, yeah, but

Kane Hsieh 2:07
what that led me to do is like I had to, I had to, like, desperately look for hacks in order to pass. And so like, I started helping out in the mechanical engineering, or the, or the electrical engineering labs, and I found a lot of opportunities to bring basic optimization and computational algorithms into different areas of engineering and really make a difference. And so that’s, you know, even an undergrad, I was seeing this pattern of taking software and CS Principles and applying them to other industries.

Nick Moran 2:41
Have you had work experience in, you know, Applied Engineering capacities to industrials or heavy industries? Or is it, you know, been more sort of big tech software side.

Kane Hsieh 2:55
So at remote IV, there was a lot of hardware engineering that was going on, not necessarily on the product side, but internally and working with our manufacturers getting with different contractors, like writing basic scripts, or API’s or, or little programs to collate and clean up and munge data together, became pretty standard for us now, at the bicycle company, we saw the same thing as well. These are industries that were working in Excel and Word and emails and files, title like certification, underscore final V to underscore signed in these roles. I was doing a lot of what could be considered software engineering, but really just to like, keep the operations going as efficiently as possible. Yes.

Nick Moran 3:46
Okay. And have you guys done any investments since the pandemic broke?

Kane Hsieh 3:52
Yeah, you know, we were pretty fortunate in our timing, and that we were about 10% deployed in our fund when the pandemic broke, we didn’t see any reason to aggressively slow our investing. In fact, like about two weeks after pandemic procreate for term sheets out, and a few of those are closing right now.

Nick Moran 4:16
Awesome, awesome. Was it tougher to get to decision?

Kane Hsieh 4:19
You know, I think our team is pretty small. There’s four of us, and we operate pretty independently. So besides the awkwardness of transitioning everything to zoom, and the eye fatigue of staring at a screen all day, there hasn’t been that much difficulty in the closing process. I would say the biggest source has been can we get comfortable with investing and deploying capital into folks that we’ve we’ve never met in person. And I mean, frankly, I think it’ll be centaurs. For all I know, I haven’t actually seen them below the chest. So that’s a little strange. But yeah, it’s it’s it’s been going and the origin of this was that when we grew from two to four people, our process suddenly took twice as long, which we realized was a bad thing. That’s not what’s supposed to happen. So we actually had to re architect the process to take between half to the same amount of time. And that was about a year and a half ago.

Nick Moran 5:19
How much time is it taking you? Like three weeks? Three weeks from initial meeting to decision?

Kane Hsieh 5:28
Yeah, yeah, we try to keep it to two, but it slipped to three, because of again, I think the awkwardness of meeting on Zoom.

Nick Moran 5:35
That’s amazing. How do you guys handle it when, you know, you engage with a founder, and it is the right fit, but maybe the timing is off, they’re not quite ready to, to, you know, raise the next round, or maybe they’re not quite ready for the level that you want to come in at? How do you deal with those,

Kane Hsieh 5:55
I would say, as a seed fund, it’s rare that people aren’t quite there yet. For us, it’s more often the case that they’re actually beyond where we would invest, in which case, if we like when we try to introduce them to later stage funds, but the reality is, like, at an early enough stage, any marginal capital is probably beneficial. And the company and founders a lot more flexible. And we try to be to, you know, those are checks anywhere between like 500,000 to a million and a half, which, you know, the nominal change isn’t that much. It’s $500,000. But if you think about the stage, we’re at that that’s anywhere between like 100 to 33% of the actual check we’re writing. And so it’s, it gives us some flexibility.

Nick Moran 6:42
What if anything has changed, about the way you guys invest? You know, since COVID, occurred and everyone’s sheltering in place,

Kane Hsieh 6:51
we’ve had to more intentionally try to spend time with potential founders, and then founders after the fact, whether it’s phone calls, casual texting, zoom, etc. Because we we’ve lost this kind of the casual or serendipitous interactions, adjacent to like the formal pitches and meetings, were actually frankly, I think a lot of metadata comes out of and so coming up with ways to interact with, with those founders, outside of the formal meetings has been the toughest part. And that’s what we’ve been working on the most.

Nick Moran 7:29
And then from a market standpoint, are there certain markets that you’re considering more considering less, you know, with these circumstances,

Kane Hsieh 7:37
and again, I consider myself lucky here, our firms fit well, and specifically, my emphasis on a lot of industrial automation, or robotic technologies, very spin between neutral to increases in demand, as a result of COVID things related to labor shortages, inability to, to get certain numbers of people on site, or just general uncertainty about supply chains. Like I think we’re seeing potential customers start to emphasize like objects efficiency more as a result. Yeah,

Nick Moran 8:14
it’s interesting, because I’ve noticed a bit of a double edged sword with the portfolio with our portfolio, we do have a lot of robotics and hard tech companies, for instance. And on one hand, I’m hearing from the founders, we can double down on product and spend a lot more time on that, which is great. On the other hands, when they only have a couple of prototypes. And especially in an office location, it’s like they can all collaborate to work together on the equipment. Are you noticing similar situations? Yeah,

Kane Hsieh 8:47
our portfolio companies that are in the process of developing hardware have to come up with ways to set shifts, and like cleanliness protocols, so engineers can come in on shifts and work on the system that has definitely slowed down productivity for the portfolio companies a bit. But there is precedents around that, especially like folks coming from companies like Apple, which generally had a lot of their engineers flying back and forth from overseas, like they, a lot of them just kind of fell back into their natural like how to figure out shifts and working on weird time zones, and working remotely. So it’s not ideal, but it is, it’s not intractable leaving

Nick Moran 9:33
about triage and dealing with the portfolio. Tell us kind of how you guys have approached that.

Kane Hsieh 9:39
So the first thing we did was go through the portfolio and get a sense for everyone’s burnout date. And then we did our best to get everyone’s burnout pushed to 18 months to 24 months, if possible. Now for the companies that had just raised in the queue. For q1, they were in a good spot. The next option was companies that had venture debt facilities, you know, we encourage them to draw down more debt than they had previously planned. And just to pause hiring until we had a better sense for how the pandemic would affect their business. And unfortunately, there was a small minority of companies, which was caught in some unfortunate timing, and ultimately had to do some furloughs, with headcount. So between a few of these levers, we had to pull we just got over one to 18 plus months of runway. That’s pretty good. And

Nick Moran 10:40
how big is the portfolio?

Kane Hsieh 10:41
The you know, I shouldn’t have exact number off the top of my head, but it’s in the high 20s. Good. And

Nick Moran 10:47
remind us, you know, what’s the thesis at route? And what’s your focus?

Kane Hsieh 10:51
The thesis route, in general is we are trying to be the first institutional capital into highly technical founders. My focus within that is in industrial automation, robotics, but I have colleagues like me, Edwards, whose focus is primarily software developer tools. And then colleagues, like Chrissy who spend a lot of time looking at hardware, logistics and manufacturing. It’s pretty broad by design. But one thing that’s in common across all the founders is they come from strong technical backgrounds, and the products they’re making include a lot of technical powers.

Nick Moran 11:28
Yeah. How are you thinking about recent events in Minneapolis and Black Lives Matter? And you know, what you can do at root to help address existing issues and racism within tech,

Kane Hsieh 11:43
the elephant in the room that’s impossible to avoid right now in any conversation is the protests, police brutality and minority representation. You know, I think it’s worth saying on the record, that as I look at our portfolio, and our partnership, we are diverse in some ways, but not in all the ways and certainly not in some very important ways, which are relevant now. And that’s on us. Like, it’s easy to blame the pipeline, or blame the system or blame biases, whoever, but fundamentally, at the end of the day, it’s our team, we make the decisions, and those decisions haven’t reflected well on diversity. And that’s something that is top of mind for us. That’s something that we don’t really have anything to hide behind on. And that’s something that we hope to change, you know, actively. Good.

Nick Moran 12:33
Interesting. Interesting. Yeah, I want to touch on a few different areas today, because, you know, you’ve got some great background and personal experience, and then, you know, investment experience, of course, in some of the, the industrial automation spaces, and 3d printing spaces and robotics and my conversations with your colleague, Avi Dawn, he’s, he’s always mentioning your name. And, you know, I’m glad we have a chance to connect on this. But, you know, maybe we’ll start out with just touching on 3d printing. You know, we, we could probably spend a whole episode on this, but super hot in the 2000s and early 2010s, and lots of ups and downs in that space. You know, I spent a lot of time early in my career in CNC and water jet and laser, in a lot of ways, even though most of those technologies are about cutting and ablation as opposed to additive. In a lot of ways, I see that as, you know, a precursor to 3d printing. Can you give us an overview of kind of your stance on 3d printing now? And, you know, are you bullish or bearish? So,

Kane Hsieh 13:36
I’m very bullish on 3d printing, and you bring up a good point, which is like if I if I think back to the early 2010s, a lot of naive investors lost a lot of money, making 3d printing bets. There was a lot of kind of breathless HYPEE, press around this, like Star Trek, Silver Bullet kind of vision of 3d printing, where it would magically solve all problems, and everyone would have a 3d printer on their desktop. And it was naive, like, you bring up CNC like CNC is an amazing process. But we still cast stuff, and we still forge stuff, and we still mold stuff, right? Like, CNC does a lot of things really well. And it’s unlocked a ton of capabilities for us. 3d printing is the same, it’s a new class of technology. But if you’re going to treat it as something that we expect to displace all those other capabilities, like you’re gonna lose money, because because it’s just that’s just not how, you know technologies, or in this case, manufacturing capabilities get adopted. If I look for founders and companies treating 3d printing, as new classic capabilities with its own intrinsic values and own intrinsic disadvantages. Those are the places where I think we’re going to start carving out real big value, because, like, that’s what it takes to to build a massive company here. It’s like understand where you’re going wanting to be extremely competitive and groundbreaking and understanding, we’re like, look, frankly, if you’re making spoons, like, molding is super efficient, we’re really good at it, there was no like competitive advantage and molding, you can kick that off to like 1000s of sciennes. Yes, your contract manufacturers and so that is, if anything, the signal to noise ratio has gotten better, because I think a lot of people that joined the hype cycle are now out of the hype cycle, and the folks I see working on it now are sharper and more driven than ever.

Nick Moran 15:33
Well, and we have seen some, some really nice success stories there. You know, without naming specific company names, you know, there’s, it’s also sort of been a key foundational technology, you know, as an enabling technology for rapid prototyping and custom components, you know, for like the the big companies like SpaceX and such, how do you frame up the categories of 3d printing and kind of where you guys are focusing from an investment standpoint,

Kane Hsieh 16:01
there have been some early successes in hardware, but the way I look at it is that the printing of hardware is actually the least interesting part to me, that’s not to say there won’t be other winners per se. But if I think about what it takes to 3d print something, you have to design it, you have to digitally design it, or part rather, you have to digitally validate that part. Once that parts validated you, you have to make that part producible on a certain machine. And every step there represents a piece of software, which, frankly, you know, I’m sure you’ve worked in CAD or similar, something like these are pieces of software that have been around for decades, there’s maybe a half dozen public companies doing 10s of billions of dollars of revenue a year that have been making the software the same way since since my grandfather was literally drafting with a pencil on a table. Yep. And so we’re actually where I see a lot of value waiting to be unlocked is in the software stack leading up to production of these of these newer advanced parts of

Nick Moran 17:08
it, I don’t want to go too deep down the 3d printing rabbit hole, because we could do that when I hit on some other areas today, you know, I used to work in industrial automation, I was a product manager founded a couple products in that space specifically. You know, from your standpoint, I saw a lot of archaic industries and a lot of really slow moving tech in this space, a lot of slow moving customers and innovation. What do you think have been some of the biggest challenges to the progress in this industrial automation space?

Kane Hsieh 17:39
So this is something that I always tell entrepreneurs that I speak to, regarding industrial automation, which is there are processes where if you, if you look at it as an outsider, especially as an engineer, you’re like, Well, this is clearly an inefficient process, I know what the deliverable should be like, therefore, there’s an obvious solution here. And the challenge there, which I’m sure you’ve seen, is that in industrial processes, there’s often parties extracting value from systemic inefficiencies. And they’re often hidden when you’re observing as an outsider. And that makes it hard. Unless you know exactly where those perverse incentives exist, you can find yourself running into a brick wall at the last second, because you’ll just be like black bag or stonewalled by some party or contractor or individual business unit that’s like, hey, that’s our golden goose, like, don’t touch it. And we see this in construction. We see this in oil and energy extractive industries, manufacturing, which is there is this cobbled together web. It’s not even a value chain. It’s like a valley web. And you have to very clearly know where you can access the inefficiencies and the clean way, and have the support of people willing to pay and influence that decision on your behalf. Because the strictly speaking, efficient engineering solution almost runs into someone who’s ready to stop you in order to preserve their revenue.

Nick Moran 19:16
While there’s there’s an inertia, right, that’s a piece of it. And then, of course, when you have like, a heavy industry or a physical system with a lot of dependencies, you know, multi system, network failures in that system can be not only really hard to diagnose, and correct, but really expensive. Not that failures aren’t expensive and software, but I just feel like the way that problems are addressed and iterated on and fixed are just faster and much more, you know, versatile in the software space.

Kane Hsieh 19:49
Yep. So you, you bring up a really good point. And I think there’s like a pretty simple matrix that we think about that captures I believe the point We’ll try to make, which is if I think about an action that a heavy industrial system makes, on one axis, there’s cost of failure per individual action. And then on the other axis, there’s frequency of that action. Right? Ideally, you want something that happens a lot, that has low individual cost of failure, right? And something to think about is soft fruit harvesting, right? picking strawberries, if you can increase the throughput, 10%, like an extra point, 5% failure rate, sure, go for it. Because the strawberry gets dropped, who cares? The exact opposite corner is like I’m docking a spacecraft to a space station. I’m not doing this that often. And I definitely cannot miss. Right? Certainly, like, you know, it’s a gradient, but to the extent that you can be in that first quadrant, which is I’m doing this thing a lot, which means I have multiple chances to like drive efficiency. And also, if I fell, it’s not that big of a deal. That’s great. So yeah, I believe that’s kind of what you’re trying to get at, ya

Nick Moran 21:11
know, 100%. And like, not a perfect science here. But I might even add a Z axis, which is like reputational risk and skill risk of the person that’s responsible for the system, right? Because in a lot of these Old World Industries, you know, people are worried about the jobs being replaced, first of all, their skill sets might be oriented to, you know, older technology that they’ve been comfortable with and using for decades. And even if they are comfortable with change, and they’re willing to implement new tech, there’s also this huge reputational risk, if something doesn’t work in their system failures, you know, their neck is on the line, whereas you take an industry, you know, like tech, where software issues, nobody likes a software issue that cost money, right. But people are much more comfortable with rapid change, experimentation, new innovation, and the people responsible are typically those that embrace, not in all cases, but at least in a lot of tech, especially out where you’re based, you know, people, the types of people are the types that embrace new technology and look for new advancement on a regular basis.

Kane Hsieh 22:27
Yeah, yeah, exactly. And it’s understandable in some industries, that there’s a fundamentally more conservative approach to technologies. Actually, something that I see pretty often getting pitched is the idea of like a fully automated, machine tended machine shop, you know, the student sees that you’re alluded to earlier, and I would love for this to exist. But at the end of the day, like, there’s huge intrinsic challenges here, where, well, let me give an example. Speaking to a machinist friend, he told me something called a tea told me about something they call the Million Dollar Club, which is, have you made an error, which causes a million dollars of damage? Wow. And that’s not either gonna be exactly. And that’s like, if you’re using, you know, CNC machines that cost hundreds of 1000s or millions of dollars, and you’re working on parts that are very high value, and you fuck something up. And you cause that tool to crash. Not only have you wrecked that machine, but you’ve taken the line down for how many days or weeks when they got a service that machine and

Nick Moran 23:33
OSHA and safety issues too, right? Yeah.

Kane Hsieh 23:37
So again, this goes back to like, the areas that I’m interested in and that I tried to push entrepreneurs towards, which is like, can you carve out atomic processes, which are, by themselves lower risk, and can be more accepting of newer technologies? And that’s a challenge.

Nick Moran 23:54
Yeah, talk more about that. You know, I feel like we’ve been kind of bearish on the industrial automation case so far. But you know, where are the opportunities and why? Why do you get excited about it?

Kane Hsieh 24:04
Well, the opportunities themselves within certain industries, right, like the there will be parts of the industries, which, you know, I want to touch with a 10 foot pole. And then there’s parts in it, where I feel pretty excited by now the industry is at large that I’ve been tracking recently include construction, the extractive industry, so it’s like mining, oil and energy, and certain classes of manufacturing. And again, it’s not specific industries, per se, but within industries, I believe we’ll find opportunities that kind of fit in the the nice corner of that quadrant. Actually, here’s a an example which continues to blow my mind, which is, we’re all familiar with just the billions of dollars being poured into autonomous passenger cars. I just saw pony I raised like $400 million way Moe Cruz, you know the baker’s dozen have others just like the sheer amount of capital invested? And frankly, like, I still don’t really trust an autonomous vehicle in the streets of SF. That said, Rio Tinto, which is, like one of the largest mining companies in the world, has been operating autonomous haul trucks since 2017.

Nick Moran 25:21
Wow, I did not know which, yeah,

Kane Hsieh 25:23
so there, if you haven’t seen a whole truck, you should google it, they’re totally wack. These are like, these are dump trucks where you could put like, 50 normal cars in the background. They’re insane. They’re huge. They run between like processing facilities and minds. You know, someone at Rio Tinto was basically like, look, you’re trying to solve this really hard problem of driving in on a city street, I have 20 miles of straight away, there’s nothing else around it. And literally any human within miles is a trained, technical expert. So I’m going to make this truck autonomous. And great application. Exactly. And this is an example where, in fact, the application that we find in heavy industrials is actually cheaper and quicker to value than the application that we’re working on here in the Valley. And so if you dig around enough, and really understand, again, all the different parties, third parties, individual business units inside some of these industries, you can find opportunities like this, you

Nick Moran 26:28
know, I just saw cruise automation got a license in either California or maybe the broader city of San Francisco to operate autonomous vehicles.

Kane Hsieh 26:37
I’m excited for it look, as a customer as a techno optimist, like, I want that to exist, I’m bullish on it. As an investor, I will gladly take the subsidized research and components and manufacturing and compute that they’ve developed and go stick it on other machines that will make the money faster.

Nick Moran 26:58
What about like, you know, we’ve seen some of these really hot tech names, hot startups, like zoom pizza, and cafe X that have these, you know, cool consumer consumption applications. And they’re taking these, you know, robotics and articulated arms and kind of dressing them up quite a bit. And we’ve seen those, I mean, some of those we’ve seen, you know, fall on really hard times here. What do you think of those applications, you know, using industrial automation and robotics in these sorts of applications.

Kane Hsieh 27:33
So it’s never good. When a company fails, right, like a lot of smart people put in good effort and ended up losing their jobs. And it’s never a happy situation, with zoom with some of the recent hardship to cafe X that said, you know, if I, if I were to try to take a pragmatic approach to understanding why things are going the way they’re going with them is like, frankly, they’re extremely naive approaches to automation. If you watch their demo videos, like you said, they’re using a six axis robot arm to do tasks in the kitchen. And, to me, this is crazy for a few reasons. One is, you know, I said this was the naive approach to automation, the analog I will give you is, imagine if, instead of designing the automobile, we try to design robot horses to drag carriages around. It doesn’t make sense. You’re designing around abstractions, and interfaces, which were in the case of like, you know, commercial kitchen, which were designed for people. If I think about the Zoom video, they always show they use a six axis arm to move a pizza to access into a pizza oven over design. It’s extremely over designed, and it’s over designed because it uses a oven which was designed for a human operator. And now they need a system to take something off of a conveyor system and put it in another device designed for humans. So you’re like forced to design around these weird abstractions like humans are good at some things. They’re not great at other things. There’s a lot of like safety and redundancy and kind of human interaction bits built into that commercial element, which is also why it’s like half the size of a room. If you have a fully automated system, and you’re trying to make it as economical as possible, you shouldn’t be using that many components designed for human operators. The issue with cat bits is slightly different, which is if you will get the machines are serving coffee out of it’s a class of espresso machine called a super automatic. There’s a con you find out like airport lounges where you hit a button and it does the grinding and the milk frothing and just like janitors together and stares at you. Yep. And granted, they put like fancy branded plates over the super automatic machines. And so you don’t see that they’re just like generic machines. But the bet with Cafe x in my mind is that it’s someone willing to pay up Premium to watch a robot arm give you a cup of coffee that you would otherwise get from this machine. If you just had a cup dispenser and a credit card swipe. That feels like a hard bet to me.

Nick Moran 30:10
It’s a fun novelty probably the first time people use it, and then don’t know.

Kane Hsieh 30:17
Yeah, and again, like, I like the vision of Cafe x, which at a high level is people should be able to get high quality food and drink at lower prices, because we can squeeze a lot of efficiency out of these systems. But both of these approaches, by using an over built arm to execute tasks around human safe machines is naive, and it costs a lot of money.

Nick Moran 30:41
Yeah, I couldn’t agree more. We actually just had Jason Calacanis on the program. I know he’s a big proponent of Cafe x, but I’m gonna have to ask him about that next time we chat. I mean, it’s shocking to me. So we’re hardware investors, we do a lot of deep tech. And we’re one of the very few between the coasts that focus between the coasts. And I can’t tell you how many I mean, we see so much right, like, everyone does the deal. Flow is crazy. But the hardware oriented and deep tech oriented deal flow is, you know, a lot of it finds its way to us, and I just see so many where the bomb is, you know, the bill of materials is just extremely long. And the points of failure on the systems are crazy. I mean, I’m a product manager that like built like three different hardware products in my time. And I could not imagine proposing some of these concepts to my management at the time, like they would have laughed me out of the room, because it just the unit economics don’t make sense. And the failure rate is, is going to be crazy. No,

Kane Hsieh 31:45
are you talking about consumer hardware or kind of hardware at large?

Nick Moran 31:50
Usually consumer, it applies to both, but usually in consumer applications, I find people are trying to bring, you know, a rocket launcher to a knife fight, right? It’s, it’s kind of this example, with this eight axis articulated arm that you said that, you know, needs to do a job. That’s two axes, right? You could use like a scarer, robot or even something simpler than that. So you know why the over design? I mean, if there is a UX consumer experience element, like, people are really going to pay a ton of money to watch a dancing robot arm that changes things, right? But if you’re trying to do a job and do it well, and do it efficiently, and trying to cut a lot of costs out of a system, by removing humans, why are you introducing so much cost into a system and introducing all these points of failure? Right? If you have a lot of like, if you have a very complex system with a lot of inter inter dependencies, when that hardware system breaks down in the field somewhere, whether it’s a an enterprise application or a consumer application, that’s really hard to fix, and you’re losing revenue. In the meantime, right? It’s not like software where you can have a remote knowledge worker fix it from anywhere.

Kane Hsieh 33:02
Yeah, on the consumer side, I’ll caveat it by saying, I don’t look at a lot of consumer hardware investments. There are challenges beyond engineering and any consumer product. And frankly, like, that’s just not an area where I have a competitive advantage. So kind of scoping this conversation to the hardware design part. Like, frankly, Apple is just set the bar really high. Yeah, I honestly just believe that’s part of it. Right? Especially to consumer software, and consumer hardware that’s subsidized by software, like iPhones have a unique feature, which I don’t think gets spoken about enough, which is only in these categories, is the best thing, also the mass thing. And no other industry do you have the thing that is the best also be the thing that is used the most. And that’s the economics and the revenue that you can extract from, you know, selling an iPhone is significantly different than what you can extract from selling a Ferrari. It’s why like, Ferrari isn’t the biggest carmaker. And I think, if you’re thinking about consumer hardware, that is often a nuance that’s missed, and people will try to design the Ferrari, once there’s a really clever capital efficient way to continue to extract value from that product. It’s not going to be the mass product.

Nick Moran 34:27
Interesting. Never never thought of it with that lens. So just to kind of pull on that thread a little further, you know, what’s your take on the capital efficiency or lack of capital efficiency and just the capital intensity of, you know, whether it be b2b or b2c? But just, you know, hard tech and deep tech oriented startups, you know, how do you think about financing these companies and then balancing traction and progress, you know, with capital efficiency?

Kane Hsieh 34:56
I’ll just say it’s the best time to be a file under one of these industries, there is a lot of capital, there was a lot of interest from investors. That said, I think the number one thing you can do to manage cash flow if you’re developing hardware is hire good talent. My partner here Chrissy her job at Apple, and then square for like a decade was making sure that things got produced on time on schedule correctly about that involves spending 200 days a year in China that involves like 20 hours Sprint’s in factories, right? Like, you can’t assume the naive thing to do would be to assume that I have an idea for hardware, therefore, I can just pass it off to see him and they’ll make it it’ll be great, right? Like the act of engineering for manufacturability. And then engineering, the actual manufacturing of it are discrete skills. And good hardware companies have armies of engineers in those buckets. And you will never in software, you will never have your UX designer also run DevOps. So why would you assume that you could build something, just because you have a cool, like cool design for hardware? Like there’s entire massive parts of managing that lifecycle that that requires good talent?

Nick Moran 36:23
So circling back to the question, you know, how do you think about financing these companies and balancing their capital efficiency? And yeah,

Kane Hsieh 36:31
tying it back together again, like, there is more invested interest in doing these things. Now, I believe that if you can demonstrate that you have a solid understanding in house in the economics of manufacturing, you know, but by the way, like, this includes not just the engineering of it, but also from negotiating supplier financing, figuring out is there like venture that that can that can be used to there’s a lot of alternative, I don’t believe that money for equity should be converted into inventory, if it doesn’t have to be. And so the way I think about it is like, do I believe these teams have the understanding of all the stones that can be turned over and levers that can be pulled to make this processes as capital efficient as possible?

Nick Moran 37:14
Do you think the milestones for you know raising rounds and are different? And how are they different? Like, if you could give us a high level that, you know, CT Plus series A maybe series be hard to generalize, you know, in such a big category, but can you give us a sense for those?

Kane Hsieh 37:33
Frankly, I don’t know if there’s even a good generalizable answer. Because every, at least you know, talking about specific industrial hardware, there’s, there’s a lot of different things about different applications. You know, it’s again, I think, it’s less about specific milestones and demonstrating that you have the knowledge and capability on the team to execute.

Nick Moran 37:55
While we’re still talking about industrial automation in the industrial space, and innovation there, how do you think jobs will be affected? I mean, is this like a net? Negative on jobs? This is a hot topic, right? A lot of VCs, a lot of other people talk about this, but you know, with with more automation and more robotics, is this a net job eliminator or job creator?

Kane Hsieh 38:16
So, look, I’m not a labor economist. So I don’t about the net job. Question there. Frankly, I just don’t know. But what I will say with confidence is that some classes or jobs will be eliminated. And this is inevitable. Well, let’s just look at some examples. Like, when’s the last time you saw crossing streets? Like these are literally people whose jobs was to remove the horseshit from crosswalks? Right? No one knows. Like we should Oh, like, we should go back to that time. It was awful, literally, like streets were covered in horseshit

Nick Moran 38:52
carriages or wagons? I mean, all those things went away, right?

Kane Hsieh 38:55
Same with like, Lamplighters when’s the last time you saw something? Exactly? When’s the last time you saw a Lamplighter? Right? Like, I’m very happy that we don’t have someone lighting explosive gas every night on every street corner. Like that was it is great that we automated that. We don’t have knocker uppers. Because now you can buy in a like your phone has an alarm clock, you don’t need to pay someone to come knock on your window every morning. Like, these are jobs which got automated away. And, you know, it’s invisible to us because we accept it as inevitable now that we’re living in it, but I acknowledge, look, the pain of these transitions is real, right? I’m not going to pretend it isn’t real. But at the same time, it’s inevitable. And what I mean by that is, companies and issues can either choose to weather, the pain, and I don’t know what will happen, someone when someone lose, but the only way to guarantee your loss is to bury your head. And like ignore automation because you want to preserve some jobs which within our lifetime will seem absurd as a Lamplighter. Yeah. Uh,

Nick Moran 40:00
no, I mean, I couldn’t, couldn’t agree more. I think that’s, that’s very fair, you know, progress in innovation is inevitable. So you’re either a part of it and part of making it productive in some way. Or, you know, you let it happen all around you. You know, what’s one piece of advice you’d get for founder’s like early stage founders in some of these automation categories? Or, you know, doing something in robotics?

Kane Hsieh 40:27
Yeah, so this goes back to the earlier point you’re talking about in the challenges of industrial automation is like, I think the critical thing is really understanding crew gets value from what you’re proposing, and who, if anyone, will kind of be an implied enemy, because you’re eating their cake. Right, understanding the perverse incentives. And the value chain in these processes, I believe is critical. Now, we, in our restaurants across maritime, and shipping, construction, and oil and energy, like every one of those companies has clearly identified like, here is the specific party we need to sell to, because by selling to them, we are able to exert influence in this process, versus if we sold to this other party, which, you know, will sandbag us because of like the systemic inefficiencies that they benefit from. It’s not enough to propose the automation, you really have to understand where that automation sits and who’s paying for it? And who’s willing to pay for it.

Nick Moran 41:26
Like that? Are there some categories within sort of the broader deep tech space? Or, you know, some of these capital intensive investment areas? Or are there some sub sectors or sub segments that you’re just avoiding?

Kane Hsieh 41:40
Yeah, a few. Pharmaceuticals? Again, this is like that has its own special class of investors for variety reasons. Yep. Material Sciences, the unit economics and materials companies can be pretty challenging. To get this isn’t to say that people will make money here. It’s just for our strategy, and font size, and talents, like these are areas we’ve chosen not to pursue for now. Oh, passenger aerospace. That’s a big one. That’s gotten more popular traction recently. Yeah. But I mean, look, I’ve been in helicopter factories. And it’s like friends with straight of helicopter pilots. And my conclusion from all that is like, never get in a vertical takeoff aircraft unless you really have to. Because the failure like the default mold of a fixed wing aircraft is like you can relatively expect it to be gliding, the default mode of a VTOL aircraft is it’s literally falling out of the sky. And there’s just like a lot of weird edge cases that occur. And look, I dream of building my own, like, kit aircraft, like I love this stuff, I will put myself in arbitrary amounts of mortal danger for the adrenaline of it. But if I’m thinking about like what it takes to certify power plants and airframes for passenger aircraft, like I can’t see within, at least within the span of, you know, the funds, I’m working at how to make those economics work.

Nick Moran 43:06
Yeah, I don’t blame you. I mean, those investors that are getting into those spaces, they are braver than I am.

Kane Hsieh 43:13
Just to be clear, I’m happy that investors are investing in it because I want I want V tall aircraft to exist, I want new generation of supersonic transport. This isn’t to say like, I don’t want this to exist or think it shouldn’t or won’t exist. I don’t believe for our fun. And for a lot of our peer funds, it is an area that makes sense to invest in, but God bless them, more power to them.

Nick Moran 43:39
So again, I heard a rumor that you’re starting a podcast.

Kane Hsieh 43:41
Well, it’s it’s 2020. And I work in tech live in San Francisco. So that’s just the reasonable problem with that. Now the difference between an idea and a follow through is you know, the important part so so I’ll let you know if the latter actually happens. Okay.

Nick Moran 43:57
Another fun thing I came across on the profile side project called transformer poetry, a book of famous poetry reimagined by open AI as GPT two language model. I don’t even know Yeah, I don’t even know what to ask specifically, but we’d love to hear you know, the inspiration for that.

Kane Hsieh 44:15
Yeah, so so it was the kind of odd moment in time where my close friend who’s the co author of that paper, David, you know, he lives with me, and that was published, I believe, end of February last year, shortly after les Murray, the famous Australian poet passed away and I happen to be reading some of his poems just because he’s a very influential contemporary poet. And I guess, you know, just sitting at home with with David there during his work and reading his poems, I I just thought it’d be interesting to mash together famous poems in GPT. Two, you know, part of this was just that random, you know, stew of events that occurred but but the other part that drove it was, personally I don’t actually I don’t believe that anything is sacred or unique or human. I’m pretty reductionist in my views that way. And I enjoy using technology to push little wedges into creative processes and seeing what happens. Now, I actually read the book elements of eloquence by Mark forces shortly after, and it’s like a technical breakdown of poetry. It’s great, I highly recommend it. But I wish I’d read that before the transformer Poetry Project because in retrospect, transfer poetry like really like as poems, they’re pretty bad. And now that I understand how poetry is, you know, but like details of how poetry is appreciated, that said, Look, five years ago, computers weren’t even in the same arena. Right? The fact that I can say, yeah, the computer poems sucked, like, at least we’re, we’re, we’re in the same arena now. Sorry, I say we as if like, I’m also computer but I get at least the computers in the same arena. And that’s interesting to me, right? Yeah. Literally, the first thing I say in transforming poetry is like this book is an experiment, it is not meant to have academic or literary value. And I believe that this was an experiment. I want to see the reactions in the world from, from the world literature from technology, like it was, I like experimenting. You know, I like I like starting little fires. Was

Nick Moran 46:18
there one outcome or one piece of poetry that came out of that that was, you know, well received and worked? Well.

Kane Hsieh 46:25
I think in a superficial reading, they all kind of work. I think it was more entertaining than well received. So for instance, like Dr. Seuss poems got really dark, which is funny to me. And then on the flip side, there was an ad Girl and The Raven became really, like hopeful. Oh, right. So it was like the system generated English, but things in poetry like meter, like, like the elements of eloquence, they were genuinely lost. And so I don’t think anyone was particularly well received. I think some were like, funny. But hey, like, you know, ask me again in 10 years, and we’ll see, we’ll see where the technology is.

Nick Moran 47:06
Right? Cane, what resource, you know, can be a book, blog, video and article a tool. But what in particular, have you found valuable and you’d recommend a listeners

Kane Hsieh 47:17
a newsletter off the top of my head, it’s called the prepared. And it’s run by a guy in New York named Spencer, right. And it’s not venture service specifically, it’s actually just covers interesting things in industrials, and manufacturing and logistics, etcetera. And I think not only does it provide very interesting insight into the industries that that I talked about, and I invest in, but the community that Spencer has put together as a tight knit group of like passionate experts. So I’m a big fan of that one. books I’ve read recently that I’ve really enjoyed. In 2019, it was their perfectionist perfectionists. I’m pretty sure you would like this spike. I’m Simon Winchester. And the book is about how humans bootstrapped precision, which is this pretty trippy thing if you think about it, because high precision machines need other high precision machines that make them and so like how do we get to high precision missions in the first place? Oh, interesting. Yeah, it starts with the steam engine and ends with the gravity wave probes, and every chapter moves in order of magnitude, and scale and precision. And he talks about the critical breakthroughs at every generation of precision that enabled us to bootstrap the next generation of precision. And it’s just like a fascinating I mean, it is by definition, has a history of disruptors. So I really liked that book. The book I haven’t read yet, but I’m very excited about this year to read. It’s sitting on my desk right now. It’s called rust, the longest war. And it’s about humanity’s efforts and history and fighting corrosion. Take it with a grain of salt, because I haven’t read it yet. But But I think this is going to be a very interesting book.

Nick Moran 48:58
Super cool. came, what do you know, you need to get better at?

Kane Hsieh 49:04
What do I need to get better at? I mean, the Blyth answer here is I need to get better at investing. Because that’s my job. But what I think, you know, we talked earlier about how people that spend a lot of time in either like process engineering, or manufacturing, engineering or mechanical engineering are, by default, more conservative and cynical than people in software. And part of that is because of the cost of failure in their practices. The fine line that I have to thread as a VC looking at these opportunities is like, when is the optimism the opportunity and when is the optimism naive? And I don’t have I mean that that’s the million dollar question. Right, wherever the billion dollar question. And so I do believe that I’m still more cynical than I should be to be fully effective. But the hard part is like ratcheting up that optimism. and seeing like, what is the equilibrium with the highest expected value?

Nick Moran 50:06
Interesting. And then finally, what’s the best way for listeners to connect with you?

Kane Hsieh 50:11
If people are working on or thinking about opportunities in the spaces or adjacent spaces have recovered, I would love to learn more. My email is just my first name cane at root like tree roots.vc. That’s cane like Citizen Kane, not like the biblical cane. And similarly on Twitter, my handle is just at Cane. So pretty, pretty easy to find on email or Twitter.

Nick Moran 50:34
There you go. Well, as an investor, you know that I love the thesis. I love the chat. It’s it’s very near and dear to me and hits close to home. So you know, thanks so much for elaborating on the thesis and going deep on on industrial automation. Yep.

Kane Hsieh 50:51
So look in bed. Last thing I would say here is, besides thanks for having me on the show is as a fund as an investor. These ideas and frameworks are constantly evolving. They have to the industries are evolving. The market is evolving on the financing side. This is what I believe right now. But this is by no means static, nor necessarily correct even. And so I welcome anyone who either wants to challenge these or to think more about them to contact me. Perfect.

Nick Moran 51:25
Awesome. Well, his his email is Canaan route.vc and his handle on Twitter is at Kane, so easy to find. And I encourage you to reach out to him Kane, thanks again for for joining us.

Kane Hsieh 51:37
This was great. Yep. Thanks a lot.

Nick Moran 51:44
That will wrap up today’s episode. Thanks for joining us here on the show. And if you’d like to get involved further, you can join our investment group for free on AngelList. Head over to angel.co and search for new stack ventures. There you can back the syndicate to see our deal flow. See how we choose startups to invest in and read our thesis on investment in each startup we choose. As always show notes and links for the interview are at full ratchet.net And until next time, remember to over prepare, choose carefully and invest confidently thanks for joining us.