Venture capitalists are investing heavily in technology, but what about deep-technology and AI companies? In today’s show, we speak to Jason Schoettler and Kevin Dunlap, both veterans in the technology sector. They have both worked in a range of areas within technology, such as working on the Mars Rover, with most of their focus on robotics and AI. They then made a career change and integrated their vast technological experience and their passion for innovation to form Calibrate Ventures, a venture capitalist company investing in advanced, emerging technologies that reshape industries and the future.
In today’s conversation, we learn about Jason and Kevin’s career backgrounds, the potential that the deep technology sector has, what ideas excite them the most, the challenges when investing in AI-based companies, what kind of technology is easily understood by the consumer, what makes a technological innovation successful, and much more.
Key Points From This Episode:
Tweetables:
“I think for me personally, the cycle-time was very long. You work on projects for a very long time. As an investor, I get to see new ideas and new concepts every day. From an intellectual curiosity standpoint, there couldn’t be a better job.” — Kevin Dunlap [0:05:17]
“So that lights me up. When I hear somebody talk about a problem that they are looking to solve and how their technology can do it uniquely with some type of competitive or differentiated advantage we think is sustainable.” — Jason Schoettler [0:08:14]
“The things that really excite us are not, where can we do better than humans but first, where are there not humans work right now where we need humans doing work.” — Jason Schoettler [0:20:44]
“Anytime that someone is doing a job that is dangerous, that is able to be solved with technology, I think we owe it to ourselves to do that.” — Kevin Dunlap [0:22:39]
Links Mentioned in Today’s Episode:
Calibrate Ventures on LinkedIn
GrayMatter Robotics on LinkedIn
Kevin Dunlap 0:00
So anytime someone's doing a job that is dangerous, that is able to be solved with technology, I think we owe it to ourselves to actually do that.
Rob Stevenson 0:12
Welcome to how AI happens, a podcast where experts explain their work at the cutting edge of artificial intelligence. You'll hear from AI researchers, data scientists, and machine learning engineers, as they get technical about the most exciting developments in their field and the challenges they're facing along the way. I'm your host, Rob Stevenson. And we're about to learn how AI happens. Joining me today on how I happens are two partners from calibrate ventures. On my left Kevin Dunlap. Kevin, welcome to the podcast. How are you today?
Kevin Dunlap 0:47
I'm doing well. Thanks for having me, Rob. Really appreciate it.
Rob Stevenson 0:50
Really pleased to have you. And on his left Jason shuttler. Jason, welcome to you as well. Thanks, Rob. Really pleased to have you both. You both have really interesting backgrounds. Maybe we start there. Kevin, do you want to maybe share a little bit about your career and background and kind of what brought you to found Calibri ventures?
Kevin Dunlap 1:04
Yeah, sure. I'm happy to give you a little bit of my background. And then Jason can give his and our stories kind of intertwine over time. But I started my career as a mechanical engineer, worked in automotive for a little bit and then actually transferred into aerospace. I was an engineer in jet propulsion laboratories. I was there when Spirit and Opportunity were actually landing on Mars, it was a great time, there's a lot of fun. But frankly, space wasn't what it is today, right from an excitement level. That was just big government. If that was different, I might still be in space. But at the time, it wasn't. But I got really inspired by early stage tech in taking those applications and background and working with founders to bring new technology to markets. And so that's really my background, how I got started. And while I was in grad school, actually met Jason and finance class, we were working together on a couple projects. And we had some great chemistry, great overlap and skill sets, and ended up joining a large family office and working together. And so I'll let Jason get his background. And
Jason Schoettler 2:07
well, Kevin's been Kevin's being a little polite, I walked into finance class, and saw this guy working Excel without a mouse, which I did not know how to do, I thought that was a pretty sweet trick. So I sat next to him. And that's where we met almost 20 years ago. And so the rest of my background just to jump in is I've been in tech my whole career. I started up in the bay area doing management consulting on the technology side during the late 90s and then joined a startup that was spinning out of Jet Propulsion Laboratory and Caltech here in Pasadena. That's what brought me down to the Pasadena area. And that was a fantastic technology that was developed to help manage their mission for the Mars Pathfinder, which was just incredible technology, literal rocket scientists that we got to work with. And out of that I was recruited to join one of the investors in that company, which was a large billion dollar family office based here in Southern California, as the first investment professional, it was a great opportunity, I got to learn the craft of investing as an early venture capitalists from one of the pioneers who been doing venture from the late 60s. So a lot of history, a lot of lessons imparted on me and Kevin, who joined shortly after I joined. So for over a decade, we worked in and made investments throughout multiple business cycles, and also across various stages of investment and various technology sets as well. We founded calibrate, really, because we became very interested in the technology in and around the year 2011, when we made our first investment in a deep technology company, we saw that in that company, the power of really low cost affordable automation and navigation systems, and how those technologies can really build affordable products for consumers. And that investment was a company called evolution robotics, based here in Pasadena, actually, that company was acquired by iRobot. And the technology that evolution built, is now embedded in all Roombas in Palo Virginia, and the founder of that company, and his team did just an amazing job, also kind of JPL DNA for that company, which was great. And so that led us to make additional investments in and around companies that we're using. Initially, our interest was in computer vision software, and sometimes that included hardware, and sometimes that included robotics as well. And so we started to make investments in and around those technology sets, including we led to a series B at ring, which is a company that was acquired by Amazon a few years back and just continue to make additional investments in consumer hardware companies and gradually began investing in more industrial technologies as well. And I think the draw for us was we watched prices come down for these technologies as functionality improved drastically. And I think that is a really wonderful recipe for software innovation in and around deep tech.
Rob Stevenson 4:59
So Having established that you both grew a little bored with landing robots on Mars. I'm curious what you find more exciting than that. When you think back from moving on from your background there, was it just like the next iteration or another application of robotics specifically, or? Or why sort of point your attention at this venture? Or this industry anyway?
Kevin Dunlap 5:19
Yeah, I mean, I think it's hard to getting bored when you're blaming robots on Mars, right. But I think, for me, personally, the cycle time was very long, right? You work on projects for a very long time. As an investor, I get to see new ideas, new concepts every day. And so from an intellectual curiosity standpoint, there couldn't be a better job, right. From a technology standpoint, you know, a lot of AI ml has been around for a long time, the concepts of it have. And what we started really realizing was that a decade ago, computer vision, and ml could work, but it was expensive. Compute was clunky. But what we started to see with ring and some other companies like ring, people don't think that it's massive computer vision company, but it really is, right, it's got a lot of data needs to train a lot of algorithms with synthetic data, because of privacy restrictions. And so we started to see computer vision starting to actually make sense and work at scale, right, there was an ROI to a customer, the perception of a synthetic data image relative to a real image was starting to, to actually narrow compute power at the edge is increasing. And we started to see those trends actually increasing an exponential level instead of a log level. And that really led us to start thinking that there was an opportunity to build real enterprise companies here, and businesses as opposed to cool tech. And that's really why we found to calibrate to focus on that, that deep tech that can scale in the next three to five years, and actually get into product, start selling to end customers and start building a business. I think there's some really interesting things happening in the longer form. But you know, I think that's why you mentioned a little bit, a lot of people you've had on the show before have one foot in academia, and one foot in enterprise, right? We're venture capitalists, so we got to be an enterprise. Yeah,
Jason Schoettler 7:15
I'll just add, for me personally, from an interest level, I think there's some real world problems that this technology has evolved to be able to solve. Now, in a really elegant way. I think rings are a great example of that there are others that are less obvious or less available to the consumer eyes, but they're certainly a really powerful set of technologies that are going to make real improvements and make a dent in the world in the decades ahead.
Rob Stevenson 7:37
So when you are receiving pitches, I guess what tends to excite you as Enza, like raise your eyebrows? So I'm sure you see a lot some of it maybe doesn't inspire you that much. Some of it's like, okay, maybe this would make money. But it doesn't excite me from innovative perspective. What kind of lights you up when you're reviewing new tech out there? For me,
Jason Schoettler 7:56
I always start with the question of who's your customer? Yeah, what do they value? And what are they willing to pay for that value? If you can't answer that question, in my mind, you don't have a really well thought out use case, you probably have some kickass technology that can do some really amazing things. But from a commercial perspective, really understanding the customer's need, the customer's use case is critical for me. And so that lights me up when I hear somebody talk about a problem that they are looking to solve, and how their technology can do it uniquely, with some sort of competitive or differentiated advantage that we think is sustainable. That's kind of the first step. And then if that customer use case is big enough, that we can see a huge market potential for that, we get really excited and start to dig in. But for me, it starts really at the customer.
Kevin Dunlap 8:46
Yeah, I'm every interview I've done. When I say it comes to people, you know, the interviewers get upset, because that's the standard answer. But it really is, right. I mean, for me, it comes down to the team. When we make an investment, we can be together for three for 10 years, right. And so I also want to work with teams that are focused on building awesome tech, that's applicable in enterprise, right. And so when you find that marriage of those two, and they're focused on building real world product, with advanced technology, that's what gets me excited as the first step. Right. And then I want to dig into how do we think about understanding the bigger problem? Is this a nice to have solution or is this a half to have solution? And so our founders come from all different types of walks. We've got two postdoc, PhD roboticist out of USC. And we've got more senior executives that have left large organizations that are starting companies. And so it can be any walk of life, but then excitement and the hunger to go build something new and exciting really has to be there too.
Jason Schoettler 9:55
I can give you an example of one that may tell her the sand a little bit. We've looked at AG Tech for many years, but not really pulled the trigger on anything until a couple years ago when we met a company called farm wise. And what we found really compelling about farm wise was that growers and farmers are facing labor shortages. I mean, so you got to look at the macro, right. So there's labor shortages, in part because of immigration, in part because of people just don't want to go work the fields anymore. So growers are facing a real world problem that they can't get the labor to go work the fields, when they can find the labor, there's rising labor costs, typically associated with that, especially here in California with minimum wage mandated to go up and in the background, or overlaid on top of that as sustainability and a desire to do more with less. And that was the macro table that they set when they came in to talk to us. And they told us about their solution, which was using computer vision and robotics and automation, to be able to go weed fields. And they can do that at the same or lower price point and build a profitable business. Doing that by going to sell their service and their technology to growers. And so for us, we found that really compelling because there's a big macro trend, a major pain point for the customer set. And we know from our travels about the computer vision, and the hardware is arriving to a price point that you can go make a real dent in delivering ROI to the customer, and build healthy margins into your business as well. So that was a really, I think, good example of weaving together a real world problem with the technologies that are available here today.
Rob Stevenson 11:34
Yeah, that makes sense. And in terms of an awareness of who a customer is a market opportunity that feels like sort of standard company evaluation that you might see for your run of the mill software company. Where would you say it differs when you are evaluating the space of AI, ml, computer vision, etc, from your standard software as a service company that might just be a flashy replacement for a spreadsheet?
Jason Schoettler 11:59
I'll jump in and answer. I mean, I think there's one real key difference right now, and that is businesses and people know how to go out and buy SAS software, right, there's hundreds, if not 1000s of you can go subscribe to have your dog walked or you can go subscribe to have, you know, help with an enterprise problem as well. And so it is a very well understood buying pattern, very well understood kind of pricing pattern very well. productize. And very well monetized. It's very familiar to everybody. I don't believe that's the case yet for AI. And I think that's a key difference. And I think when we look at the market for these technologies, we're kind of where we were 20 years ago, when SAS was just becoming a thing. When Salesforce came in and was disrupting some old line software companies with this idea that you never needed to maintain software anymore, you just use the app, I think we're at the kind of at that stage right now for AI. So it's still very specialized. Customers are typically pretty hardcore data science teams, academics, very forward thinking technology companies. And we think that as the use cases become more prevalent, as the technology becomes more productized and easier to use, we'll see more widespread adoption. And so that's an important differentiator as compared to SaaS companies.
Kevin Dunlap 13:19
And I think the reason that we're at that point in time is really allow the applications for AI and automation are mission critical to businesses, right. So you've got longer sales cycles. If I simply want a new CRM system, there's dozens I can go get, I can pull my info out, I can load it back in, it's pretty, pretty simple now, right? But if you're going to someone and say, we're going to automate a portion of your middle assembly line for someone like Tyson Foods, or you're going to Toyota and say, you should really be reducing your cost to collect data and improving your development time, with a company like parallel domain that does synthetic data, those are things that take longer to evaluate. The sales cycles are longer. But once you're in, you're usually, right, the switching cost is higher. And companies are starting to realize that they need to implement these solutions to improve their speed development, reduce their dev time, and actually get product to market faster. Right. And as they work through those, I think that's something investors have to be focused on too, right? It's not an immediate sales cycle. And so understanding that it takes longer, but you could ultimately be more valuable, I think, is one of the keys that excites us from an economic perspective as an investor.
Rob Stevenson 14:37
Is that longer sales cycle a result of what Jason You were just kind of explaining, like a lack of familiarity with how to buy these products?
Jason Schoettler 14:44
I think that's part of it. I think that we just don't have as many people in these organizations on the purchasing side who are as familiar with these technologies that will change. I mean, that is changing. It will change over a generation, which is why we're excited about Being on the ground floor. But I think there are other Kevin touched on another really important reason why the sales cycles are longer, which is they're often going to very mission critical parts of businesses where if something goes wrong or goes haywire, like, there's major problems. And so I think it's more reality of selling into enterprise into more mission critical parts of businesses. But also, it's coupled with just the general adoption curve for these technologies into the more mainstream.
Kevin Dunlap 15:27
And we've got slightly different business models, right? Sometimes you're selling data. And so you're tearing your level of data, sometimes you're selling robots as a service, sometimes you're selling the upfront system with a subscription. And now you got to get finance teams at large organizations to think a little bit differently than they may have in the past. Right. And so there's some education that companies need to do, not necessarily with the end customer, that's gonna be using it every day. But actually, with the finance teams, help them understand the ROI, help them see the value of that. But once you get through that hurdle, you're kind of through it, right? They get comfortable with a contract, the purchasing power and everything else.
Jason Schoettler 16:07
I mean, example that is just as a company, our portfolio that was going in to talk to a customer about using their technology as a service, or the finance team has never, you know, they have to run their ROI calculation, and they had never come across something where there's no upfront capital investment. So they didn't have a way to within that organization, translate, there is no upfront cost to just pay as you go into their ROI model and the way that they think about and make decisions about their organization. And that's, that's one of many, many examples of the business side of things really needing to be understood and adjusted to so you can get through and get the technology adopted,
Rob Stevenson 16:46
that feels like a significant hurdle. Particularly if you were a different investment firm, you know, a different investment firm looking at it, the same kind of company that you might be looking at might say, Hey, you have a great product seems like there's a real pain point here. I don't think people are ready to buy this. I don't think people know how to buy this. So we're gonna pass. I'm sure that comes up constantly for you like, is that just an assumed part of the space for you now?
Kevin Dunlap 17:07
Yeah, I think it is. I mean, that's why we always talk about deep tech at scale within three to five years, right? Because we know that we need to take a little bit of lag in and faith in the product, and the actual value that can be delivered with that product. But we build a portfolio of companies, right. And so we have investments in lots of different verticals, really to kind of alleviate some of that, right? Because we won't get it right all the time. But we also won a lot of deals because we have this experience, right? So we beat out larger, more well known firms, because when we sit down with the founders and start explaining the business aspects, they're like, Oh, I haven't necessarily thought through some of those things where this other firm didn't bring that up. They're just looking at my assumed customer acquisition costs right now, which, in the earlier stages, you know, it's really a while it's a WAG, right, you're trying to figure it out. But you're right. I mean, I think we attracted a handful of founders, because of the experience and where we focus.
Rob Stevenson 18:04
Yeah, also prioritizing or indexing on that timeframe, probably one way around that as well than not like what is impactful on this slightly shorter timeframe. So what are some examples maybe of the types of tech people do understand are a little more comfortable pulling the trigger on,
Jason Schoettler 18:21
I can give you one example from our portfolio, there's probably another one in a company that we know that that we're not investors in. But let me start with a company called regard, which is in the healthcare space, and I love my doctor, he's wonderful, he does a really nice job, but he can't spend his time scouring all the data that there is about me. And one of the things that we see as compelling opportunities in technology is machines can can ingest and learn lots of different datasets and make conclusions that it's very hard for humans to do with they've given an enormous amount of time, but I only get like 15 minutes with my doctor sometimes. And so regarde works with hospitalists to pull in all the information from a medical record, you know, labs and all that, and evaluate all that data, and essentially recommend a diagnosis for what's going on based on what it can read. And it's reading notes. It's reading datasets about blood pressure, different vital signs. And what we found is the doctors miss stuff because they're human. And I think one of the things that I love about my doctor, is it he is human, but boy, do I really want to have him not missing anything. And so the idea that we can have a machine as an assistant, not to replace the doctor, but to augment that doctor. It's a really interesting use case where technology can be leveraged in a particular setting to deliver a service. Like that's one example where it's a win win saves the doctor time, I get a better outcome because we're not missing anything because the doctor is tired or overworked or whatever it is. They all are. And so I think that's one case that may not be obvious, but certainly would be embraced by a All participants in that ecosystem?
Rob Stevenson 20:02
Yeah, it's kind of a cui bono sort of question, right? Like the stakes are high for you. And that scenario, Jason right, even if you didn't have this background in AI, so someone would want to have that sort of safety net almost, if their health was the question was what was being evaluated there. And you brought up something that I wanted to dig into a little bit, which is this notion of, okay, the tech isn't replacing the doctor, but supplementing their work augmenting them. We speak a decent amount about human in the loop on this show. And it also is just related to this larger question of in what cases do we expect tech to work alongside humans versus automate away work from human? So I'm curious where you all come down on what jobs should be automated can be automated versus augmented via human the loop type approaches and how you sort of evaluate that balance?
Jason Schoettler 20:49
I mean, I'll jump in. And I'll first say that the things that really excited us is are not where can we do better than humans? But first, where are there not humans doing work right now, where we need humans doing work. And so if you look at the demographic trends in a couple of different industries, I mentioned ag earlier, certainly there are fewer people going into ag construction, and other trades, some aspects of manufacturing, are just not seeing the talent pools available. And we are really interested in those macro trends that we don't think are going to reverse overnight, if at all, and investing behind technologies that solve a problem where there's an acute labor shortage. So in that it's kind of a different bucket than what you pose. But that's a really important one. And one that really, really excites us, because we've seen technologies developed to a point where we can actually insert them into scenarios where there's a real acute labor shortage to have have a great win win win, as it relates to human in the loop, or what should we automate, where Shouldn't we automate? The major point is that we want to see, humans do what they're best at. And we want to see technology do whether it's really good at and in many cases, technology can take away the routine, and the mundane and free people up to do more creative and meaningful work.
Kevin Dunlap 22:08
And I think we've seen that happen in the very high volume, low mix scenarios in manufacturing, we're starting to see that in the middle volume hire mix. Alright, we have a company called gray matter robotics, they do autonomous scanning and sanding for material removal. The people that are doing this job right now are getting carpal tunnel syndrome, within a year to three years of starting these jobs, they're very repetitive, they're inhaling dust and debris into their lungs. So they shouldn't be, you know, this is something that's ripe for disruption. That wasn't possible five, six years ago, right. And so anytime someone's doing a job that is dangerous, that is able to be solved with technology, I think we owe it to ourselves to actually do that. There's lots of jobs that where we need other people. And if we could find other places for them where they can add more value. I think we should be doing that as a society.
Rob Stevenson 23:06
Yeah, yeah, that makes all the sense in the world. We are creeping up on optimal podcast length here. Gentlemen, before I let you go, I'm going to do a quick whip around for both of you to hear a little advice to our AI entrepreneurs out there. If I'm building in the space, I'm a practitioner, I have a side hustle, I have an idea I have a project, what advice would you give for them to make that into a reality into a viable business?
Kevin Dunlap 23:29
I would start by understanding your customers needs. And I think that's the most important thing when you're building a company is understanding the value that you can provide to an end customer, right, you need to understand their pain points, need to understand what their budgets are, you need to understand how long term problem it actually is for them. And then you need to understand who else is doing something similar or potentially moving into your space. Because it's not just about capital in venture, it's also about your speed of execution. And so those were the places I'd begin with. And as you get further along, then you've got to better understand your unit economics, you've got to understand your business model. How can you charge How can you make money? How do you need to grow and support an organization. But those are the quickest bits there.
Jason Schoettler 24:18
I would agree with both of those. And I would add, surround yourself with people that don't necessarily think like you who come from different walks of life, particularly if you've been in technology for most of your career, most of your education, getting to network and find mentors and find friends and just start building out a much richer and diverse set of a network. I think that can go a long way to help inform some of what Kevin I think rightly mentioned, in mentorship, I think there's no substitute for it. So finding great mentors has served us well, and I think it would serve any young young or older entrepreneur. Well,
Kevin Dunlap 24:53
yeah, in addition to Jason's comments, I think community is also important, right? And as part of that, We're doing an event in October of this year where we're bringing together builders implementers and academics to actually talk about perception and learn from each other to better understand what needs the implementers have what the builders are building, and also tying that all together with academia.
Rob Stevenson 25:18
That sounds awesome. Is there a signup link we can drop in the in the show notes can you provide that I'm
Jason Schoettler 25:23
sure we get that it's called the edge of now or eon for short,
Rob Stevenson 25:27
eon edge of now make sure you all check it out out there in podcast land. Kevin Jason, this has been a wonderful chat. Thank you so much for being with me today and sharing your experience and your views on a space I've loved learning from you both great. How AI happens is brought to you by sama. Sama provides accurate data for ambitious AI specializing in image video and sensor data and notation and validation for machine learning algorithms in industries such as transportation, retail, ecommerce, media, med tech, robotics and agriculture. More information, head to sama.com