Collin Stedman - CTO at Predata

(Joel Beasley at 00:00:02) This is the Modern CTO Podcast. Hey, buddy. How are you?

(Colin Moody at 00:00:14) I'm doing alright. How are you?

(Joel Beasley at 00:00:16) Fantastic. Alright. Are you out of WeWork?

(Colin Moody at 00:00:18) I am. Yeah. And as soon as you joined, somebody started talking out there. So, unfortunately, I can't get to a completely silent place. I apologize.

(Joel Beasley at 00:00:27) No worries. Where are you at specifically?

(Colin Moody at 00:00:31) We're in Manhattan.

(Joel Beasley at 00:00:33) Oh, I've been in that WeWork before.

(Colin Moody at 00:00:35) Yeah. Around Soho, one of the WeWorks. There's so many of them. We're just in one of the older Soho WeWorks, actually.

(Joel Beasley at 00:00:44) Oh, there's more than one in Manhattan?

(Colin Moody at 00:00:46) There are a lot now. Many.

(Joel Beasley at 00:00:48) Oh, okay.

(Colin Moody at 00:00:48) Yeah. WeWork has taken over the world. I think they're one of the largest real estate companies in all of New York now. Yeah. So they're pretty large. But they're just renters too. They don't own any of this. They just have a massive, they're one of the biggest renters, I guess.

(Joel Beasley at 00:01:05) Yeah. I saw an article the other day about how they took a play from the McDonald's playbook about how they booted up. Ray Kroc, I believe is his name. Have you ever watched the Netflix? It's a movie, but it's the McDonald's show.

(Colin Moody at 00:01:21) Yeah. I actually haven't seen it. It's on my list, but I haven't watched it.

(Joel Beasley at 00:01:26) I suggest it. I think it's really good. Yeah. It's very clever.

(Colin Moody at 00:01:31) Is this how they franchised it in order to make it pop up faster?

(Joel Beasley at 00:01:36) Yeah. It's the whole story of the business.

(Colin Moody at 00:01:38) Okay.

(Joel Beasley at 00:01:39) And it's quite possibly the most interesting thing you're going to watch this month.

(Colin Moody at 00:01:45) Sounds good. Yeah.

(Joel Beasley at 00:01:46) Good hype. I mean, look, it's on every corner. It's kind of interesting. You want to know their origin story, right?

(Colin Moody at 00:01:51) Yeah. The other one I've been meaning to watch. It just got on Netflix. I forget. It's called American Company or something like that. American Factory. And it's about a Chinese investor investing in an American, I think glass factory that was sort of winding down, and they're trying to create some cross-specific business with glass factories in China, but also in the U.S. It looks interesting. Seemingly interesting given the current political climate, but it just came on Netflix.

(Joel Beasley at 00:02:29) Yeah. I just saw the Jack Ma and Elon Musk clip on YouTube from some international summit that was in China, and Musk was blown away by China. He was like, this is unbelievably advanced. I've never seen people build something with such high quality so fast. It's really just mind blowing.

(Colin Moody at 00:02:49) Yeah. That's it's interesting to hear Musk say that because one of the things that I've been thinking about a lot recently with the work Musk is doing is, can we automate so much of the process of manufacturing to a point where the labor costs are less significant and we can manufacture anywhere? You know, in a world where the labor costs are one of the largest drivers, of course so much manufacturing has gone offshore. And if labor costs become much less significant and instead the access to know-how for automation becomes much more significant, and the U.S. is one of the largest markets for that experience, will it make more sense to move factories back or at least much closer? And I think that, you know, trade war aside, and that's certainly already driving a lot of manufacturing out of China into other developing markets where labor costs are lower, but the next step might be to bring factories much closer to the end consumer. You know, wherever the main costs will be, the access to the raw materials, the cost of shipping goods to the end consumer, and the cost to maintain and improve an automated system. And in that world, I wonder if a lot more manufacturing will come home or close to home.

(Joel Beasley at 00:04:08) I think you're right. I think, ultimately, after this long journey, we end up where we started in these small communities that are self-sustaining. Right? Because that's where we started, small self-sustaining communities, and we were hunter-gatherers, and we were moving about all the time. And then we separated and isolated specific things to be manufactured in specific places. And I think ultimately, we'll all come back to super advanced 3D printing machines and local gardens that are run by autonomous things. Everything will just be in these self-contained. It's like distributed systems, right? That's the most stable system you can have is a collection of self-contained systems.

(Colin Moody at 00:04:44) Yeah. I'm hopeful that we'll get to that future. I mean, I still think that one of the very interesting conflicts in technology is whether or not it makes more sense to distribute it or to centralize, and I'm not sure that it's always one way or the other. So, for example, in renewable energy, a lot of solar has been extremely decentralized. It's residential. We put individual solar panels on various homeowners' roofs. And that actually turns out to be very difficult to manage, and there's a potential advantage to centralizing more solar, to have large solar farms where you can have a dedicated interconnection to the grid. Actually, having distributed assets connected to the grid causes a lot of trouble for the grid. So this is just an example of how I think that decentralization makes systems much more resilient, but it can also come at an efficiency cost, for example. Same thing in Bitcoin and cryptocurrencies. Look how much we spend in energy, or whatever the proof is, to have this decentralized system. It turns out that if you have a system where trust is assumed and managed by a couple of central authorities, that's a very efficient system. It might not be the most resilient if it turns out that that trust gets broken, but it's very efficient compared to the trustless alternative.

(Joel Beasley at 00:06:19) Be an interesting tech talk, the spectrum of decentralization, right? Because technically, once you split it into two, it's kind of decentralized. So it's like, where does the efficiency, and where is the spectrum? Because, in our state, we have this company called FPL, Florida Power and Light, and they started to create these miniature farms near schools, right, where you could buy into it. And it was a couple acres of land, but they're solar farms. Miniature ones for that community, and they would centralize them for little communities versus having them in the middle of the state where no one sees them. That way people can feel like they're part of it. So there's a spectrum of how much centralization or decentralization is too much versus its effectiveness.

(Colin Moody at 00:07:04) Yeah. I actually quite like that idea. I mean, that lets individuals feel like they're part of this. It's not something that's hidden, managed only by the grid operator or the energy company. But it also is probably much easier to maintain and make sure that those panels are working and that the interconnection with the grid is well managed.

(Joel Beasley at 00:07:26) Yes. Let's take it over to Predata. What does Predata do?

(Colin Moody at 00:07:30) Yeah. So Predata is a predictive analytics company that uses machine learning to relate online activity to geopolitical and economic trends. The problems that we're trying to solve, the clients that we work with today, fall into two primary verticals. So we have a finance angle. The applications there are perhaps the most obvious: make money or manage your risk to not lose money. And then we have a security business as well.

(Joel Beasley at 00:07:56) So you're tracking security threats and how they spread, or what do you do with the security side?

(Colin Moody at 00:08:02) Yeah. So a lot of this is it's still trends. It's identifying things like, is attention in the conflict in Kashmir on the rise, or specific aspects of that conflict, and being able to track that on both sides of that border and around the world.

(Joel Beasley at 00:08:23) Are you the only people doing this? Who else is doing this right now?

(Colin Moody at 00:08:30) Yeah. So I actually think that there aren't a lot of people doing this in an automated way. So I'd say that there are a lot of people who are trying to track global and economic trends, but especially on the geopolitical side, the state of the art is still very much employing human experts, domain-specific experts, analysts who try to ingest as much information as possible in order to come to some conclusion about where they think things are heading. So Predata, one of the key insights that Predata brings to the table is that there's just too much work for humans to do this job. There's limited hours in the day. You can't have an unlimited team of unlimited size, but there's so much to read. Humans can miss unobvious patterns in the data. And so Predata is bringing some quantitative rigor to a space where up to this point, it's been largely qualitative, but also we can just search through much more data. That being said, we're not trying to replace these human domain-specific experts. There's a lot of uncertainty in the models that we're building. That's because the geopolitical space is one where you have problems of international scope, lots of hidden information or information that would be very difficult to measure. And so there's a lot of uncertainty that you're not going to be able to remove from these models. It's not facial recognition. So rather than getting to a black box model where we think that we can tell our end client, just do what the, trust whatever the model says, use that in your decision making, what we instead want to do is augment the ability of human experts to make sure that they don't miss information that they might otherwise miss. We want to make sure that our models are showing their work in a sense. So we want the end user to be able to see why Predata came to the conclusion that it did, what went into that, and allow them to sort of confirm for themselves that they think that we're capturing something interesting or, alternatively, a bit spurious and that they can ignore it. But on the spectrum of AI is going to replace humans to AI is going to upgrade humans, we fall heavily on the latter.

(Joel Beasley at 00:10:53) And then we invest in Neuralink.

(Colin Moody at 00:10:55) Yeah. Right. Well, I hope so. Yeah. I mean, I want to remain relevant.

(Joel Beasley at 00:11:00) Me too. I'm team relevance as well. So when you were in elementary school, did you dream of geopolitical data?

(Colin Moody at 00:11:10) To be honest, no. I don't think I'd say that. My interest was very much on the technology side. You know, my entire life, I grew up inspired by astrophysics, things coming out of NASA. I thought I wanted to be an aerospace engineer at one point. And I built Rube Goldberg machines. I was very much an engineering sort of kid growing up. And then when I got to college, that was my first taste of computer science and programming, and that changed my life. I knew within a couple of weeks of taking my first computer science class that that was what I always wanted to do. That was the thing I didn't know that I always wanted to do because the time to iteration, the cycle time was so short, and I just fell in love with that. You know, any idea that I had or any problem that I faced, I could sort of rapidly overcome it and try something new and see if that worked. So that cycle time, the ability to refine a product very quickly that way was intoxicating to me. But I didn't want to rule out the liberal arts and being a person with a broader understanding of what was happening in the world. So I didn't actually, I didn't go to a purely technical school. I went to Princeton, and one of the reasons I chose Princeton was precisely because it had a fantastic liberal arts program and it had a Woodrow Wilson School of Public Policy. And so I've always pictured using technology to make the world a better place, and I understood that it was very important to have a strong humanities foundation. And I think that's why I ultimately ended up where I am today. And I found like-minded people at Princeton, and that's what became Predata.

(Joel Beasley at 00:13:12) That's amazing. Now, were you one of the co-founders of Predata, or did you come in after it was already established?

(Colin Moody at 00:13:19) Yeah. So I came in shortly after it was established. The co-founders were one of my good friends at Princeton and then a lecturer at Princeton. And I had taken his class on innovation, basically global innovation at the interface of technology and policy. So it was very much on how do you innovate radically in global markets in the face of various regulation around the world. And, you know, my friend had taken that class a year before, so he and the lecturer started Predata, and I realized that that was something I was very excited about. So I actually started working in college for the company.

(Joel Beasley at 00:14:11) That's exciting. So you went to college, and you actually got a job. You're the 0.1%, right?

(Colin Moody at 00:14:17) Yeah. I guess so. Yeah. I was working on this before, yeah, even before graduation. So yeah. Pretty early. Pretty lucky.

(Joel Beasley at 00:14:26) But that shows you the importance of relationships throughout your career, right? You meet people. It's very important.

(Colin Moody at 00:14:32) Yeah. Absolutely. It's very important to communicate, yeah. You have to be able to communicate with other people and bounce ideas off of each other and see where other people's pain points are. You can't innovate in a vacuum.

(Joel Beasley at 00:14:45) So how does your engineering team look right now? Do you local, remote, hybrid?

(Colin Moody at 00:14:50) Yeah. So it's all local, really. Though I'm a big fan of being able to potentially work wherever I am. And so we allow people to work remotely a certain amount of time each month, but we're currently located all in Manhattan. I think it's important for us to, I've been following trends in remote work very, very closely, actually, because I value having interaction in person. I think that it's just a faster communication link. You can learn much more quickly, share what you've learned much more quickly. But at the same time, I do think that it's nice to be able to work anywhere where we have a laptop and Wi-Fi.

(Joel Beasley at 00:15:36) What do you guys use in your stack?

(Colin Moody at 00:15:38) We're a Python shop, I guess, primarily because machine learning, you know, so much of that ecosystem is now based around Python. And on the server side, we use Django. We use Celery, if you're familiar with that, for task orchestration. So it's a well-established stack in the Python ecosystem.

(Joel Beasley at 00:16:03) So how large is the team now that you're currently managing for engineering?

(Colin Moody at 00:16:07) It is eight.

(Joel Beasley at 00:16:08) And what—

(Colin Moody at 00:16:10) Go ahead.

(Joel Beasley at 00:16:10) Yeah. And so, like, oh, you're hiring? Is that what you're gonna say?

(Colin Moody at 00:16:13) Oh, yeah, we are. Yes.

(Joel Beasley at 00:16:14) Oh, nice. Yeah.

(Colin Moody at 00:16:15) We're so—

(Joel Beasley at 00:16:15) People wanna learn more about that, where would they go? Just predata.com?

(Colin Moody at 00:16:19) Yes. If you go to predata.com, we should have a link to—we have a profile on the Muse, and that's where most of our, all of our information should be there.

(Joel Beasley at 00:16:30) Oh, nice. So if people are interested in working at a company that's changing the world, small startup-style company, right, you're growing, lots of energy, solving big problems, then head over to predata.com, check it out?

(Colin Moody at 00:16:42) Exactly. Nice. We should be hiring throughout the rest of the year, through the next year, as far as we can see.

(Joel Beasley at 00:16:48) What type of positions are you looking for?

(Colin Moody at 00:16:50) Yeah. So the engineering team is currently split into roughly three different roles. So we have—we're hiring for machine learning engineers, platform engineers, and front-end engineers. I would also say that given how small we are, these are not extreme. It's not like anything is walled off. Everybody is cross-functional to some extent. You know, and as much as people want to be, really, we want people to do anything that they're capable and productive at doing. So, yeah, we have members of our data science team on the machine learning side who are currently working on delivering, you know, building microservices for delivering new kinds of data to our clients or to our internal analyst team. We have members of our platform team who are currently actually experimenting with new ways of representing the data that we've collected in order to make things like search easier. So it's very cross-functional.

(Joel Beasley at 00:17:54) And so has when it started, was it just you in engineering and now you've built this team of eight? Or how did that evolve?

(Colin Moody at 00:18:02) So yeah, it started with my friend, the co-founder, myself, one other engineer, and we've grown this team up to eight.

(Joel Beasley at 00:18:10) Right. So you're learning a lot then right now.

(Colin Moody at 00:18:13) Yeah. Absolutely. You never—no. Yeah. You never stop learning. Right? We—one of the things that we hire for is we're looking for autodidacts, people who are able to learn new things in a very self-directed way and are constantly driven to read and research and, you know, try out new ideas or implement new papers, for example. Autodidacts, constantly learning.

(Joel Beasley at 00:18:38) So autodidacts means constantly learning?

(Colin Moody at 00:18:41) Self-learning. Yeah.

(Joel Beasley at 00:18:43) Oh, I got a new word. It's great because, like, the show got popular, and so then we got a lot of big-named guests on. So, like, if you listen, like, earlier episodes, a lot of startup companies, medium-sized companies, like, later episodes have been, like, 10,000 employees stuff. And so when I ask them to go back, they're giving me advice from, like, 15, 20 years ago. And so I'm like, let's bring in some high-energy startup companies so we can get that advice, like, hear from them what's going on right now. And that's why, you know, trying to bring as much value with the show as possible, hearing from extremely experienced people that are experienced and just going through it for the first time right now. I love it, man. It's exciting.

(Colin Moody at 00:19:22) Yeah. Yeah. It was fun to listen to—I saw—I've listened to the show, and it's great to—that you have that mix. It's definitely not the case that, you know, at a small company, you know, some of the things that are typical for us are, for example, having the best metrics to, you know, just show what, you know, knowing all the metrics that we need to focus on and how to record those. That can be very hard for a startup. Right? You don't necessarily know what those metrics are or how to record them. And, yeah, building the team, you know, the—I still have, you know, I spend a lot of my time thinking about, you know, what the team is going to look like. And once you get larger, I feel like there's some of those challenges become maybe not the CTO's responsibility. But—

(Joel Beasley at 00:20:07) It's always changing. It's, like, it's energetic. It's fun. It's unique. Like, that's what keeps it so exciting. Right?

(Colin Moody at 00:20:15) Yeah. Yeah. You wear many hats.

(Joel Beasley at 00:20:18) Oh, yeah. Yeah. I actually just Slacked Chloe. Here we go. Brought it in. So we've actually—I wanted to share this with you because we were just talking. You just mentioned it, so I Slacked her real quick. So this is actually every process that happens in my business. It's the standard operating procedure. And what happened was we had wanted to do this for, like, a couple months. You got a year into the business, and it's, like, everything from, like, job descriptions, every role, every responsibility, and it's all laid out. Now some people hold multiple roles. Right? But eventually, it'll scale out. And it's, like, as soon as we get a procedure to where a point where, like, it's somewhat known to have some useful result, then it goes in here and we put, like, everything in it from sales to releasing software, like everything we do. And I'll tell you the biggest difference in the world in my business came when we started printing out these Google Docs and putting them into this binder.

(Colin Moody at 00:21:15) So that's very interesting. I'm also a huge proponent for writing, you know, writing down processes. I mean, that currently, my job is mostly, I'd say, process refinement. Right? It's figuring out what process needs to be put in place and then how to execute on it. So—

(Joel Beasley at 00:21:34) You print it out into a binder?

(Colin Moody at 00:21:36) Well, I haven't gotten that done. Yeah. I need to try that. I mean, we share a lot of information, you know, through Google Drive.

(Joel Beasley at 00:21:43) That's what we do too. But we just started printing it out. And, like, there's something that happens in the universe. Like, when you print it out, right, and then it's there and you hold it, and you can, like, it's just different.

(Colin Moody at 00:21:57) I don't know why because—

(Joel Beasley at 00:21:57) Like, once you print it out and then you go make change, you get into this concept of, like, you want it to, like, look better. It's like doing a report in school. Like, you start to gain pride over it versus it just being, like, another Google Doc you can clone and hit plus. And, like, something happens, at least for me and our team.

(Colin Moody at 00:22:13) I really believe that, you know, until you write these things down, you can't really refine them. Right? In the same way that you can't optimize what isn't measured, I would say that, you know—so I'm a big fan of that saying, but I think it extends to things like processes, even if, you know, even if it isn't necessarily quantitative measurement, just having that, you know, that process recorded makes you think about what pieces are actually essential, what pieces maybe aren't always applicable, you know, what could still use work, and it's a discussion point to have with other people.

(Joel Beasley at 00:22:48) Yeah. So I know you guys are getting started and getting out there. Anyone listening to the show, you think that, like, who would be listening to the show that would have a use to, like, reach out and be a customer of yours?

(Colin Moody at 00:23:01) Yeah. I mean, I think anybody who has some sort of geopolitical exposure should reach out today anyway. I mean, I think in the future, we'll, you know, we'll have the ability to measure trends about things that aren't only geopolitical. And in that world, I think there's a reason for, you know, corporate intelligence customers or, you know, business intelligence customers to reach out as well. But today, the focus is geopolitical. So if you have international supply chains or if you're a trader with global exposure, you know, or if you have infrastructure or resources in parts of the world where you're concerned about security, those are all reasons to reach out to us.

(Joel Beasley at 00:23:49) So have you gotten to the point yet where you're helping, like, you're creating the career ladders at your company for, like, how your engineers can progress?

(Colin Moody at 00:23:57) Yes. I don't think that we're imposing a lot of hierarchy right now. But, yeah, I'm very focused on how career progression is going to look and making sure that people are constantly learning and moving on to, you know, the new projects, the next step in their, you know, their development as engineers. I mean, it's still, it's an eight-person team. And, you know, as we hire here, I don't want us to have an incredibly, you know, complex ornate tree, but I do want to make sure that we have a way to ramp people up from, you know, you're just starting at Predata, you know, here are the ropes, here's how the technology works, to becoming, you know, a big driver of what the product is and what the technology is.

(Joel Beasley at 00:24:45) Yeah. How do you onboard? How do you pair? How do you get people into the company successfully?

(Colin Moody at 00:24:50) Yeah. I mean, how do we train engineers to think about product, to understand, you know, the needs of our end user? How do we expect them to provide effective code review to their peers? How do we work across teams? So, you know, allowing people to be cross-functional. And so, for example, we have members of our front-end team who are working with members of our data science team in order to improve the way that we represent our data in our SaaS platform. That might not be the very first project that a member of our front-end team takes on, but as we, you know, as people become more sophisticated, we can take on more complicated projects, bigger challenges.

(Joel Beasley at 00:25:34) Yeah. And usually, when I talk to companies, usually when engineering hits around 30 people, that's when dynamics start to change because then you won't have a personal relationship with everybody and you're the CTO. And it's like, how do you communicate in a way where you can engage with everybody? And people will start asking for things and wanting, you know, a more mature company at around 30 people in engineering is usually when that happens. Like, right now, it's, like, nice to have. You gotta keep everybody, like, focused, but then, like, the pressure builds. Right?

(Colin Moody at 00:26:03) Yeah. Well, we do have—so we do have separate teams here. I mean, we have the data science team and the rest of engineering are semi-autonomous. They have separate annual goals, for example. But, yeah, 30 actually seems a little late to me. I mean, I feel like by the time you get to eight or 10, it becomes fairly difficult to have a personal, you know, interaction with every member of the team. I mean, it's possible, but it would, you know, definitely be a full-time job. And I do have, you know, there are other members of the team who focus on more of the day-to-day of, you know, how data science is operating, how engineering is operating. And I try—I try to spend my time on either things at a strategic level and a product level or things at a process level. So whatever the day-to-day of those teams is, making sure that those teams can be productive. And then finally, I do spend time working one-on-one with especially, especially newer members of the team in order to onboard them.

(Joel Beasley at 00:27:07) Dude, that's awesome. I'm excited. You're, like, avoiding 80% of the mistakes. Right? Like, that's like, we just by hearing how you structure your day, it's like, whoo. That's like you're—it looks like you're on the right path. That's very exciting.

(Colin Moody at 00:27:21) Yeah. Well, thank you. I hope so. I feel like it's working. I know that I—I still feel like I write more code than I should and less than I'd like to, and that is—that's one of the challenging things for me. I started this job because I love that, but now it's time for me to think more about team processes and making sure that we're hiring effectively and hiring people that are, you know, going to be a good fit for, you know, for our culture and for the task and then making sure that they ultimately are successful.

(Joel Beasley at 00:27:55) Yeah. Like, I think of it like visualization as like an interstate, and you can get off at whatever exit, like, you want. So you could just decide that I only wanna write code and then you hire out someone to do the other stuff and to keep it going, but you can go really, really far.

(Colin Moody at 00:28:11) Yeah. I mean, I try to—when I code, I try to pick things that are either going to be enablers for the rest of the team. So for example, when I code these days, a lot of the time it's tests, which might not even be the most fun thing to do, but it's an enabler for the rest of the team. Or sometimes I'll pick things where, every once in a while, something will come up where I feel like I just have the right expertise to take on a project and do it quickly enough that it makes sense for me to do it. But I try to focus on things that are going to make other people more productive.

(Joel Beasley at 00:28:46) I love it. So what's the culture like there? Like, what do you value?

(Colin Moody at 00:28:51) Yeah. Well, so as I said, we look for autodidacts. We look for explorers, people who are excited to take on new problems because we always, as, you know, as a very small startup, it's inevitable that problems arise that nobody on the team has deep experience solving. And so it's important for us to be able to hire engineers who can—who are excited to take on these problems, you know, venture out with a little bit of guidance, but sometimes not as much as I'd like, and then come back and say, you know, these are the things that I've tried. These are the things still left to try. Like, here, let's exchange notes. So that sort of explorer mentality is important to us. And along with that, I think it's very important that we find engineers who are able to get started quickly with a, you know, sort of low cost of, you know, low price of failure. So whenever I see somebody who is able to, you know, rapidly iterate and they're not afraid—I think there are two mistakes that people make here. They're either afraid to move forward because they're afraid of breaking something, or they're just so gung-ho that they break things indiscriminately. And that balance, the people that are in between, are the best engineers in my mind. They're the people who figure out how to put on all the appropriate, you know, training wheels, so that they, you know, if they do make a mistake, the cost is very low. And then once they have that certainty that the cost of failure is low, they just—they can move very, very quickly.

(Joel Beasley at 00:30:27) I love it. That's like people listening. Right? They hear that. They're like, that's the culture. That's the type of company I wanna be a part of. And then you can attract your tribe that way.

(Colin Moody at 00:30:36) I will also say that as a geopolitics analysis company, everybody on the team is interested in knowing more about what is happening in the world. So we have lots of communication in Slack channels of like, "Oh, here's interesting news coming out of Hong Kong," or things happening in Venezuela, things happening in the EU or Brexit. So we're looking for—we value engineers who ultimately care about not just technology, but again, what is happening in the world. How do we identify real trends, real information, and derive insight that we can get to customers as quickly as possible?

(Joel Beasley at 00:31:23) Boom. I love it. You guys are all passion. You're all driven towards this goal.

(Colin Moody at 00:31:27) Yeah, nice.

(Joel Beasley at 00:31:28) So as we start to wrap up, what are you most excited about today? Like, what's the thing that you're really pumped up about today?

(Colin Moody at 00:31:35) In general, I mean, I guess I'd still say machine learning. I think that—I mean, so we're a machine learning company, but I think that there are so many different opportunities in ML that are going to have major impacts on what the world looks like, and many of them potentially in the relatively short term. Sometimes, though, I mean, I'm excited about it. I'm also nervous.

(Colin Moody at 00:32:00) I mean, I think that there's a lot of opportunity for abuse with many of the advances that I'm seeing. For example, we've been playing around a lot with a model called Grover, which does text generation, basically. And you can use this to make amazing fake news, and potentially very difficult to identify fake news. And so it's very exciting and fun to see, you know, to use Grover as a chatbot, but it also does make me nervous.

(Colin Moody at 00:32:32) So I know that the world is going to change. There's no—we can't put, we can't close Pandora's box here. The technology exists. We're going to have to contend with it. I believe it can make the world a better place, but we need to be on our guard.

(Colin Moody at 00:32:48) What I worry about is that ML will have some sort of Chernobyl moment where with all of the innovation going on, some angle has been missed or some ethical consideration hasn't been taken into account, and then there's a major disruption that causes regulation to come into place that makes it hard to do this work in the States. And I mean, I want the innovation to happen here, but I think that we need to make sure that we don't cause a major disruption that would affect the ability to do that research in the States as opposed to somewhere else in the world.

(Joel Beasley at 00:33:33) So I saw you had an office in Washington, right?

(Colin Moody at 00:33:37) Yes.

(Joel Beasley at 00:33:37) And that's where my sister lives. But also that's where one of our guests was, Anish. He was the first CTO of the United States of America.

(Colin Moody at 00:33:45) Okay.

(Joel Beasley at 00:33:46) And now he's doing something with analytics. But he had mentioned to me—I would listen to that episode with Anish Chopra because he mentioned to me this organization, this group that gets together that helps decide future technologies for the United States, and it's a way for you to get involved at a very high level with the policy and technology. I don't know much about it, but he was talking about it, and it was super interesting.

(Colin Moody at 00:34:15) Yeah. That sounds very interesting. It's something where I think we operate in the right space to add something to the conversation.

(Joel Beasley at 00:34:25) Right. The best way to predict the future is to create it, right? So if you guys are like, "We want to make sure policy heads in the right direction for this," go get involved in one of the groups.

(Colin Moody at 00:34:34) Yeah. I also believe that the best way to predict the future is to understand the present. And maybe that's the same with innovation, right? The future is already here. It's just not evenly distributed. So yeah, being able to understand what's happening today and responding as quickly as possible to that information.

(Joel Beasley at 00:34:53) I understand what's happening today. Predata is killing it. I love it. Any other ways we can help today?

(Colin Moody at 00:35:00) We have a weekly newsletter.

(Joel Beasley at 00:35:02) Oh, cool.

(Colin Moody at 00:35:02) Which can also be found on our website. So if anybody's listening who's interested in what machine learning can do to identify geopolitical trends, you should check out that newsletter.

(Joel Beasley at 00:35:15) Very cool. Geopolitical Machine Learning Trends. Awesome. Have a fantastic day, Colin. I'll have Jake and Chloe loop back with you.

(Colin Moody at 00:35:22) This is fun, Joel. Thank you so much.

(Joel Beasley at 00:35:24) Thank you. Bye.