WEBVTT NOTE This file was generated by Descript 00:00:02.040 --> 00:00:04.140 Mike Bifulco: Hello and welcome back to APIs you won't hate. 00:00:04.230 --> 00:00:05.280 My name is Mike Biko. 00:00:05.280 --> 00:00:09.120 And today I'm sitting down for a conversation with a friend of mine from 00:00:09.150 --> 00:00:12.534 actually last summer, which I guess we'll get into in, in a little while here. 00:00:12.534 --> 00:00:16.415 But talking a little bit about Sky Verne with Ton Singh from 00:00:16.645 --> 00:00:17.575 the founding team of Sky Verson. 00:00:18.055 --> 00:00:18.595 How are you today? 00:00:19.435 --> 00:00:19.915 Track 1: I'm good. 00:00:19.915 --> 00:00:20.365 I'm good. 00:00:20.665 --> 00:00:21.925 You know, it's bright and early here. 00:00:23.245 --> 00:00:24.085 Mike Bifulco: Yeah, that's it. 00:00:24.115 --> 00:00:25.495 Well, are are you on the West Coast? 00:00:25.840 --> 00:00:27.100 Track 1: I'm actually in the east coast here. 00:00:27.160 --> 00:00:27.490 Yeah. 00:00:27.550 --> 00:00:30.485 But it's I live in Canada and we're a nice, like sunny day today, so 00:00:30.505 --> 00:00:31.645 I'm pretty, pretty excited about 00:00:31.705 --> 00:00:32.394 Mike Bifulco: oh, nice. 00:00:32.995 --> 00:00:33.415 Very cool. 00:00:33.415 --> 00:00:36.445 Yeah, I'm a fellow East coaster, so at least we're on a similar time zone. 00:00:36.445 --> 00:00:38.605 I was worried I made you get up at the crack of dawn for this. 00:00:39.015 --> 00:00:39.345 Yeah. 00:00:39.345 --> 00:00:39.585 Cool. 00:00:39.590 --> 00:00:39.885 Thank you. 00:00:39.885 --> 00:00:41.085 So why don't we start here. 00:00:41.090 --> 00:00:43.644 Tell me a little bit about about yourself and about Sky Verne. 00:00:43.644 --> 00:00:44.785 So let, let's start with Sky Verne. 00:00:44.785 --> 00:00:46.315 What's the elevator pitch for what you're building? 00:00:46.930 --> 00:00:51.030 Track 1: So Sky is a a browser based automation tool and we 00:00:51.030 --> 00:00:53.770 use AI to help like companies automate workflows in the browser. 00:00:54.375 --> 00:00:57.105 Specifically what we do is we help companies automate, like really boring 00:00:57.105 --> 00:01:00.740 things like form filling particularly interacting with government websites 00:01:00.740 --> 00:01:02.600 or generating insurance quotes online. 00:01:02.990 --> 00:01:06.410 So really targeting like the back office companies, things that people don't even 00:01:06.410 --> 00:01:10.400 know happen in a company, which is kind of why we get, you know, excited about it. 00:01:12.540 --> 00:01:15.530 Mike Bifulco: Yeah, the classic startup of of making a boring 00:01:15.530 --> 00:01:16.990 problem into a business. 00:01:17.520 --> 00:01:17.729 Track 1: that's 00:01:17.740 --> 00:01:18.009 Mike Bifulco: cool. 00:01:18.009 --> 00:01:19.929 So I definitely wanna talk about that. 00:01:19.960 --> 00:01:23.304 Want, want to dig into it and learn more about how you got there and how 00:01:23.365 --> 00:01:27.145 like, like many of us, you fixated on an exciting, boring problem. 00:01:27.425 --> 00:01:29.224 But before we do that, tell me a little bit about yourself. 00:01:29.224 --> 00:01:31.354 What was your background sort of before you. 00:01:31.829 --> 00:01:34.499 Started building Sky Verne, what did your career kinda look like? 00:01:34.499 --> 00:01:37.810 And, and yeah, I don't know if, hit me with the the exciting, interesting points. 00:01:38.259 --> 00:01:38.500 Track 1: Yeah. 00:01:38.500 --> 00:01:40.669 So I I was I've been programming for a long time. 00:01:40.699 --> 00:01:42.679 I started programming when I was like a, a kid. 00:01:42.799 --> 00:01:46.189 My dad tried, actually taught me c plus plus when I was in grade seven because 00:01:46.189 --> 00:01:47.509 he was like, this is gonna be the future. 00:01:47.779 --> 00:01:48.499 And I was like, okay, sure. 00:01:48.499 --> 00:01:48.799 Whatever. 00:01:48.804 --> 00:01:49.219 I don't care. 00:01:49.549 --> 00:01:49.729 Yeah. 00:01:49.729 --> 00:01:51.559 And we made like some really dumb programs back then. 00:01:51.559 --> 00:01:54.379 And then when I was in high school, I was really into this game called 00:01:54.379 --> 00:01:58.909 Scape, which many, I'm sure many of the listeners are into or into in the past. 00:01:59.474 --> 00:02:03.524 And I think I really hit my stride because I started using, I started 00:02:03.524 --> 00:02:07.994 writing bots for Inscape, using some like client libraries that were exposed. 00:02:08.474 --> 00:02:11.024 And that really, like, I think some of the source codes still 00:02:11.029 --> 00:02:12.495 open, open source on my GitHub. 00:02:12.495 --> 00:02:14.654 I don't, I don't think it runs anymore, but it's, it's still there. 00:02:14.954 --> 00:02:18.464 And I always think back to it is like a really fond memory of when like, I really 00:02:18.464 --> 00:02:22.034 like embrace programming and I always loved like solving problems with it. 00:02:22.034 --> 00:02:22.815 And, um. 00:02:23.394 --> 00:02:26.154 Coding in general, and it just happened that it could also turn it 00:02:26.154 --> 00:02:28.194 into a career, which is kind of lucky. 00:02:28.194 --> 00:02:29.515 Lucky happenstance, right? 00:02:29.934 --> 00:02:34.274 And so I went through university here in Canada and kind of just 00:02:34.279 --> 00:02:36.734 fell into programming as a career. 00:02:36.835 --> 00:02:38.815 And spent a few years working at a few companies. 00:02:38.845 --> 00:02:42.894 And my career really, I would say, took a shape when I started 00:02:42.894 --> 00:02:44.005 working at this company called Fair. 00:02:44.649 --> 00:02:48.159 Here, here in Canada and I got to own their machine learning platform there. 00:02:48.159 --> 00:02:50.630 I got to build it up from ground up having no machine learning background 00:02:50.630 --> 00:02:54.810 at all at the time, and it really made me fall in love with what turned 00:02:54.810 --> 00:02:56.580 out to be the next wave of computing. 00:02:56.585 --> 00:02:58.650 I didn't, I didn't know that back then either, but it turned out to 00:02:58.650 --> 00:03:01.720 be the next wave of computing and really helped shape my career. 00:03:02.515 --> 00:03:05.125 So I got the opportunity to, to build their search and discovery system 00:03:05.125 --> 00:03:08.995 from ground up and make, you know, millions of dollars of impact there. 00:03:08.995 --> 00:03:12.265 And I took that learning to another company afterwards, called goPuff 00:03:12.270 --> 00:03:13.135 and did the same thing there. 00:03:13.135 --> 00:03:14.905 And after that decided to start my own company. 00:03:14.905 --> 00:03:15.780 And here, here we're today, 00:03:17.635 --> 00:03:17.995 Mike Bifulco: Got it. 00:03:18.055 --> 00:03:18.325 Wow. 00:03:18.325 --> 00:03:18.565 Yeah. 00:03:18.565 --> 00:03:21.355 So it sounds like based on that timeline, you were probably building 00:03:21.355 --> 00:03:25.765 ml things like before people were even breathing the word GPT into the air. 00:03:25.765 --> 00:03:26.005 Right. 00:03:26.010 --> 00:03:27.475 How, how long ago did you start working 00:03:27.475 --> 00:03:28.165 with ML stuff? 00:03:29.447 --> 00:03:29.563 Track 1: I. 00:03:29.668 --> 00:03:32.068 It was like, it was around the time that GPT two was out, 00:03:32.218 --> 00:03:33.688 but GP three wasn't out yet. 00:03:33.688 --> 00:03:35.548 And so people didn't really know what it was at the time. 00:03:35.608 --> 00:03:35.938 Yeah. 00:03:36.748 --> 00:03:40.078 And ML was still, you know, like more, less neural networks 00:03:40.078 --> 00:03:41.638 and more like other algorithms. 00:03:41.688 --> 00:03:44.358 Today, you know, people don't really like the other ones that much anymore. 00:03:44.388 --> 00:03:45.918 So yeah, things have changed quite a bit. 00:03:45.948 --> 00:03:46.398 Mike Bifulco: Yeah. 00:03:47.178 --> 00:03:47.508 Cool. 00:03:47.508 --> 00:03:47.808 Yeah. 00:03:47.838 --> 00:03:48.768 Oh, that's really interesting. 00:03:49.068 --> 00:03:52.548 So you built an ML platforms at, at two companies and then you kind of 00:03:52.553 --> 00:03:55.588 made the leap into something a little more entrepreneurial and, and doing 00:03:55.588 --> 00:03:56.848 it for yourself from the sounds of 00:03:56.848 --> 00:03:57.028 it. 00:03:57.028 --> 00:04:01.198 And so what, what was, what's Sky Run's story like what, what was the first 00:04:01.448 --> 00:04:02.558 I don't know, attempt at a product? 00:04:02.558 --> 00:04:03.968 How did you decide that you were gonna do 00:04:04.028 --> 00:04:06.968 Track 1: So Sky's actually our third pivot, believe it or not. 00:04:06.968 --> 00:04:09.278 So we've been building startups for a while. 00:04:09.488 --> 00:04:13.118 My buddy, my buddy, my co-founder Shu and I, we decided to start a company 00:04:13.118 --> 00:04:14.408 together on like one random trip. 00:04:14.408 --> 00:04:15.863 We're like going surfing on the car drive. 00:04:15.863 --> 00:04:18.608 We're like, Hey, we should start a company and, you know, like, any good stories so 00:04:18.788 --> 00:04:20.648 that, that we ended up materializing that. 00:04:21.193 --> 00:04:24.433 And the first product we built was actually a tool to help onboard software 00:04:24.433 --> 00:04:27.883 engineers because I had recently become a manager at a company, and I went through 00:04:27.883 --> 00:04:29.203 the pain that was poor onboarding. 00:04:29.623 --> 00:04:31.963 I later learned that that's a very common target ideas for 00:04:31.993 --> 00:04:33.253 engineers to build work on. 00:04:33.793 --> 00:04:35.473 But we went through the process of building a startup. 00:04:35.478 --> 00:04:37.183 We made every single mistake you could imagine. 00:04:37.388 --> 00:04:37.508 I. 00:04:37.973 --> 00:04:41.933 I think Y Combinator has been very helpful in our career, in our career because 00:04:41.933 --> 00:04:45.203 they put out a lot of content online talking about how to build startups, and 00:04:45.263 --> 00:04:48.143 we later watched all those videos and they recanted all the mistakes we'd made. 00:04:48.413 --> 00:04:49.613 We'd already made up to that point. 00:04:49.618 --> 00:04:51.803 So we built a product for us without talking to any users. 00:04:52.203 --> 00:04:54.243 We're like, we know exactly what people need to solve this problem. 00:04:54.813 --> 00:04:57.993 So we built a solution and then we started trying to find people who'd buy it and 00:04:57.993 --> 00:05:01.563 turns like nobody wanted to buy it because we were not really solving a problem the 00:05:01.563 --> 00:05:02.853 way that people were having the problem. 00:05:03.963 --> 00:05:08.428 And it was, it was not obvious at the, at the outset that that was that, that, 00:05:08.428 --> 00:05:09.478 that that's where we're going down. 00:05:09.538 --> 00:05:11.998 But I'm, I'm still glad we did it because it taught us, you 00:05:12.003 --> 00:05:13.198 know, how to ship code fast. 00:05:13.198 --> 00:05:14.278 We're working part-time. 00:05:14.278 --> 00:05:16.018 Like, you know, we had, we had our full-time jobs. 00:05:16.018 --> 00:05:18.208 We were just coding on the side, trying to get a product out. 00:05:18.208 --> 00:05:21.108 And all the lessons we learned after we built the product, like how to 00:05:21.108 --> 00:05:23.788 do sales, how to talk to people, how to make sure that the problem 00:05:23.788 --> 00:05:26.098 we're solving is aligned with the problem people are actually having. 00:05:26.273 --> 00:05:27.958 Were, were things that we wouldn't have learned had we 00:05:27.958 --> 00:05:29.008 not gone through that journey. 00:05:29.968 --> 00:05:30.958 Mike Bifulco: Yeah, definitely. 00:05:31.198 --> 00:05:31.738 I would love to 00:05:31.738 --> 00:05:35.708 see, I'm sure someone somewhere has a chart on correlation between 00:05:35.708 --> 00:05:39.288 startup founders and having had a, a, a failed experience in the 00:05:39.288 --> 00:05:41.508 past, or at least something that didn't go according to plan. 00:05:41.848 --> 00:05:44.278 I'm definitely one of those, the, the first couple of attempts at 00:05:44.283 --> 00:05:47.948 building a company for me were not not, not successful experiences in 00:05:47.948 --> 00:05:50.888 the traditional sense, but definitely in the sense that I took away lots 00:05:50.888 --> 00:05:51.908 of learnings from it and like. 00:05:52.238 --> 00:05:55.538 Often harken back to the things I did poorly in the past as 00:05:55.538 --> 00:05:56.678 I'm making decisions now. 00:05:56.948 --> 00:05:57.218 Yeah. 00:05:57.248 --> 00:05:57.518 Okay. 00:05:57.518 --> 00:05:59.708 So, so you said you're in your third pivot at this point? 00:05:59.708 --> 00:06:00.038 Yeah. 00:06:00.338 --> 00:06:00.938 Track 1: That's right. 00:06:00.998 --> 00:06:02.108 Yeah, that's right. 00:06:02.108 --> 00:06:06.218 So after that, we pivoted into something that was much more in our domain, which 00:06:06.218 --> 00:06:09.368 was since I've helped two companies build their machine learning platform for, 00:06:09.368 --> 00:06:12.338 particularly for search and discovery experiences, to help like marketplaces 00:06:12.343 --> 00:06:16.723 increase their revenue I had a, we basically found a company that wanted, 00:06:16.723 --> 00:06:18.313 was interested in buying that product. 00:06:18.968 --> 00:06:22.778 And so we pivoted so we could build it and sell it to that company. 00:06:22.978 --> 00:06:24.538 And we got a commitment very early. 00:06:24.538 --> 00:06:27.418 We, we, we actually changed how we did software development. 00:06:27.418 --> 00:06:27.838 In this case. 00:06:27.843 --> 00:06:28.828 We did sales first. 00:06:29.268 --> 00:06:30.888 Made sure the shape of the product made sense. 00:06:30.888 --> 00:06:34.128 We were a lot of documents to align with the people who wanted to buy it. 00:06:34.218 --> 00:06:36.248 Talked to many other companies along the way. 00:06:37.163 --> 00:06:38.153 And then we started building the product. 00:06:38.543 --> 00:06:41.953 And so we, we did a hard pivot at that point because we found that there was 00:06:41.953 --> 00:06:45.253 actually, this, this problem I had solved at two companies was actually a 00:06:45.253 --> 00:06:47.083 problem that many other companies had. 00:06:47.693 --> 00:06:50.423 Started exploring the general space of like search and discovery. 00:06:50.673 --> 00:06:53.043 Went down a few lanes that other companies had also gone down. 00:06:53.043 --> 00:06:58.143 So one very common idea that people who have worked in, in this field go into is 00:06:58.143 --> 00:07:00.003 they try and build search as a service. 00:07:00.033 --> 00:07:04.413 Like they build like a plugin for Shopify where people can plug, put in a search 00:07:04.413 --> 00:07:06.153 engine and make more money that way. 00:07:06.453 --> 00:07:08.973 We went down that path, we went down doing the same thing for recommendations, 00:07:08.973 --> 00:07:12.133 and it turns out both those ideas are very tough to execute on because 00:07:12.133 --> 00:07:14.653 many people have tried before and the value of the product isn't there. 00:07:15.088 --> 00:07:19.228 Yet isn't like the shops are too early to realize the value of the product. 00:07:19.228 --> 00:07:22.318 And so this, this is I think another area of possible tar ideas. 00:07:22.318 --> 00:07:26.398 It's unclear whether it's actually a tar idea today, but a tar idea being an 00:07:26.398 --> 00:07:29.968 idea that's seems attractive from the outset, but it's not has lots of people 00:07:29.968 --> 00:07:31.168 who've tried it before and all failed. 00:07:31.618 --> 00:07:34.468 And so we talked to a bunch of founders who had tried it and had failed with 00:07:34.468 --> 00:07:36.088 those and kind of had had some learnings. 00:07:36.448 --> 00:07:39.658 So we went down this path of building a different platform. 00:07:39.898 --> 00:07:40.198 Yeah. 00:07:40.959 --> 00:07:43.349 Mike Bifulco: so why don't you tell me a little bit about the first version 00:07:43.349 --> 00:07:46.524 of what has now become Sky Verne and how you decided on that pivot. 00:07:47.214 --> 00:07:47.604 Track 1: Yeah. 00:07:47.664 --> 00:07:52.614 So we were just to quick, quickly finish the loop on Vern, we built a 00:07:52.614 --> 00:07:56.104 product and what we learned was that we actually learned during sales. 00:07:56.104 --> 00:07:57.844 We got about three commitments from companies and. 00:07:58.789 --> 00:08:00.919 We learned that the companies were excited about our product. 00:08:01.129 --> 00:08:04.279 All had a division between their data science team and engineering teams, 00:08:04.399 --> 00:08:05.749 which made sales very challenging. 00:08:06.239 --> 00:08:08.099 Particularly if a company had that division. 00:08:08.099 --> 00:08:09.989 They were really excited about our product and if they didn't have 00:08:09.989 --> 00:08:12.629 that division, they were not excited about our product at all, which made 00:08:12.629 --> 00:08:14.039 selling the product very challenging. 00:08:14.309 --> 00:08:16.409 And so we again, decided to pivot the third time. 00:08:16.409 --> 00:08:17.549 This one was way more deliberate. 00:08:17.549 --> 00:08:18.149 We had revenue. 00:08:18.149 --> 00:08:22.289 We were like, had some some traction, but we decided to pivot again and 00:08:22.289 --> 00:08:25.679 we wanted to build something with AI agents and we didn't really understand. 00:08:26.294 --> 00:08:29.144 We knew from a technical perspective, okay, we're like, this is cool stuff. 00:08:29.144 --> 00:08:31.304 We wanted to build cool stuff with this, you know, as engineers do. 00:08:31.784 --> 00:08:33.974 But we didn't really understand how users wanna use this product. 00:08:33.974 --> 00:08:37.364 So what we started doing was reaching out to a lot of people to see how 00:08:37.694 --> 00:08:39.254 people would wanna use AI agents. 00:08:39.614 --> 00:08:42.494 And the thing that kept coming up over and over again was that people were 00:08:42.494 --> 00:08:45.914 really excited about the possibility of automating, like basically controlling 00:08:45.914 --> 00:08:48.024 the computer to do something in a company. 00:08:48.174 --> 00:08:48.384 Okay. 00:08:48.384 --> 00:08:50.304 And that something was a big, big question mark. 00:08:51.704 --> 00:08:53.889 So we talked, talked to a few companies that actually building products in 00:08:53.889 --> 00:08:56.079 the space, but we also talked to a bunch of people who were very excited 00:08:56.079 --> 00:08:58.509 about, like excited about this by reaching out to random people on the 00:08:58.509 --> 00:09:01.299 internet, like looking at forums, looking at Discords and so on. 00:09:01.899 --> 00:09:05.469 And we found a few companies that were really, really excited to automate 00:09:06.159 --> 00:09:07.059 workflows in their background. 00:09:07.149 --> 00:09:09.759 And the first one we ended up working with was a company who wanted to automate 00:09:09.879 --> 00:09:12.219 insurance code generation using AI agents. 00:09:12.579 --> 00:09:13.989 You can imagine that's really boring. 00:09:14.049 --> 00:09:16.629 But the product they, they sell is, they're an insurance reseller. 00:09:16.629 --> 00:09:19.619 You go to their website, you fill out their workflow and then they basically 00:09:19.619 --> 00:09:23.999 fan out to a bunch of insurance providers like State Farm, Geico, Allstate, and 00:09:23.999 --> 00:09:26.279 generate quotes for you and that stuff. 00:09:26.399 --> 00:09:30.119 If anybody, if any of your listeners have experience building scrapers 00:09:30.124 --> 00:09:33.869 before, that stuff breaks all the time because the layout changes a little bit. 00:09:33.869 --> 00:09:36.569 They have like all these nuances, and so the question was, could 00:09:36.569 --> 00:09:38.069 you make that better using ai? 00:09:38.909 --> 00:09:39.899 Could you make that better? 00:09:39.899 --> 00:09:44.943 Using computer vision and some form of like GPT-4 S things. 00:09:45.153 --> 00:09:47.523 And so that's, that's the, we took, tried to take a stab at that. 00:09:47.523 --> 00:09:50.523 So at first we asked, the few people we ran into we're like, Hey, what have 00:09:50.523 --> 00:09:51.723 you tried to solve this problem today? 00:09:51.993 --> 00:09:54.453 They're like, oh, we talked to these companies, nobody solved it for us yet. 00:09:55.023 --> 00:09:56.073 And we're like, okay, let's try it out. 00:09:56.823 --> 00:09:59.463 And that's kind of how um, was born. 00:09:59.733 --> 00:10:01.923 So we ended up working with a handful of these companies. 00:10:02.508 --> 00:10:04.308 Just to begin with, just to get the shape of our product. 00:10:04.308 --> 00:10:04.548 Right. 00:10:04.548 --> 00:10:06.558 And then now we're starting to scale out to more and more companies. 00:10:06.908 --> 00:10:09.818 Mike Bifulco: I would imagine that many people listening to this podcast have 00:10:09.818 --> 00:10:14.288 tried to write a screen scrapper in the past, either for like a assignment 00:10:14.288 --> 00:10:17.378 for a, you know, a course or a code school or whatever, or for the, you 00:10:17.378 --> 00:10:20.228 know, some real business reason that they wanted to have something automated. 00:10:20.228 --> 00:10:23.788 And I think having done it once is enough to tell you how painful 00:10:23.788 --> 00:10:26.368 it can be when things break, and how hard it is to build something 00:10:26.468 --> 00:10:28.508 Track 1: And it's funny, I've done it too, you know, like when I was 00:10:28.508 --> 00:10:32.618 kid, I was writing, it's the same idea except now using way better 00:10:32.618 --> 00:10:34.208 technology, you know, it's, it's great. 00:10:36.173 --> 00:10:37.343 Mike Bifulco: Yeah, exactly. 00:10:37.343 --> 00:10:41.603 And so you, you pivoted to, to this sort of third place where you're 00:10:41.603 --> 00:10:43.763 starting to, to build for automation. 00:10:43.983 --> 00:10:46.803 If I remember correctly, you had applied to and gotten into yc. 00:10:46.803 --> 00:10:47.793 So this is how we met. 00:10:47.798 --> 00:10:49.323 I sort of alluded to this in the intro. 00:10:49.533 --> 00:10:52.113 We both went through Y Combinator's summer program last year. 00:10:52.113 --> 00:10:53.703 We were in the the S 23 batch. 00:10:53.993 --> 00:10:56.543 And at the time you were y Vern, right? 00:10:56.543 --> 00:10:57.383 So you were in, in, 00:10:57.383 --> 00:11:01.193 your second pivot building marketplace tooling and, and pivoted there. 00:11:01.193 --> 00:11:04.133 Can you tell me a little bit about what the YY Combinator experience was like 00:11:04.133 --> 00:11:06.083 for you and, and some of the things you might have gotten outta that? 00:11:06.828 --> 00:11:09.648 Track 1: Yeah, so the Wes experience is pretty cool. 00:11:09.648 --> 00:11:13.948 So I, I, I feel like Wesley's community. 00:11:14.308 --> 00:11:16.438 It was just like a bunch of really technical people who 00:11:16.438 --> 00:11:17.638 want to build something cool. 00:11:17.638 --> 00:11:21.178 And I think that I, I don't really think I real recognize that, but when we 00:11:21.178 --> 00:11:24.628 applied, we applied because the videos are very, very helpful to us in our first 00:11:25.078 --> 00:11:27.358 pivot journey into our second pivot. 00:11:27.668 --> 00:11:30.248 But the experience itself is very cool because you're in this like 00:11:30.938 --> 00:11:34.898 concentrated group of highly technical people working on really interesting 00:11:34.898 --> 00:11:37.778 problems and that that is like nothing short of inspirational. 00:11:38.573 --> 00:11:41.513 Because anytime you meet up with somebody, like even even me and you, right? 00:11:41.513 --> 00:11:44.723 Every, when we had our convers, when we chatted, I would just be inspired 00:11:44.723 --> 00:11:48.413 to work hard and solve problems and you know, you could talk with people 00:11:48.418 --> 00:11:52.043 at a higher level and I think that's something that YC unlocked for everybody. 00:11:53.243 --> 00:11:55.733 In terms of my experience specifically, my experience was like a little bit 00:11:55.733 --> 00:12:00.053 unique because I actually had a kid during the YC batch, which I learned 00:12:00.053 --> 00:12:01.193 I wasn't the only one first of all. 00:12:01.373 --> 00:12:05.123 But I also don't recommend if anybody's listening and thinking about that. 00:12:05.128 --> 00:12:06.623 I don't recommend going through 00:12:06.788 --> 00:12:07.778 Mike Bifulco: I can't imagine. 00:12:07.868 --> 00:12:08.588 Yeah. 00:12:08.618 --> 00:12:09.038 Track 1: Yeah. 00:12:09.428 --> 00:12:09.668 But 00:12:10.248 --> 00:12:11.388 Mike Bifulco: So that must have complicated thing. 00:12:11.388 --> 00:12:13.038 So your kid was born during the batch. 00:12:13.218 --> 00:12:14.718 Track 1: it was, yeah, she was born during the batch. 00:12:14.718 --> 00:12:16.638 She was born actually exactly halfway through the batch. 00:12:17.028 --> 00:12:17.388 Mike Bifulco: Wow. 00:12:17.388 --> 00:12:18.228 Well, first of all, 00:12:18.228 --> 00:12:19.248 congratulations. 00:12:20.118 --> 00:12:21.708 Track 1: I didn't get, I didn't necessarily get the full experience 00:12:21.708 --> 00:12:24.798 of yc, but it was also kind of extra focusing 'cause we really 00:12:24.798 --> 00:12:25.698 had time for nothing else. 00:12:25.848 --> 00:12:28.548 Like you really had time for nothing beyond startup and, and life. 00:12:28.548 --> 00:12:28.938 And that's it. 00:12:29.480 --> 00:12:31.910 I think, I think the month after she was born, I lost like 10 pounds 00:12:31.940 --> 00:12:34.220 because I was sleeping like three hours a night and just like grinding 00:12:34.220 --> 00:12:37.520 during the day and like, you know, baby taking care of the baby at night. 00:12:38.051 --> 00:12:39.071 Mike Bifulco: What a wild motivator. 00:12:39.071 --> 00:12:40.001 That's really interesting. 00:12:40.031 --> 00:12:43.811 I feel like few people in the world can relate to a desire for things to be 00:12:43.811 --> 00:12:45.781 automated, like parents of a newborn. 00:12:46.141 --> 00:12:49.441 Uh, you know, like if, if everything was just a little bit easier, your life 00:12:49.621 --> 00:12:51.811 Track 1: Well, it's funny because like there, there was a whole class 00:12:51.811 --> 00:12:54.511 of ideas we were exploring that were like born of that pain, right? 00:12:54.511 --> 00:12:57.121 Because I was like, okay, so I'm taking care of baby while programming 00:12:57.121 --> 00:13:00.271 or while doing sales, how can I control my computer with my voice? 00:13:00.361 --> 00:13:03.571 There was like a whole subset of whole subset of like applications we were 00:13:03.571 --> 00:13:07.591 exploring, like voice of text, like, you know, computer automation, all based 00:13:07.621 --> 00:13:10.621 on this huge problem where I was just like, my hands are busy, but I still 00:13:10.621 --> 00:13:12.631 wanted to, to be productive, right? 00:13:12.631 --> 00:13:15.571 And so we ended up not, not going down those paths, but something 00:13:15.571 --> 00:13:17.431 similar I guess is what we landed on. 00:13:18.241 --> 00:13:18.781 Mike Bifulco: Sure. 00:13:18.781 --> 00:13:19.201 Yeah. 00:13:19.231 --> 00:13:19.681 Wow. 00:13:19.921 --> 00:13:21.581 Well it, it's really interesting. 00:13:21.586 --> 00:13:24.701 So you, you definitely had a non-traditional path in yc, right? 00:13:24.706 --> 00:13:27.971 Like probably in the per percent of a percent of people that who've 00:13:27.971 --> 00:13:30.451 ever been through YC have probably had a kid during the batch. 00:13:30.931 --> 00:13:32.311 Track 1: Actually it's more common than you think. 00:13:32.311 --> 00:13:33.211 It's more common than you think. 00:13:33.216 --> 00:13:34.231 That's what I learned during the batch. 00:13:34.231 --> 00:13:36.391 It's when you are in a situation, you find out other people 00:13:36.391 --> 00:13:37.351 are also in that situation. 00:13:37.351 --> 00:13:41.521 And I learned that it actually, maybe a percent of the batch, I would say is 00:13:41.521 --> 00:13:41.941 like that. 00:13:43.321 --> 00:13:44.011 Mike Bifulco: That's fascinating. 00:13:44.011 --> 00:13:47.551 Yeah, I definitely know some, some friends who I met through IC who 00:13:47.831 --> 00:13:51.341 got married before, during, or after the batch, and that's a, a similarly 00:13:51.341 --> 00:13:52.811 like life shifting experience. 00:13:53.121 --> 00:13:56.211 Usually something that's hard to explain to, you know, a, a potential 00:13:56.216 --> 00:13:59.301 partner that like, Hey, I know we're getting married like in a month, but I'm 00:13:59.301 --> 00:14:00.771 have to move to San Francisco now. 00:14:01.491 --> 00:14:02.721 It's a hard thing to deal with. 00:14:03.046 --> 00:14:07.056 like you, I had an uncommon YC experience in as much as my, my company craft work 00:14:07.056 --> 00:14:09.456 is very, very geographically focused. 00:14:09.456 --> 00:14:09.636 Right. 00:14:09.636 --> 00:14:12.186 We are in Charlotte, in North Carolina, and so I spent all 00:14:12.191 --> 00:14:15.256 summer flying back and forth to San Francisco for for meetings because 00:14:15.256 --> 00:14:17.026 our business is so importantly here, 00:14:17.356 --> 00:14:20.236 um, I can definitely relate to like, not fitting the pattern. 00:14:20.236 --> 00:14:20.806 Exactly. 00:14:21.256 --> 00:14:21.526 Okay. 00:14:21.526 --> 00:14:24.906 So, so you made it through yc, your pivot into skiver. 00:14:24.906 --> 00:14:27.716 Then was was also during yc or was that after the batch ended? 00:14:28.286 --> 00:14:29.966 Track 1: It was after, so it was about a month after. 00:14:29.966 --> 00:14:32.066 So what happened during the Bachelors were, were. 00:14:32.621 --> 00:14:36.191 Basically one thing YC focuses you a lot on is like picking one 00:14:36.191 --> 00:14:37.841 metric and growing that metric only. 00:14:38.111 --> 00:14:39.881 And so the metric we we had picked was revenue. 00:14:40.211 --> 00:14:42.971 And so we're like singularly focused on growing, growing that metric. 00:14:42.971 --> 00:14:46.301 And what happened during the batch was we had one customer that was paying us like 00:14:46.301 --> 00:14:48.036 some amount of money and we had actually. 00:14:48.971 --> 00:14:51.851 One commitment from another customer for a bigger amount of money. 00:14:51.851 --> 00:14:52.601 So we're very excited. 00:14:52.601 --> 00:14:55.751 Things are going well, and after the batch, we got a third commitment. 00:14:55.751 --> 00:14:58.121 So things are going amazingly well, right? 00:14:58.121 --> 00:14:59.741 And we're like, yeah, this is gonna be a big product. 00:14:59.741 --> 00:15:00.911 It's gonna be huge, you know, amazing. 00:15:01.101 --> 00:15:05.601 But what happened was right after that, we started looking into our sales data. 00:15:05.606 --> 00:15:07.851 See, okay, so now that we have three commitments, like what is 00:15:07.851 --> 00:15:09.561 repeatable about this process? 00:15:09.831 --> 00:15:11.511 Like what is common in these customers? 00:15:11.511 --> 00:15:14.601 What is un uncommon in the cus common in the customers that rejected us? 00:15:14.606 --> 00:15:17.581 So we can start like selecting earlier, if that makes sense. 00:15:18.136 --> 00:15:20.806 And that was about, like, that was about three weeks after the batch. 00:15:20.806 --> 00:15:25.096 And we, in digging into that data, that's when we realized that commonality where 00:15:25.196 --> 00:15:28.106 the companies that were excited about our product were ones that had like 00:15:28.106 --> 00:15:30.836 their data teams and engineering team structure in like a very specific way 00:15:30.836 --> 00:15:32.156 where they didn't communicate effectively. 00:15:32.486 --> 00:15:35.186 And if you look at the total number of companies like that in the 00:15:35.186 --> 00:15:36.596 world, one, there aren't that many. 00:15:36.596 --> 00:15:40.676 And two, even if there are many, it's, it, it, it becomes a much less exciting 00:15:40.676 --> 00:15:44.216 problem to solve when you're not able to help many companies, you know? 00:15:44.816 --> 00:15:45.716 Be more effective. 00:15:46.046 --> 00:15:46.856 And so that's why we actually 00:15:46.856 --> 00:15:47.246 pivoted. 00:15:47.756 --> 00:15:48.056 Yeah. 00:15:48.386 --> 00:15:48.716 So it was 00:15:48.716 --> 00:15:49.346 after the batch, it was 00:15:49.351 --> 00:15:50.216 about a month after the batch. 00:15:51.101 --> 00:15:53.081 Mike Bifulco: the scale story is, is dramatically different 00:15:53.081 --> 00:15:55.511 if you're targeting, you know, half a dozen companies or less, 00:15:55.841 --> 00:15:58.841 uh, you may still be able to reach an exciting revenue number, but harder. 00:15:59.516 --> 00:15:59.726 Track 1: Yeah. 00:15:59.726 --> 00:16:01.436 And then when you're pivoting, you start asking yourself questions. 00:16:01.436 --> 00:16:01.586 Right? 00:16:01.591 --> 00:16:04.226 You're like, okay, well is there a way, is there a shape of the product that I can 00:16:04.226 --> 00:16:06.446 build that maybe appeals to more people? 00:16:06.626 --> 00:16:06.896 Right. 00:16:06.896 --> 00:16:10.436 Or is there, is there, or is can we work on this problem somehow? 00:16:11.156 --> 00:16:13.556 But I think the question that we always ask ourselves when we're pivoting is, 00:16:13.976 --> 00:16:15.596 do we wanna still go after this market? 00:16:15.986 --> 00:16:17.666 Do we still want to go after this thing? 00:16:17.666 --> 00:16:21.386 If we are thinking right, do we take one step back or do we 00:16:21.386 --> 00:16:22.556 wanna take this full step back? 00:16:22.556 --> 00:16:26.486 I take the full picture and this entire time, you know, we saw the, the evolution 00:16:26.486 --> 00:16:29.426 of language models we're getting more and more excited by it and had a lot of 00:16:29.426 --> 00:16:32.306 experience with it because I'd worked in search and language models have been 00:16:32.306 --> 00:16:33.986 part of search for, for pretty long time. 00:16:34.376 --> 00:16:36.026 I would say transformers was, you know. 00:16:36.716 --> 00:16:38.906 Created to help with natural language processing, which is like 00:16:38.966 --> 00:16:40.346 core to Google search algorithm. 00:16:40.706 --> 00:16:44.666 And so that's why we decided to take a leap forward and 00:16:44.756 --> 00:16:46.376 and do the full pivot instead. 00:16:47.291 --> 00:16:47.831 Mike Bifulco: Yeah. 00:16:47.861 --> 00:16:48.101 Okay. 00:16:48.101 --> 00:16:50.141 And so now we're in the world of Sky Verne and you're 00:16:50.141 --> 00:16:52.281 automating things with AI agents. 00:16:52.331 --> 00:16:55.241 , from what I can tell, the point you're at right now is sort of early on 00:16:55.246 --> 00:16:57.041 in the product building experience. 00:16:57.041 --> 00:17:01.011 And so can you tell me a little bit about what, the current version of HelloWorld 00:17:01.016 --> 00:17:02.571 looks like for someone using Sky Verne? 00:17:03.576 --> 00:17:05.651 Track 1: So we're at, we're at a stage where we're actually manually 00:17:05.681 --> 00:17:08.481 onboarding customers because the user interface is something we're 00:17:08.481 --> 00:17:09.736 still trying to get, like correct. 00:17:10.541 --> 00:17:12.581 We're looking in the process of onboarding. 00:17:12.581 --> 00:17:14.681 We're learning about how people would want to interact with our product. 00:17:14.681 --> 00:17:17.081 But today, sky Ver is an API that you call. 00:17:17.081 --> 00:17:18.191 It's a pure API product. 00:17:18.881 --> 00:17:23.321 And what you pass it is basically a URL you want to go to. 00:17:23.771 --> 00:17:25.631 So we have four customers that are live. 00:17:25.636 --> 00:17:26.951 They all pass us to different URL. 00:17:27.161 --> 00:17:29.681 The insurance reseller passes us something like echo.com. 00:17:30.386 --> 00:17:33.666 Other customers pass their own URLs and and then they pass us 00:17:33.666 --> 00:17:34.866 what we call a navigation goal. 00:17:35.461 --> 00:17:39.961 And that basically tells our Charles Skiver to go do something. 00:17:40.231 --> 00:17:43.561 In the case of generating an insurance quote, it's like, Hey, go, go generate 00:17:43.561 --> 00:17:45.211 an auto insur or auto insurance quote. 00:17:45.661 --> 00:17:46.951 Don't generate a home insurance quote. 00:17:47.431 --> 00:17:49.711 You're done when you have gotten to the quote page. 00:17:50.161 --> 00:17:50.671 And then what? 00:17:50.731 --> 00:17:52.951 What we're able to do is take that instruction and basically 00:17:52.956 --> 00:17:56.731 start@geico.com and just keep going until it hits that goal. 00:17:56.731 --> 00:17:57.481 So it just takes a look. 00:17:57.481 --> 00:18:00.451 Look at every action that's possible on a screen and just decides, 00:18:00.451 --> 00:18:02.431 hey, this is the most likely thing to get us towards this goal. 00:18:03.136 --> 00:18:05.086 Over and over and over again until it gets there. 00:18:05.596 --> 00:18:08.476 So it'll click on auto, it'll, you know, fill in, fill in information. 00:18:09.556 --> 00:18:13.616 And then what we also ask our customers to pass in is basically some information 00:18:13.766 --> 00:18:14.846 that's necessary to complete the goal. 00:18:14.846 --> 00:18:18.236 So in the insurance case, all of your information basically. 00:18:18.241 --> 00:18:21.266 And then we just keep using, mapping that information to whatever's on the 00:18:21.266 --> 00:18:22.586 screen in real time and just like. 00:18:23.321 --> 00:18:24.761 Boom, boom, boom, until it gets to the goal. 00:18:25.361 --> 00:18:27.191 So to go back to your original question, which is where are we 00:18:27.191 --> 00:18:28.121 in the development lifecycle? 00:18:28.121 --> 00:18:31.271 So we have four customers we're live with today, and we're looking, we're, 00:18:31.286 --> 00:18:34.631 we're onboarding about three a month at, I would say, at this, at this point. 00:18:35.188 --> 00:18:37.833 Mike Bifulco: Are there so that's a reasonably small sample set, but 00:18:37.838 --> 00:18:40.893 in terms of the customers you're courting right now or talking to, 00:18:40.893 --> 00:18:42.243 are there commonalities between them? 00:18:43.338 --> 00:18:44.718 Track 1: No, actually, and that's by design. 00:18:45.358 --> 00:18:46.198 That's actually by design. 00:18:46.198 --> 00:18:49.598 So whenever you build a product that has like too many applications, you 00:18:49.598 --> 00:18:53.918 always hit this like decision point where you wanna have you wanna onboard 00:18:53.918 --> 00:18:56.888 people, like you wanna solve one problem well and onboard lots of people 00:18:56.888 --> 00:18:58.418 in that, that are in that vertical. 00:18:58.778 --> 00:19:01.118 Or do you wanna solve a lot of people's problems? 00:19:01.358 --> 00:19:03.878 Kind of all, and we're actually in the second, we're doing 00:19:03.878 --> 00:19:04.898 the second one on purpose. 00:19:05.228 --> 00:19:07.368 And the reason for that is as you can imagine, interacting 00:19:07.368 --> 00:19:08.508 with the web is very messy. 00:19:09.048 --> 00:19:10.908 Every single website is like, designed differently. 00:19:11.148 --> 00:19:14.418 Some, like we talked to this one customer where their workflow in 00:19:14.598 --> 00:19:18.568 involved interacting with this like website in in, in India, like the state, 00:19:18.568 --> 00:19:19.798 state government website in India. 00:19:20.128 --> 00:19:22.678 And halfway through the workflow to refresh the page five times 00:19:22.678 --> 00:19:23.488 to get a button to show up. 00:19:24.358 --> 00:19:26.878 Like, how do you, how do you instruct an agent to do that? 00:19:27.118 --> 00:19:29.023 You gotta like, you know what I mean? 00:19:29.253 --> 00:19:32.458 Like how do you instruct, how do you instruct any AI to know that, 00:19:32.518 --> 00:19:34.258 okay, you gotta to this page, you don't see a button you need to click 00:19:34.258 --> 00:19:35.458 on, you gotta refresh the page. 00:19:35.458 --> 00:19:35.878 Like, wow. 00:19:36.178 --> 00:19:36.508 But 00:19:36.838 --> 00:19:39.988 the reason, the reason we're going horizontal though, is to, to kind 00:19:39.988 --> 00:19:41.248 of learn about cases like that. 00:19:41.548 --> 00:19:44.338 And what we found with our product is every time we onboard a new case, 00:19:44.848 --> 00:19:47.218 we actually end up solving like 10 other cases we didn't know about. 00:19:47.488 --> 00:19:49.258 And so as we onboard more and more websites. 00:19:50.563 --> 00:19:53.263 Our coverage increases, and that's kind of the thing we're going after right now. 00:19:53.623 --> 00:19:56.683 And there'll be a tipping point, whereas coverage will be, you know, 00:19:56.683 --> 00:19:59.893 maybe 50% of the web or 70% of the web, where then we can start 00:19:59.893 --> 00:20:01.303 onboarding customers extremely fast. 00:20:01.963 --> 00:20:03.343 We can open it to everybody. 00:20:03.823 --> 00:20:06.103 They can have a recent, like a moderately good experience that you 00:20:06.103 --> 00:20:08.113 would expect, and that's kind of what we're pushing towards today. 00:20:09.868 --> 00:20:10.828 Mike Bifulco: Oh, that makes a lot of sense. 00:20:10.828 --> 00:20:14.938 I, I think you, you're more likely to get more broad feedback than to focus 00:20:14.938 --> 00:20:18.598 on, you know, just solving problems for the, the insurance adopters or 00:20:18.598 --> 00:20:20.008 whatever it may be in the world. 00:20:20.398 --> 00:20:20.608 Okay. 00:20:20.608 --> 00:20:22.948 So let, let's talk about this from the perspective of, of an API 00:20:22.953 --> 00:20:26.178 developer knowing that our audience for the podcast is generally people 00:20:26.178 --> 00:20:27.498 who are building and designing APIs. 00:20:27.498 --> 00:20:30.708 I'm kind of interested to hear, two, two things from you. 00:20:30.708 --> 00:20:34.678 One is like, what's an interesting use case for for using Sky ver that 00:20:34.678 --> 00:20:36.358 might stand out to API developers? 00:20:36.608 --> 00:20:39.758 And then I, I want to ask essentially the same question, but about building 00:20:39.758 --> 00:20:42.308 Sky ver things you've learned about APIs, but, we'll, we'll do that next. 00:20:42.308 --> 00:20:45.578 Tell me about like what's an interesting use case for an API developer 00:20:45.628 --> 00:20:47.038 that they might use Sky ver for 00:20:47.738 --> 00:20:48.663 Track 1: Yeah, so. 00:20:49.323 --> 00:20:52.893 I would, I'm gonna answer this question from, from the perspective of an API user. 00:20:53.073 --> 00:20:57.343 So I would say an API developer would be looking to, you know, build APIs on 00:20:57.343 --> 00:20:58.693 things that don't already have, have APIs. 00:20:58.693 --> 00:20:59.023 Right? 00:20:59.443 --> 00:21:02.983 And actually, if we were to distill why like Sky Run's goal down to what it's 00:21:02.983 --> 00:21:05.743 actually trying to do, we're trying to solve that exact problem, which is 00:21:05.773 --> 00:21:10.813 how can we build an API for websites that don't have APIs and might never 00:21:10.813 --> 00:21:11.023 have APIs. 00:21:11.863 --> 00:21:12.853 And might never have APIs. 00:21:13.393 --> 00:21:15.823 And then there's two categories of websites that are like that, right? 00:21:15.853 --> 00:21:19.063 One is, maybe, there might be more, but two, two big ones that I think about. 00:21:19.483 --> 00:21:23.083 One is websites that don't want you to have an API. 00:21:23.383 --> 00:21:27.043 And then there's obviously like, like you could, you could put like insurance 00:21:27.043 --> 00:21:28.093 code generation in that, right? 00:21:28.093 --> 00:21:31.138 Geico probably doesn't want to expose an API because even though they 00:21:31.303 --> 00:21:34.633 have to make the algorithm, they, they used to generate quotes public. 00:21:35.353 --> 00:21:37.753 They obfuscate it because they don't want people to replicate it. 00:21:37.753 --> 00:21:37.933 Right. 00:21:37.933 --> 00:21:41.293 So they, so there's a lot of websites that don't want to expose an API, LinkedIn 00:21:41.293 --> 00:21:44.293 might be this, you know, Twitter's recently become this with their, the 00:21:44.298 --> 00:21:46.243 way they charge for APIs and, and so on. 00:21:47.503 --> 00:21:49.993 So we can help companies interact with websites that 00:21:49.993 --> 00:21:51.613 don't want you to have an API. 00:21:51.613 --> 00:21:54.403 Obviously, we don't wanna violate any terms of services or anything like that. 00:21:54.403 --> 00:21:56.743 So we, we make sure that the, the use cases are reasonable. 00:21:57.403 --> 00:22:00.583 And then the, and the second type is the websites that. 00:22:01.813 --> 00:22:04.543 Can't have an API and, and that's like generally I would 00:22:04.543 --> 00:22:07.303 consider government websites that fall in that category, right? 00:22:07.303 --> 00:22:10.063 Like legacy websites that are relatively unmaintained. 00:22:11.083 --> 00:22:13.903 We, we've been talking to like a bunch of companies that are like 00:22:14.003 --> 00:22:19.168 dealing with old school, like supply chain procurement companies, right? 00:22:19.168 --> 00:22:22.118 And they have like these content management systems they were using 00:22:22.748 --> 00:22:25.808 for their purchase orders that were built in like the nineties or the two 00:22:25.813 --> 00:22:27.698 thousands that are unmaintained, but. 00:22:28.313 --> 00:22:31.913 You know, or the definition of a moat where your data is so far in 00:22:31.913 --> 00:22:34.013 it that it's so hard to get it out that nobody's gonna do it, which is 00:22:34.013 --> 00:22:36.803 like the textbook definition of what a good software moat could be like. 00:22:37.163 --> 00:22:37.553 Right? 00:22:38.123 --> 00:22:39.803 So those are websites that would never have an API. 00:22:39.803 --> 00:22:43.733 And so we really are trying to solve the problem of building an a 00:22:43.733 --> 00:22:45.113 p on websites that don't have one. 00:22:45.693 --> 00:22:48.153 So from an API developer's perspective, it's like how do you 00:22:48.453 --> 00:22:52.083 develop an API that is the API to any website that doesn't have one? 00:22:52.088 --> 00:22:53.853 And that's a challenging problem, right? 00:22:54.303 --> 00:22:55.353 That's a really challenging problem. 00:22:55.353 --> 00:22:57.183 And how can you like build an interface that. 00:22:57.798 --> 00:23:01.578 Conforms to the, the peculiarities of every website. 00:23:02.358 --> 00:23:06.228 And what we landed on was actually we just have four fields that kind of, and we use 00:23:06.228 --> 00:23:07.518 language models to do everything else. 00:23:07.918 --> 00:23:10.198 So we have one field that's the URL, where, where do you wanna start? 00:23:10.588 --> 00:23:12.628 Second field, which is what we call navigation goal, which is what 00:23:12.633 --> 00:23:13.828 do you wanna do on the website? 00:23:13.833 --> 00:23:15.328 So what navigation actions do you wanna take? 00:23:15.778 --> 00:23:18.238 And then third is what do you data extraction goal, which 00:23:18.238 --> 00:23:19.228 is what do you wanna extract? 00:23:20.233 --> 00:23:24.593 Then fourth is what we call payload which is what data do I need to do 00:23:24.593 --> 00:23:25.583 everything that you asked me to do? 00:23:26.093 --> 00:23:30.263 And so since, since the last three fields are unstructured, they are 00:23:30.263 --> 00:23:34.523 like, like natural language fields, we can then use language models on 00:23:34.523 --> 00:23:39.743 the backend to map whatever they pass in to things that need to be done. 00:23:40.073 --> 00:23:42.713 And so the API from from the user perspective is actually very simple. 00:23:42.713 --> 00:23:44.153 All of our customers use the same one. 00:23:44.873 --> 00:23:48.623 The things that passed in are totally different, and then we kind of map it back 00:23:48.893 --> 00:23:51.923 and that creates its own challenges, of course, but I'm happy to get into that as 00:23:51.923 --> 00:23:52.193 well. 00:23:52.943 --> 00:23:53.243 Mike Bifulco: yes. 00:23:53.243 --> 00:23:56.243 I think I definitely would like to, this is, this is very juicy and I think 00:23:56.243 --> 00:23:58.823 you've just given me the title for this episode of you're building the 00:23:58.823 --> 00:24:00.683 API for websites that don't have APIs. 00:24:00.683 --> 00:24:03.033 That's like very, very juicy and very good. 00:24:03.083 --> 00:24:03.293 Okay. 00:24:03.293 --> 00:24:05.513 So that, that's a perfect dovetail actually into the next step then. 00:24:05.513 --> 00:24:10.643 So like you, you have a, a wildly simple at least shape of your API, but three 00:24:10.643 --> 00:24:12.903 of the four fields are NLP fields. 00:24:13.053 --> 00:24:13.353 What is 00:24:13.358 --> 00:24:16.533 it like to build something where you're asking someone for a very abstract input 00:24:16.533 --> 00:24:18.003 and trying to map that to a real goal? 00:24:18.008 --> 00:24:19.053 Like how, how do you build around that? 00:24:20.703 --> 00:24:23.703 Track 1: This is something we're actively iterating on because the problem with the, 00:24:23.703 --> 00:24:27.453 the benefit of abstract inputs is that the demos are very, very attractive, right? 00:24:27.453 --> 00:24:28.923 You, You, can make, you can make, 00:24:28.923 --> 00:24:32.073 something work pretty effectively, but how do you make it work a hundred percent. 00:24:33.453 --> 00:24:35.703 And we found that the more precise you are in the goals, it's same 00:24:35.703 --> 00:24:36.753 as when you talk to like chat pd. 00:24:36.753 --> 00:24:40.173 The more precise you are with what you want to get done, the more accurate it 00:24:40.173 --> 00:24:45.003 is that being able to get that done and then you can start abstracting that as 00:24:45.003 --> 00:24:49.053 concepts a way to like more deterministic things like I would say the most of our 00:24:49.413 --> 00:24:53.373 time is not spent iterating on how we understand the logic coming in from, 00:24:53.403 --> 00:24:55.143 from our customers, but it's, it's. 00:24:55.908 --> 00:24:59.568 How do we make sure everything that's happening with this thing that is 00:24:59.568 --> 00:25:04.548 not deterministic, every binding to it is deterministic, and how can we 00:25:04.548 --> 00:25:05.868 conform it to be more deterministic? 00:25:05.868 --> 00:25:09.198 So all of our time is actually spent in like massaging, I guess the, 00:25:09.418 --> 00:25:13.483 the, the HTML, the, the JavaScript that you see on a, on a browser 00:25:13.783 --> 00:25:14.743 that, that runs in a browser. 00:25:15.298 --> 00:25:18.508 To conform to what the language model can interact with and how it 00:25:18.508 --> 00:25:19.948 relates all the way back to a goal. 00:25:20.218 --> 00:25:22.678 And so we spend a lot of time helping our customers, one, create the prompts 00:25:22.678 --> 00:25:27.118 and create the payloads they wanna send to us, but two, finessing the 00:25:27.178 --> 00:25:29.158 language model to output a specific way. 00:25:29.398 --> 00:25:31.318 And actually, if you look at how like GitHub copilot is 00:25:31.323 --> 00:25:32.128 made, they do the same thing. 00:25:32.308 --> 00:25:35.848 Like most of their remote, I would say is not really the language model. 00:25:35.908 --> 00:25:39.298 I mean, of course they benefit from improvements in it, but it's how do you 00:25:39.808 --> 00:25:44.068 force the language model and, and have all these like edge guards in place. 00:25:44.743 --> 00:25:45.943 So that the output is productive, 00:25:46.723 --> 00:25:47.893 the output is very productive. 00:25:47.893 --> 00:25:48.103 Right. 00:25:48.108 --> 00:25:49.423 It's like extraordinarily productive. 00:25:50.143 --> 00:25:50.473 Yeah. 00:25:51.868 --> 00:25:54.988 Mike Bifulco: I, I love this sort of problem for, for so many reasons that 00:25:54.988 --> 00:25:59.108 like, this is, to me, the core problem of dealing with large language models right 00:25:59.108 --> 00:26:01.178 now is the non-deterministic angle of it. 00:26:01.178 --> 00:26:04.643 And I'm sure you've seen articles that have come out recently about I think 00:26:04.648 --> 00:26:08.543 it was Ford who had a customer ask for a car for free or something like that, 00:26:08.543 --> 00:26:11.723 and they were like legally obligated to give away a car because their, their 00:26:11.723 --> 00:26:13.403 chat bot on their site said they could do that. 00:26:13.403 --> 00:26:13.823 And I think. 00:26:14.053 --> 00:26:16.843 If there's maybe an airline that did something similar recently and, and 00:26:16.843 --> 00:26:19.363 all of these things that are coming up that is like the real consequence 00:26:19.363 --> 00:26:22.848 of using something non-deterministic is that you can offer up a result to 00:26:22.853 --> 00:26:27.478 someone that is completely not based in reality and, and be legally bound to it. 00:26:27.528 --> 00:26:29.953 And there's a lot of really challenging things that come along with that. 00:26:30.368 --> 00:26:34.538 I was talking to another founder recently who, who was musing about LLMs and, 00:26:34.538 --> 00:26:37.638 and was saying something like you know, trying to imagine themselves as a, as 00:26:37.638 --> 00:26:40.188 a kid picking up their first program and trying to imagine how they got 00:26:40.193 --> 00:26:43.778 from like adding up the perimeter of a square to you know, something where 00:26:43.783 --> 00:26:45.098 it's like, ask an abstract question. 00:26:45.103 --> 00:26:47.948 I'll do anything you want is, is a pretty wild response. 00:26:48.168 --> 00:26:50.418 And it's some of the magic of where we're getting to today, but definitely 00:26:50.418 --> 00:26:54.198 like, involves a, a reasonable amount of, of risk and sort of uncertainty. 00:26:54.228 --> 00:26:57.618 And maybe that's why someone building a, a startup like you are 00:26:57.618 --> 00:26:59.088 is isn't a good place to tackle this. 00:26:59.088 --> 00:27:02.038 That you can kind of deal with that uncertainty and take chances that 00:27:02.068 --> 00:27:05.398 probably like, I don't know, Delta Airlines or whoever it was, probably 00:27:05.398 --> 00:27:06.628 shouldn't be taking right now. 00:27:07.258 --> 00:27:07.618 Track 1: Yeah. 00:27:07.623 --> 00:27:09.963 . Just to, just to touch on that a little bit, the other side of risk 00:27:09.963 --> 00:27:13.293 management though, is like also thinking about how the thing is used, right? 00:27:13.548 --> 00:27:13.768 Mike Bifulco: Mm 00:27:14.553 --> 00:27:18.123 Track 1: Like, and, and that's the part that we, we think about a lot internally 00:27:18.123 --> 00:27:22.203 is like, you know, the cost of getting an insurance quote wrong is low. 00:27:23.383 --> 00:27:27.223 It's like, it's not, it is not like you're processing a transac transaction. 00:27:27.223 --> 00:27:30.903 It's not like you are, you know submitting like a birth certificate where the cost 00:27:30.903 --> 00:27:32.463 of an error is like catastrophic, right? 00:27:32.943 --> 00:27:36.603 And so we generally steer ourselves and ourselves and of course our customers 00:27:36.603 --> 00:27:40.713 as well to handling use cases where the cost of an error is low and the air error 00:27:41.223 --> 00:27:44.433 error is not necessarily in like the particular text you input, but the, the 00:27:44.433 --> 00:27:46.593 result of the combination of effects. 00:27:47.043 --> 00:27:49.323 And so one thing we don't do, for example, is we don't do deal 00:27:49.413 --> 00:27:50.703 with any transaction events. 00:27:51.423 --> 00:27:54.633 Like if you wanted to, for, if you wanted to use our product to go buy, 00:27:55.083 --> 00:27:58.173 automate some procurement pipelines on websites that don't have APIs, 00:27:58.563 --> 00:28:01.413 and you wanted to order hundreds of thousands of dollars for the product, 00:28:01.503 --> 00:28:03.663 we would say we'll build a card for you. 00:28:03.723 --> 00:28:06.543 We won't check out, you know, we won't check out. 00:28:06.548 --> 00:28:08.943 And, and that's like a reasonable safeguard to put into place because 00:28:08.943 --> 00:28:11.523 then the value still generated for you where you don't, you don't have to go 00:28:11.523 --> 00:28:15.873 through the effort of building a cart for your procurement pipeline, but, but 00:28:15.873 --> 00:28:18.393 the risk, you can still have a human in the loop to do the final validation. 00:28:18.873 --> 00:28:21.093 Which is something that is necessary today. 00:28:21.123 --> 00:28:24.903 Maybe, maybe in five years it might not not be necessary, but today certainly is. 00:28:25.878 --> 00:28:26.268 Mike Bifulco: Yeah. 00:28:26.388 --> 00:28:26.718 Brilliant. 00:28:26.748 --> 00:28:29.778 That, that is a great way to handle risk and that's actually 00:28:29.778 --> 00:28:31.273 a lot of, of how I talk to. 00:28:31.848 --> 00:28:34.638 So in a past life I worked at Stripe as a developer advocate. 00:28:34.638 --> 00:28:37.998 And this was the, the conversation I would have with a lot of people building early 00:28:37.998 --> 00:28:41.238 stage products is that essentially around credit card information specifically, 00:28:41.238 --> 00:28:45.348 like you don't ever wanna store a credit card number in your database, and Stripe 00:28:45.348 --> 00:28:48.858 has a massive team of very qualified people worried about data privacy, and 00:28:48.858 --> 00:28:52.608 that's why you should offload it there and like trust that stripes the holding 00:28:52.608 --> 00:28:55.578 of your customer's credit card information is better than you could ever do it. 00:28:55.883 --> 00:28:58.013 You're almost the inverse of that, where it's kinda like, well, we just 00:28:58.013 --> 00:29:00.673 wanna do the things , that we are relatively sure the consequence of 00:29:00.673 --> 00:29:03.763 it is, is reasonable both for, you know, sky ver and for your business, 00:29:03.763 --> 00:29:05.203 but for your customer's business too. 00:29:05.653 --> 00:29:08.773 Yeah, I love that that's, that's nuanced and valuable perspective to 00:29:08.773 --> 00:29:09.973 give to your customers along the way. 00:29:10.993 --> 00:29:14.723 In terms of use cases for skyr, so you've got a small amount of customers onboarded. 00:29:14.933 --> 00:29:18.083 Are there use cases that are interesting to you, or ones that you would like to see 00:29:18.333 --> 00:29:21.123 customers jump on board with that you're maybe particularly excited about or you 00:29:21.123 --> 00:29:25.163 think are creative or uses for skyr that, that are maybe non-intuitive at first? 00:29:25.908 --> 00:29:27.828 Track 1: Yeah, I think we've found a bunch of use cases. 00:29:27.828 --> 00:29:30.258 The one that we're actually actively focused on right now is 00:29:30.338 --> 00:29:33.158 in the general field of interacting with government websites. 00:29:33.638 --> 00:29:37.018 So you're a founder when you were, when you incorporated, you can imagine 00:29:37.018 --> 00:29:40.078 the delusion paperwork you had to fill out with the state of Delaware 00:29:40.298 --> 00:29:42.068 with, you know, state of California. 00:29:42.408 --> 00:29:44.538 And we found that there's like a bunch of companies like 00:29:44.538 --> 00:29:45.738 accounting firms, law firms. 00:29:46.463 --> 00:29:50.693 Banks even that ha, that have this like back office that either has 00:29:50.693 --> 00:29:52.913 people doing it or they have like really haphazard pipelines that 00:29:53.153 --> 00:29:56.363 automate it, where they just interact with those government websites. 00:29:56.363 --> 00:29:59.393 So something that is probably on top of many founders' minds right now is 00:29:59.633 --> 00:30:01.253 trialing for Delaware franchise taxes. 00:30:02.348 --> 00:30:04.178 It is due in seven days. 00:30:04.178 --> 00:30:04.808 I haven't done it. 00:30:04.808 --> 00:30:07.628 I don't know if you, your company's done it yet, but you know, it's on 00:30:07.628 --> 00:30:09.758 my to-do list, but it's something you have to do, and if you do it wrong, 00:30:09.758 --> 00:30:12.848 you're gonna end up paying like tens of thousands of dollars worth of taxes 00:30:12.848 --> 00:30:14.438 that are not necessary or get fined. 00:30:15.698 --> 00:30:16.088 Mike Bifulco: Right, 00:30:16.238 --> 00:30:16.418 right. 00:30:17.003 --> 00:30:17.843 Track 1: people have to do manually. 00:30:17.933 --> 00:30:18.173 Right? 00:30:18.173 --> 00:30:20.903 And, and the question, is it, does that have to be something 00:30:20.903 --> 00:30:23.333 that people do manually or can it be automated in the future? 00:30:24.203 --> 00:30:27.083 And so that, that whole area of use case is something that we were very excited 00:30:27.088 --> 00:30:29.803 about very excited about automating, which is kind of funny because like, 00:30:29.803 --> 00:30:30.583 who would be excited about that? 00:30:30.583 --> 00:30:30.793 Right? 00:30:30.793 --> 00:30:34.243 I, I certainly wasn't from like a user perspective, but from a 00:30:34.243 --> 00:30:36.583 a business perspective, I'm very excited about automating that stuff. 00:30:37.033 --> 00:30:37.303 Yeah. 00:30:37.948 --> 00:30:38.728 Mike Bifulco: That's, that's the thing. 00:30:38.728 --> 00:30:41.338 Those are the automations that, that are most valuable is the ones that 00:30:41.338 --> 00:30:44.188 people want to do the least or that are the most painful for them to do. 00:30:44.598 --> 00:30:45.258 Super cool. 00:30:45.258 --> 00:30:46.218 That makes a ton of sense. 00:30:46.268 --> 00:30:48.068 I, the thing I always I. 00:30:48.303 --> 00:30:48.693 Tell. 00:30:48.753 --> 00:30:51.753 Well, so the, the first thing that came to mind for me when you said automating 00:30:51.753 --> 00:30:55.033 government websites is something that I've had to do a few times, which is 00:30:55.243 --> 00:30:58.063 baffling when you have to do it is looking up an EIN for a company you 00:30:58.063 --> 00:31:03.113 own, but you've misplaced and that is so painful to do and such a nightmare, 00:31:03.113 --> 00:31:05.843 and you can understand all the red tape involved in all this stupid 00:31:05.848 --> 00:31:07.343 government websites you have to navigate. 00:31:07.908 --> 00:31:12.088 But I've certainly lost hours of my life to just looking up a eight 00:31:12.088 --> 00:31:13.468 digit code or whatever that is. 00:31:13.568 --> 00:31:13.868 Yeah. 00:31:13.873 --> 00:31:14.828 that's a great use case. 00:31:14.828 --> 00:31:15.878 I'm super into that. 00:31:15.908 --> 00:31:18.188 , and looking forward to seeing, you know, the realities of 00:31:18.188 --> 00:31:19.028 what you're building there. 00:31:19.518 --> 00:31:20.568 , can you tell me about your team? 00:31:20.628 --> 00:31:23.748 So like, how big is the company and what is your stack, what's 00:31:23.748 --> 00:31:24.648 your architecture look like? 00:31:25.223 --> 00:31:28.913 Track 1: Yeah, so we at the end of yc, we did raise like a small pre-seed 00:31:28.918 --> 00:31:30.733 round for Wyvern before we pivoted. 00:31:30.733 --> 00:31:32.353 We ended up stopping the fundraising after that. 00:31:32.353 --> 00:31:34.333 But we are a team of three people right now. 00:31:34.393 --> 00:31:36.553 All three, three of us are co-founders. 00:31:36.553 --> 00:31:39.433 We are all technical people, so we all write code every day. 00:31:39.943 --> 00:31:41.173 Which is, which is cool. 00:31:41.353 --> 00:31:43.903 Obviously the, the responsibilities are not equally distributed. 00:31:43.903 --> 00:31:46.793 Like I, I also write code, but I tend to write the least amount of code because I 00:31:46.793 --> 00:31:48.493 spend most of my time doing sales as well. 00:31:49.163 --> 00:31:49.433 Mike Bifulco: Yeah. 00:31:49.463 --> 00:31:49.733 Okay. 00:31:50.498 --> 00:31:52.088 Track 1: yeah, from an architecture perspective. 00:31:52.093 --> 00:31:55.508 So we decided to keep our product relatively lean as, as it is today. 00:31:55.598 --> 00:31:56.138 This might change. 00:31:56.168 --> 00:31:57.458 This will definitely change in the future. 00:31:57.698 --> 00:31:59.978 And so our product is entirely in the, in Python. 00:32:00.278 --> 00:32:01.088 It's an API product. 00:32:01.088 --> 00:32:02.438 We actually don't have a user interface. 00:32:02.903 --> 00:32:05.633 I think this is the truth about startups that people don't necessarily talk 00:32:05.633 --> 00:32:08.843 about is many companies don't invest in things that aren't necessary. 00:32:09.173 --> 00:32:12.753 Like we haven't had a user interface for four months now. 00:32:12.758 --> 00:32:15.063 We've been onboarding customers and it turns out there's a subset 00:32:15.063 --> 00:32:17.823 of customers out there that are okay with that, that are actually 00:32:17.823 --> 00:32:19.293 totally okay with not having that. 00:32:19.293 --> 00:32:21.333 And so we didn't build it yet. 00:32:21.453 --> 00:32:24.303 We'll build in the future when we have more time and more, more, you 00:32:24.303 --> 00:32:26.283 know al time allocation to build it. 00:32:27.123 --> 00:32:28.503 But our stack is fully in Python today. 00:32:28.503 --> 00:32:32.123 We use like Fast API as our primary API driver. 00:32:32.493 --> 00:32:36.033 And under the hood we use Postgres just as our database, super 00:32:36.033 --> 00:32:37.773 base, as our database data store. 00:32:38.703 --> 00:32:40.939 We run everything on AWS Nothing too 00:32:40.939 --> 00:32:41.360 crazy. 00:32:42.093 --> 00:32:45.403 Mike Bifulco: Certainly familiar shape and clever application of, of the tools 00:32:45.403 --> 00:32:48.503 that you're using to build towards something that makes a lot of sense. 00:32:48.803 --> 00:32:48.923 I. 00:32:49.593 --> 00:32:49.893 Okay. 00:32:49.893 --> 00:32:52.143 And so you're at the point, you've, you've onboarding customers, 00:32:52.143 --> 00:32:53.553 you're starting to ramp that up. 00:32:53.863 --> 00:32:55.323 So tell me about what's next. 00:32:55.323 --> 00:32:57.723 What is the next milestone for Sky Ver that you're looking towards? 00:32:58.473 --> 00:32:58.653 Track 1: Yeah. 00:32:58.653 --> 00:33:02.883 So something we've been thinking a lot about internally is how can we solve a few 00:33:02.883 --> 00:33:06.733 very specific problems that actually have come up in, in our sales conversations. 00:33:06.733 --> 00:33:10.033 And so there's two, two problems that are pretty, that stand out a lot. 00:33:10.273 --> 00:33:10.963 So one is. 00:33:11.308 --> 00:33:14.698 I talk about interacting with government websites a lot as like a use case, but 00:33:14.698 --> 00:33:18.368 one that I didn't touch on is a vertical that has another set of problems, 00:33:18.368 --> 00:33:19.838 which is the healthcare vertical. 00:33:20.508 --> 00:33:23.988 Particularly healthcare and health tech companies do a lot of boring 00:33:23.993 --> 00:33:27.418 form filling as well, largely because of like compliance requirements. 00:33:27.418 --> 00:33:29.788 So they have like disparate systems that each store data, and 00:33:29.788 --> 00:33:32.098 because of compliance, they can't actually like, share, share data. 00:33:32.098 --> 00:33:35.128 So they have like processes that involve taking data from system 00:33:35.128 --> 00:33:36.628 A and transcribing the system B. 00:33:37.288 --> 00:33:39.658 And the problem with these systems is that they require. 00:33:40.228 --> 00:33:42.808 Every vendor to be compliant, like in some way, whether it's HIPAA 00:33:42.808 --> 00:33:45.158 compliance, whether it's SOC two compliance many actually even 00:33:45.158 --> 00:33:46.838 require a self hosted solution. 00:33:47.198 --> 00:33:51.338 And so an idea that we've been going pretty far down is actually open sourcing 00:33:51.343 --> 00:33:53.448 a product to one, solve that problem. 00:33:53.448 --> 00:33:56.358 But two, also just share the cool stuff that we're building with the world. 00:33:56.358 --> 00:33:59.557 You know, as a developer, playing around with a tool that can interact 00:33:59.557 --> 00:34:01.057 with the web would be pretty awesome. 00:34:01.057 --> 00:34:03.067 I mean, there's some websites that certainly wouldn't work 00:34:03.067 --> 00:34:05.017 for, like if you try to use it on LinkedIn, it just doesn't work. 00:34:05.017 --> 00:34:06.037 We don't, we don't allow that. 00:34:06.217 --> 00:34:09.207 But, but there's many websites where it would work for and very cool. 00:34:09.207 --> 00:34:15.267 So one, it shares the cool of the product with the world and gets developers 00:34:15.267 --> 00:34:16.617 pretty excited about what we're building. 00:34:16.617 --> 00:34:19.797 And two, it actually solves a real problem that we've experienced talking 00:34:19.797 --> 00:34:23.157 to our customers is how do we create a self hosted, reliable version 00:34:23.157 --> 00:34:25.617 of our product where our customers actually have full transparency on 00:34:25.617 --> 00:34:29.067 how it works, and open source actually solves that problem completely. 00:34:29.067 --> 00:34:31.267 And so we're pretty excited about launching an open 00:34:31.272 --> 00:34:32.377 source version of our product. 00:34:32.764 --> 00:34:33.034 Mike Bifulco: Wow. 00:34:33.034 --> 00:34:33.934 That's, that's really cool. 00:34:33.934 --> 00:34:37.944 That is super exciting as a not only a founder of a startup, but I tend to 00:34:37.944 --> 00:34:41.574 get into like dangerous side projects on the weekends when I have a silly idea. 00:34:41.884 --> 00:34:44.314 And there's probably half a dozen things that have floated across my 00:34:44.314 --> 00:34:45.514 brain as you and I have been chatting. 00:34:45.514 --> 00:34:47.584 That'd be like, man, I would love to have an automation solution 00:34:47.584 --> 00:34:50.404 for this thing that I is a, a part of my life for whatever reason. 00:34:51.109 --> 00:34:53.149 I think open source is a really interesting angle for this too. 00:34:53.179 --> 00:34:55.309 'cause no doubt you'll see people with creative ideas that you haven't 00:34:55.309 --> 00:34:58.459 even thought of yet for feature additions or, you know, tweaks 00:34:58.464 --> 00:35:01.519 to the API and functionality or shape or whatever the case may be. 00:35:01.849 --> 00:35:05.149 Track 1: One thing we're really excited about open sourcing specifically is as I 00:35:05.149 --> 00:35:08.389 mentioned before, like one of the problems we have is the coverage of, of the web 00:35:08.419 --> 00:35:10.129 with our product is increasing over time. 00:35:10.609 --> 00:35:12.649 And the truth is we are limited, right? 00:35:12.649 --> 00:35:13.609 We have three people working. 00:35:13.609 --> 00:35:15.919 We can only talk to so many people and experience so many websites. 00:35:16.669 --> 00:35:19.099 One thing we are excited about open sourcing is that maybe that 00:35:19.339 --> 00:35:21.859 you will try it out for this idea that you had over the weekend. 00:35:22.489 --> 00:35:25.249 And you'll find an issue with the website and maybe you'll fix it. 00:35:25.759 --> 00:35:26.119 Right. 00:35:26.299 --> 00:35:29.569 And that then, then we can really like harness power of the open source community 00:35:29.569 --> 00:35:33.199 to build a really cool product that can interact with the entirety of the web. 00:35:33.469 --> 00:35:33.649 Right. 00:35:33.649 --> 00:35:35.089 That is like the dream we're going towards. 00:35:35.089 --> 00:35:36.359 And, that could be possible, right? 00:35:36.359 --> 00:35:37.949 With the, with the open source angle. 00:35:38.339 --> 00:35:38.429 Yeah. 00:35:38.879 --> 00:35:41.699 Mike Bifulco: You're enabling a feedback loop for more of the intranet to, to 00:35:41.699 --> 00:35:43.469 make things more broadly applicable. 00:35:43.739 --> 00:35:45.269 I would love to be able to, to poke with that. 00:35:45.519 --> 00:35:48.884 And so you're, you're opening up some open source projects. 00:35:48.944 --> 00:35:51.854 Presumably those are gonna be based on similar stacks, so Python, 00:35:52.244 --> 00:35:54.374 uh, fast, API and Postgres in whatever shape. 00:35:54.654 --> 00:35:55.344 That's really great. 00:35:55.344 --> 00:35:57.534 And so what about in terms of your team? 00:35:57.534 --> 00:35:59.004 Are, are you thinking about hiring? 00:35:59.004 --> 00:36:01.494 Is that something that's gonna come into focus for you anytime soon? 00:36:02.194 --> 00:36:04.569 Track 1: I, I think probably in six months we'd be looking to hire some people, 00:36:04.569 --> 00:36:05.919 but for now we're trying to stay lean. 00:36:06.249 --> 00:36:06.549 Mike Bifulco: Sure. 00:36:06.954 --> 00:36:09.969 Track 1: we're actually like optimizing for high, high frequency communication 00:36:09.969 --> 00:36:14.004 and like just when you have a small team, you have only a small number 00:36:14.004 --> 00:36:15.084 of projects you never work on. 00:36:15.351 --> 00:36:15.981 And, and that 00:36:16.311 --> 00:36:20.451 small number of projects actually for startups can be a benefit 00:36:20.481 --> 00:36:23.421 because it really makes you focus on what is good or bad. 00:36:24.111 --> 00:36:26.661 And that's something that people don't necessarily always talk about is like 00:36:26.661 --> 00:36:30.051 the, the real advantage of the small team, which is high, high amount of 00:36:30.056 --> 00:36:31.371 ruthless prioritization is needed. 00:36:31.371 --> 00:36:32.961 Otherwise you end up working on things that don't matter. 00:36:33.351 --> 00:36:35.721 And if you work on things that don't matter to startup, then you 00:36:35.721 --> 00:36:38.031 launch things next week instead of launching things this week. 00:36:38.481 --> 00:36:41.451 And that effect compounds negatively. 00:36:41.961 --> 00:36:43.071 And so we are actually quite a. 00:36:43.881 --> 00:36:45.771 Happy with how lean our team is today. 00:36:45.861 --> 00:36:48.801 Obviously it won't be sustainable as we ramp up sales and acquire more customers. 00:36:48.861 --> 00:36:52.231 You know, there's some use cases we're solving for right now that are one 00:36:52.236 --> 00:36:53.221 of the companies we're helping with. 00:36:53.221 --> 00:36:56.311 We help 'em apply to like, job applications online and, and at night 00:36:56.311 --> 00:36:59.491 we'll wake up and we, and some jobs will fail and we have to manually 00:36:59.491 --> 00:37:03.031 resubmit them, you know, and so as we acquire more customers, we won't be 00:37:03.031 --> 00:37:04.621 able to scale that side of it forever. 00:37:04.621 --> 00:37:07.441 So we'll definitely need to hire for some help, but that's not something we're 00:37:07.441 --> 00:37:08.971 planning any time in, in the short term. 00:37:09.591 --> 00:37:10.881 Mike Bifulco: that's really good perspective too. 00:37:11.151 --> 00:37:14.661 A lot of the successful founders I've spoken to are really good at that 00:37:14.666 --> 00:37:18.691 prioritization and things like cutting scope dramatically is like a real skill. 00:37:18.941 --> 00:37:21.156 And just like you said, shipping this week is always more important than 00:37:21.156 --> 00:37:23.916 shipping next week, especially if you're able to learn from it and sort 00:37:23.916 --> 00:37:25.326 of iterate more quickly as a result. 00:37:25.326 --> 00:37:25.716 That's great. 00:37:25.716 --> 00:37:29.416 So for folks listening to the, the show right now where's the best place 00:37:29.416 --> 00:37:30.796 to go to get started with Sky Ver. 00:37:31.401 --> 00:37:34.891 Track 1: So if you go to sky.com, you'll see a link to our GitHub. 00:37:34.921 --> 00:37:38.461 I would recommend checking it out and you know, if you could clone the 00:37:38.461 --> 00:37:41.881 repository, play around with it, try a couple websites, maybe open a bug report, 00:37:41.911 --> 00:37:43.531 that would be super helpful for us. 00:37:44.011 --> 00:37:46.141 Also, any stars are of course, always appreciated. 00:37:46.141 --> 00:37:48.601 , Mike Bifulco: for people who may be interested in sort of like. 00:37:49.486 --> 00:37:50.416 The customer side of things. 00:37:50.416 --> 00:37:54.466 So if, if people are working on teams that they feel like might be a candidate 00:37:54.466 --> 00:37:58.256 for someone who would be a good customer for you is there a place to reach 00:37:58.261 --> 00:37:59.756 out on, on the website there as well? 00:38:00.371 --> 00:38:02.081 Track 1: Yeah, so if you go to our landing page, there should be 00:38:02.081 --> 00:38:04.241 a button where you can actually book a meeting directly with me. 00:38:04.571 --> 00:38:07.331 Feel free to feel free to do that anytime, even if you don't have 00:38:07.336 --> 00:38:10.061 any customer use, 'cause I'm always happy to chat and share and talk. 00:38:10.541 --> 00:38:13.151 But if you also wanna get in touch with me, you can always email me 00:38:13.151 --> 00:38:18.611 at ton at we or@sc.com, so@sky.com and I'll be happy to respond. 00:38:19.032 --> 00:38:19.422 Mike Bifulco: Okay. 00:38:19.482 --> 00:38:21.102 Well switching to thank you so much for joining today. 00:38:21.102 --> 00:38:22.482 It's been a pleasure talking to you. 00:38:22.482 --> 00:38:26.532 I'm really, really excited to see that you're opening up some open source angles 00:38:26.537 --> 00:38:28.122 for, for the community to tackle here. 00:38:28.372 --> 00:38:29.182 And I wish you the best. 00:38:29.212 --> 00:38:31.932 Please feel free to, to come back and join us anytime if you have 00:38:31.937 --> 00:38:34.122 other things you wanna chat about or if you have launches upcoming. 00:38:34.182 --> 00:38:35.592 And tha thanks a lot for joining me today. 00:38:35.592 --> 00:38:36.072 I appreciate it. 00:38:36.582 --> 00:38:37.182 Track 1: Oh no, it was . Great. 00:38:39.462 --> 00:38:39.762 Mike Bifulco: Sure. 00:38:39.882 --> 00:38:40.332 Talk to you soon.