The End of Manual Debugging

Aaron Epstein 0:00 [upbeat music] I am excited to welcome Sherwood from Sazabi, who has exited his first YC company and is coming back through the YC batch again for a second time. Sherwood, thanks so much for joining.

Sherwood Callaway 0:15 Yeah, really excited to participate in the upcoming YC batch. Uh, be back here for a second time. I think it's gonna be a really different experience, and to, uh, introduce Sazabi to, to the batch and to,

Sherwood Callaway 0:24 and to the world, so.

Aaron Epstein 0:25 Yeah, amazing. So maybe to start off, um, tell us a little about what Sazabi does.

Sherwood Callaway 0:29 Sazabi is an AI-native observability platform. It's specifically designed for, uh, fast-moving engineering teams. Um, and one way to think about it is like, it's like a Datadog or a Sentry, but instead of being built in twenty-ten or twenty-twelve, it's built in twenty-twenty six, and it's aggressively AI native or uses a lot of AI agents. Um, and the idea basically came from this, uh, experience I've had, uh, over the course of my engineering career, where every time there's an outage in production, every time there's a bug, I would spend hours digging through dashboards and log searches and looking at flame graphs and kind of not really knowing what's going on, and eventually I might get to the root cause, maybe not. Uh, I felt like AI coding tools had gotten very powerful, and the, the part of my job that involved creating new software was, was, uh, really changing, but the part of my, uh, job that involved maintaining existing software wasn't really changing. With Sazabi, customers can basically ask questions about how their production system is doing. Uh, they can ask things like, uh, "Why is production down?" Or, "What does this error mean?" Or, "Which customers are affected?" Or, "Which commit is responsible?" Uh, and it makes it very easy to, to find and fix production

Sherwood Callaway 1:36 issues.

Aaron Epstein 1:36 Amazing. And you, you kind of have a controversial take on this and how you're building it, that it's just built off of the log files, right? Tell us a little bit more about that.

Sherwood Callaway 1:44 Yes. So one of the things that we say at Sazabi is, "Logs are all you need." And it's actually one of three different, uh, sort of hot takes that we have. We call them our core principles, and they're a part of, uh, our manifesto, which we, we published. Logs are all you need specifically refers to this idea that we think you only need logs, like only need this specific type of telemetry, uh, to do observability well. Um, and that's controversial because in the past, for a long time, uh, people would say that you need something called the three pillars of observability. You need logs, metrics, and traces, and, uh, that's because, like, these different data types serve different types of queries, and, uh, some data types are, are large and some da- some data types are small. And so for different types of things, you, you, you end, end up using these different kinds of data. But the result is that you have three different things that you need to instrument. So for every engineering team out there, they need to implement, uh, logs, metrics, and traces for all of their services. That's a huge pain in the butt. Uh, logs relative to metrics and traces are so much easier to use. Like, every developer knows how to do a print statement, um, and every developer knows how to read a, a log stream. And so we think that, uh, logs kind of represent the Occam's razor of observability. Like, it's... Anyone can do it. Like, it's the simplest possible way of, of doing it. Um, then there's another thing which makes this true, uh, which is this AI paradigm shift. So in the past, logs were kind of the least valuable type of telemetry because they're mostly unstructured, unstructured. Um, that's, that refers to the fact that, like, the log line contains, like, natural language, and it's-- and you can't, like, have a machine read over it and derive a lot of insight. Well, obviously, AI changes what's possible there, and, uh, you can now have AI agents read all of your log lines and tell you what's going on. So that's the idea behind logs are all you need.

Aaron Epstein 3:33 Very cool. And obviously, observability is changing a lot over the last couple decades, but, um, talk about how you got your start early on, uh, setting up observability at Brex.

Sherwood Callaway 3:43 Yeah. So I started my career as a front-end engineer, but I was, like, the junior most front-end engineer at, at Crunchbase originally, and, um, because you're a junior, like, you don't get to work on any of the important stuff. So my team was complaining about CI/CD and about how long their builds and deploys would take. So I went down this rabbit hole, which is CI/CD, and figured out how to make our deploys faster, uh, and that led me to infrastructure and DevOps, uh, which turned out to be something that was, uh, really fascinating for me. I got basically obsessed with this problem of, like, how do we make production reliable, and how do we make developers more productive and allow them to, to, to ship faster and to ship higher quality code? So I was already kind of primed and interested in that space, and then I went to join Brex, where I was lucky to be, um, basically the third infrastructure engineer at Brex. Um, I think it was roughly employee seventy, and they had just stood up what, what they call their foundation team. My... The second infrastructure engineer I think had joined the same week.

Aaron Epstein 4:39 [laughs]

Sherwood Callaway 4:39 So I can't quite call myself the second, but, like-

Aaron Epstein 4:41 [laughs]

Sherwood Callaway 4:42 ... really close. Uh, and actually, he works at Sazabi today, which is great.

Aaron Epstein 4:46 Oh, very cool.

Sherwood Callaway 4:47 While I was at Brex, first we built out, uh, all of Brex's infrastructure that supported the company through hypergrowth, so all of our, uh, staging and, and production environments, the microservice framework, CI/CD system. Eventually, when we had rolled out, like, a lot of this basic infrastructure and the foundation team had grown a lot, we were in this position where we were starting to splinter the foundation team into different subteams and focus on different specializations. And we had found ourselves in a position where we had, like, fifty different microservices running in Kubernetes as a part of our production cluster, and they were all owned by different teams, and some teams would own multiple microservices, and it was getting hard to kind of know what was going on in production. And so the answer I didn't know at the time because I was too junior, but the answer is observability. Like, this is the reason that observability exists, is to help, uh, developers or engineering teams ask questions about, uh, their production systems. And so I was fortunate to be one of the team members that helped start the observability team at Brex. Uh, and then while working on, on observability at Brex, I became very observability pilled. [laughs] Basically, like, the obs-observability philosophy is that you can, you can have, uh, integration tests, you can have unit tests, you can have QA processes, you can have, like, release processes. You can do everything you want-Like static analysis before the code goes to production to try to make sure that it's right and that it's gonna, uh, behave well. But, like nothing really prepares you for production. It, it sort of reminds me of that, um, that Mike Tyson quote where he's like, you know, "Everyone's got a plan till they get punched in the face." [laughs]

Aaron Epstein 6:17 Right.

Sherwood Callaway 6:18 That's so true of, of shipping software. Uh, so once you're in a production, you just can't predict what's gonna happen. You can only prepare yourself to respond, uh, and that's what observability is.

Aaron Epstein 6:29 And how long did you spend, uh, building that out at Brex?

Sherwood Callaway 6:32 I worked on observability at Brex for about a year. Uh, in total I worked at Brex for about two and a half years.

Aaron Epstein 6:37 Mm-hmm.

Sherwood Callaway 6:38 The main things that we did while we were on that observability team was setting up auto instrumentation so all of, every microservice that gets created, uh, automatically receives a bunch of metrics and traces and logs, um, installing all the infrastructure that's responsible for capturing and forwarding that information to our observability systems, setting up tools like Datadog and configuring them. Like, I went, uh... I mean, Datadog is like hundreds of modules, but I feel like I've explored every single one and every knob and configuration option and, and tuned it appropriately for, for the team at Brex. Uh, and then now have done that at a couple different companies. Um, and other things like making sure that every service has, uh, like a dashboard that covers all of the key metrics and, um, and basic monitors.

Aaron Epstein 7:22 Yeah.

Sherwood Callaway 7:22 And driving, uh, SLO and SLI adoption, which is a little bit, um, more... That's, uh, some observability lingo that is more common in the, the enterprise and SRE space, but basically, uh, making sure that every team knows how to use observability tools and is measuring their services.

Aaron Epstein 7:37 Mm-hmm. And talk about your decision to ultimately leave Brex and start your own company.

Sherwood Callaway 7:42 Yeah. Honestly, I feel like I'd been latent for a long time. Yeah, like I started my career as a, uh, through a web development boot camp, and, uh, that boot camp was here in San Francisco. I grew up in North Carolina, so pretty far cry from like Silicon Valley, startup tech culture. And while I was at this boot camp, everyone I knew was reading Hacker News and, uh, wanted to work at the hot YC startup. It was like Stripe and Gusto and, uh, Instacart at the time. Uh, and it was just an amazing experience for me and it, it made me really want to work in startups and specifically to do my own YC venture backed tech startup. So from the very beginning of my career, like that had been my long-term goal. While I was working at Brex in... I was living in New York with my roommate, uh, who was also an early Brex team member. He and I had relocated to New York together with Brex from San Francisco to help start this New York office, but the, the pandemic happens, and so we got stuck in New York with like a two-year lease. [laughs] And of course, like, what are you, what do two like software engineers do like when they have a lot of free time and no, nothing going on? So we started to talk about startup ideas and felt that we had a really special team and wanted to do something together and, uh, applied for YC and that was, uh, I guess the rest is history.

Aaron Epstein 8:50 Yeah.

Sherwood Callaway 8:51 Yeah.

Aaron Epstein 8:51 Um, talk a little bit about how you came up with the idea that you originally applied with.

Sherwood Callaway 8:54 My first company was called Opkit. We went through YC Summer '21, and which was one of the COVID batches. And what Opkit was, was a voice AI for a healthcare company. So we build LLM-based voice agents to automate calls to insurance companies for things like, uh, e- eligibility, which is basically when the, the provider, the, tries to determine whether you have coverage or not for your, your procedure, and for, uh, uh, prior authorization, which is another hoop that insurance companies make you jump through to get certain procedures.

Aaron Epstein 9:24 Mm-hmm.

Sherwood Callaway 9:25 Uh, and also, uh, to do, uh, claim, to check on the status of, of claims that have been submitted. Um, it's a pretty esoteric field. It was a very be- big field, but at a far cry from anything I'd done previous.

Aaron Epstein 9:36 Yeah. Why did you decide to work on that?

Sherwood Callaway 9:38 Yeah. It's, I was... Well, I guess, to start with, my, my dad was a doctor growing up.

Aaron Epstein 9:43 Mm.

Sherwood Callaway 9:43 Um, so I had a little bit of a lens of like, you know, if, if I know, if what I know is like just software engineering and tech startups and, you know, I need to have an idea that's like out in the real world, right? Like, that, that impacts real businesses, not like just tech businesses. So the only field that I really knew anything about outside of, um, uh, technology was, was healthcare, and I felt like I had access to my dad, who's a subject matter expert, and to his practice and his, uh, his colleagues, and I could go interview the, the staff and the, the administrators there. Um, so it felt like an opening, and it meanwhile, like I was coming from a lens of having worked at Brex and having been very interested in fintech during fintech's like real rise from like during the second half of the 2010s, and, uh, was thinking about like what's next for fintech, you know. It's seems like we've done a lot of cr- like consumer fintech and now we have like the Brexs of the world, the Brexs and the Mercurys. Um, maybe verticalized fintech would be what's next. And so I thought, why not? Like, healthcare seems like a big vertical. Like there's a lot of money moving there. Um, a- in retrospect, I think, uh, Opkit was kind of like an MBA case study approach to, to picking a company or picking an idea. It wasn't based on personal experience or insight or, uh, like, or passion, candidly. It was based on this, um, this real desire to start something and, uh, and, you know, grasping basically for, for whatever I think I, wherever there was an opportunity.

Aaron Epstein 11:12 Yeah. Did you ever consider moving outside of healthcare or for, you know, the la- You spent years working on Opkit.

Sherwood Callaway 11:19 Yeah. Yeah.

Aaron Epstein 11:20 Wa- was it always like, this is a healthcare company and we must stay within healthcare?

Sherwood Callaway 11:24 Yeah. We, you know, I regret not evaluating that like fu- fundamental condition more carefully. Um, I mean, actually I don't have a ton of regrets because I am here where I am today, and like, uh, those, there were a lot of hard lessons learned from Opkit, but now Sazabi gets to benefit from all of them. But we went in like with a healthcare idea and I, you know, my c- my co-founder Justin and I love working together and we felt like we were learning a lot and making some progress. Slow, slow and steady progress, um, and peeling back layers of the healthcare ecosystem and learning about new problems within the space. So we, we enjoyed it, but I don't think that it was the right thing for us to be working on. And, um, one thing that I wish we had, uh, thought, thought more about is like whether this is really playing to our strengths.

Aaron Epstein 12:08 Mm-hmm.

Sherwood Callaway 12:10 And I think that's like a great piece of advice for, for YC founders and founders in general is like look for your personal expertise or, um, or b- or build f-based something, something based on your prior experience.

Sherwood Callaway 12:21 You probably have more prior experience than, um, than you think, and there are probably more openings there than you realize.

Aaron Epstein 12:27 Yeah.

Sherwood Callaway 12:28 You don't need to build a business in some other field just because it feels like whatever you're, you've been doing is, is not real.

Aaron Epstein 12:34 Yeah.

Sherwood Callaway 12:35 Um-

Aaron Epstein 12:35 Yeah, we tend to see that a lot, people that have a lot of experience in an area. For you it was, you know, observability and even in fintech, you probably had a lot of fintech experience, much more than, than other, uh, people would.

Sherwood Callaway 12:47 Yeah.

Aaron Epstein 12:47 And then explicitly saying, "I don't wanna do that because I, either I know too much about it so I know, you know, all the problems or what's gonna be difficult about it, um, or I'm kind of burned out on it-

Sherwood Callaway 12:58 Yeah

Aaron Epstein 12:58 ... and so I'm gonna go and do something that I've never done before and I don't have the, the depth of experience or insights on." Was there a moment when you were building Opkit, maybe in hindsight looking back, where you're like, "I should have known at this moment that this was not the right thing for us"?

Sherwood Callaway 13:14 I think what's funny, it was like we were definitely not the right team to start a healthcare fintech company or a healthcare AI company when we started, but progressively we became that team.

Aaron Epstein 13:24 Mm-hmm.

Sherwood Callaway 13:24 So ironically, like if I had, if I... if Sazabi were a, a different kind of company it could be a healthcare or a health tech company because we have now four years of hard-earned experience in that space.

Aaron Epstein 13:34 Right.

Sherwood Callaway 13:34 And at some point it was like, oh, we've got one year of experience in this space, and we're like, "No, more than anyone we talk to." Um, and I guess this is the definition of the sunk cost fallacy, but, uh, there were, there were times when we, especially during YC, which was at the very beginning of our company, uh, which really kind of like kicked us out of the, out of the nest so- to some extent, um, where we evaluated it, but we... I don't think we went deep enough about that or really reckoned with like what it would look like for us to spend the next 10 years of our life working on a healthcare or health tech company. Um, yeah, we should have thought harder about that.

Aaron Epstein 14:07 Yeah.

Sherwood Callaway 14:08 [laughs]

Aaron Epstein 14:08 Well, it's like the curse of the second time founder a lot of times is-

Sherwood Callaway 14:11 [clears throat]

Aaron Epstein 14:11 ... coming up with the idea to work on because you know too much, and-

Sherwood Callaway 14:15 Yeah

Aaron Epstein 14:15 ... one of... I mean, you're describing naivete of jumping into this space that you knew nothing about-

Sherwood Callaway 14:20 Yeah

Aaron Epstein 14:20 ... and trying to learn as quickly as you can, but sometimes that's what's needed too just to be able to-

Sherwood Callaway 14:25 Yeah. I, I think that's true. I mean, I think that if you're, especially if you're like a really young person and you're ready, you wanna get started, you... like you can be an expert at anything, but you have to be prepared to put a couple of years of experience in, um, to, to learn the ropes, to ramp up, to build a network, um-

Aaron Epstein 14:40 Yeah

Sherwood Callaway 14:40 ... to gain trust with, with people in the space.

Aaron Epstein 14:42 Yeah.

Sherwood Callaway 14:43 'Cause, um, th- those things are required I think. They're preconditions for success.

Aaron Epstein 14:47 Yeah.

Sherwood Callaway 14:47 But if you have some experience, like it doesn't need to be hard. Like you can just trade on the experience that you've already gained thus far, and for Sazabi, I, you know, if everything about Opkit was out of alignment with who I am and my background and my experience and my passion, with Sazabi like the first thing I wanted to do was to r- align everything with who I am. And so I was like, "I'm gonna build a tool for people like me, um, based on my personal experience. Like it's gonna be an observability thing, it's gonna be a developer tool. I'm gonna sell it to gr- like early and growth stage tech startups, which is like the only companies ever- I've ever worked at. Um, and I'm gonna name it after, um, Mobile Suit Gundam," which is a 1970s, '80s era sci-fi anime which I love. [laughs] So I was like, "This is a symbolic act-

Aaron Epstein 15:33 Yeah

Sherwood Callaway 15:33 ... about making Sazabi, um, in alignment with who I am."

Aaron Epstein 15:36 Yeah.

Sherwood Callaway 15:37 And, uh, that's probably the most important thing that we... most important mis-mistake I think we made with, with Opkit.

Aaron Epstein 15:42 Yeah. Yeah, talk about that maybe like, uh, the tail end of the Opkit journey and some of the lessons that you learned there, and then what brought you back to starting another company, um, and deciding to work on this.

Sherwood Callaway 15:55 I guess to catch people up on Opkit, we were working for about two-ish years and we had built what's called a revenue cycle management software company where we were basically helping medical practices, uh, check patients' insurance when they come in, um, keep that information on file, and then prepare and submit claims to insurance companies. We had built kind of a traditional SaaS tool for this. Um, LLMs obviously happened in that time, and one, one of the things we learned while working on Opkit is that there's a lot of operations that healthcare providers can only do over the phone with insurance companies, and that's... we can... maybe that's by design, maybe that's not. Uh, and so we had built a call center in the Philippines of people who are calling insurance companies, uh, to support our, the, the rest of our product. Um, and we started using LLMs to, to quality assurance on those calls and to do data entry and data extraction from the call transcripts, and then we built an LLM voi-voice agent. So I, I think we were one of the first teams to actually build and commercialize an LLM-based voice agent in healthcare.

Aaron Epstein 16:53 Mm-hmm.

Sherwood Callaway 16:53 Um, it was very early. It was very hard. Uh, I think we really enjoyed the engineering challenge [laughs] more than anything else, which probably points to, to what we should've been doing.

Aaron Epstein 17:02 Yeah.

Sherwood Callaway 17:02 And we did that pivot in our last, or in our second year at Opkit. Um, and so then our... we had about a year worth of runway left to build and commercialize that solution, basically retool the company, reframe it, and then, uh, and go out and try to raise the seed extension to see if we can kind of keep things going. And got to the point where we had about six more months left of runway. We had prepared all of our materials. We had launched the products. Uh, went out to fundraise. Some of our early conversations were, were not that exciting. Um, kind of didn't seem like investors were super, super interested, and that led us to reconsider whether this was the right thing for us to be doing.

Aaron Epstein 17:39 Mm-hmm.

Sherwood Callaway 17:39 And so, um, our decision was that it was not, and there's still plenty of time for us to go do something better.

Aaron Epstein 17:44 Mm-hmm.

Sherwood Callaway 17:45 So we then started to seek out a, a- acquirers, and talked to a lot of growth stage healthcare companies and growth stage fintech companies, and that, to be honest, that felt like more of the same. You know, we'd be, we'd be in a larger organization doing some of the, working on some of the same problems we've already been working on.

Aaron Epstein 18:01 Mm-hmm.

Sherwood Callaway 18:01 Uh, and then we talked to one AI sales tech company called 11X, which is based here in San Francisco, um, that was growing super fast. They were building a voice agent. We, we knew their CTO from, from Brex.

Aaron Epstein 18:14 Mm-hmm.

Sherwood Callaway 18:14 And it seemed like, uh, a much more exciting-Place for us to be. We wanted to get closer to, to AI, and so we decided to join Eleven-X.

Aaron Epstein 18:22 Very cool. And yeah, talk about the decision to leave that and then start Sazabi again and

Sherwood Callaway 18:27 Yeah, I mean Eleven-X was am- amazing and, and, uh, I'm really proud of the product we built there, and the team we... Our, our team came in and, and did a lot of great engineering work there. But, uh, first thing we did was rebuild their AI SDR product, their sales rep, and that was probably like a three to four-month-long project of, you know, building it from scratch, uh, cutting over from the old platform to the new platform and, and scaling it and kind of getting it to steady state. And in month like five, where most of that work had been done, we're now in sort of maintenance mode, and that's like, okay, well, I've now set up Datadog again.

Aaron Epstein 19:01 Mm-hmm.

Sherwood Callaway 19:01 I've set up all of the observability tools and CI/CD tools that I always set up, and because I'm tech lead for this project, anytime something goes wrong, like I get a Slack message, or someone, like a, a go-to-market person comes and like taps on my shoulder. Um, so now I'm like the on-call guy.

Aaron Epstein 19:16 Mm-hmm.

Sherwood Callaway 19:16 And I found myself... We had, we're building a futuristic AI product using futuristic AI coding tools, and then when I would go to debug it, and it was the same painful manual experience that I've had for my entire career. Uh, and so it was just, it, it, it was painfully clear to me that, uh, we were going to do to, uh, observability what was already being done to, uh, like code generation or, or, or the, the act of creating new code.

Aaron Epstein 19:43 Mm-hmm.

Sherwood Callaway 19:44 Uh, and I also realized, wait a second, like this is like the perfect company for me, right?

Aaron Epstein 19:49 Mm.

Sherwood Callaway 19:49 Like it's, it, it's completely aligned with my interests, my experience, my combination of experience and, and now AI and, uh, and infrastructure and dev tools and observability and, you know, I'm a s- I'm, I, I've already been a founder, so I kind of, it's not, it wouldn't be a scary, new thing.

Aaron Epstein 20:07 Yeah.

Sherwood Callaway 20:07 Like I know how to fundraise, I know how to build a team, uh, I know how to do a big launch. I know how to do press. Um, so all of the, I, I felt like I had all of the right i-ingredients to do this company and that this was like the company I was born to, to build.

Aaron Epstein 20:23 Yeah.

Sherwood Callaway 20:24 So...

Aaron Epstein 20:24 Talk about anything that you specifically want to do differently with this company based on-

Sherwood Callaway 20:28 Yeah

Aaron Epstein 20:28 ... lessons that you learned from building Opkit.

Sherwood Callaway 20:30 Yeah, I mean, the, the big one was like, build something I know.

Aaron Epstein 20:33 Mm-hmm.

Sherwood Callaway 20:33 Uh, leverage my personal experience, pl- play to my strengths, call it Sazabi, because it should be, uh, it should be a representation of who I am. Um, another one is that I thought it should be fun, and I think as fun as Opkit was, like I, I loved working with my co-founder and my best friends. I loved going through YC. Uh, healthcare was like intellectually rewarding and, and interesting and, um, there were a lot of new experiences, which were, which was, which was great. I can't say that it was like fun.

Aaron Epstein 21:00 Mm-hmm.

Sherwood Callaway 21:00 You know? Like we were pushing that boulder and boulder never really, they tell you like the boulder, you get over the hill and the boulder starts to roll. Like it never really rolled for us, and it was really hard. Um, so my decision to start Sazabi was like, when I start this new company, it's gonna be fun.

Aaron Epstein 21:17 Mm-hmm.

Sherwood Callaway 21:18 Um, and so I, I, I think that, uh, calling it, uh, something that I, I love and, and building in a space that I love and working with, um, frankly like customers and, and other developers that I would be grabbing beers with anyways if I wasn't, uh, doing this, if they weren't my customers-

Aaron Epstein 21:34 Yeah

Sherwood Callaway 21:34 ... uh, was another, uh, thing, another thing that I wanted to do to make it, um, feel like, uh, like a better time.

Aaron Epstein 21:41 Yep. Yep. Talk about your future vision for Sazabi. Where is it going five to 10 years from now?

Sherwood Callaway 21:46 Well, our, our mission is to become the default observability tool for all companies, uh, in what we call the AI era-

Aaron Epstein 21:53 Mm-hmm

Sherwood Callaway 21:53 ... which I would say is, um, probably starts with, uh, agentic, like the year of agents in 2025 and, and forwards. Uh, our vision is to create this world of, uh, self-healing software where software, um, i-improves i-itself without human, uh, intervention or direction. And, uh, I think one thing that I, that I remark on and reflect on is the fact like companies like Cursor are automating this part of the software development role, which is like creating new features and apps, and that's obviously like a big part of the job, and software engineering is like a trillion-dollar industry globally. So right, like big opportunity. Cursor's a big company.

Aaron Epstein 22:29 Mm-hmm.

Sherwood Callaway 22:30 But a much bigger part of my job as a developer is maintaining software that's already been written, and I think that is the opportunity for Sazabi. That's what we're focused on.

Aaron Epstein 22:39 Mm-hmm.

Sherwood Callaway 22:39 So, um, I, I think we definitely have the, uh, the chance to be like a really incredible and generational company, um, building what we're building.

Aaron Epstein 22:47 Yeah. Amazing. Talk about the decision to come back and, and do YC again

Aaron Epstein 22:51 for a second time.

Sherwood Callaway 22:51 Yeah. This one is interesting. I feel like there's like a lot of reasons not to, right? It's like, uh, I've already done YC. I'm already a part of the network. You know, I've already got access to Book- Bookface. Um, you know, it's on my resume, which I feel like that's what a lot of people want.

Aaron Epstein 23:06 Mm-hmm.

Sherwood Callaway 23:07 So it's a little counterintuitive. Um, I guess from a business perspective, the first thing that comes to mind is that YC is an accelerator and, uh, we need to accelerate, like Sazabi needs to accelerate. Uh, our opportunity is here today, like will not be here forever. There are other companies that will try to do it.

Aaron Epstein 23:24 Mm-hmm.

Sherwood Callaway 23:25 And, uh, I don't want to waste any time. Like it feels like very ephemeral to me. So I, I want my team to move fast. I think YC will help us move faster. Um, another thing is kind of, um, culturally, I w- I, I want like to establish a culture at Sazabi that is kind of based on some of the, or is imbued with some of the YC values. Like YC will create, like it creates a deadline for us very early. It has many, there are many smaller deadlines along the way. So we just have to get used to this, uh, this cadence of shipping quickly. And, um, YC also pushes you to get to market super fast, which I don't... We have a team of engineers, right? We're like eight engineers or something. So yeah, it's, it would be very easy for us to just build forever-

Aaron Epstein 24:04 Mm-hmm

Sherwood Callaway 24:05 ... and not commercialize. So I want that. Um, and I, I guess on a, the, the last note related to, um, from, from a business perspective is g- distribution and go to market. Um, and this is like a little more tactical, but, you know, every YC company is a software company. Every software company has an observability solution.Sazabi should be that solution. Um, I'll be following in a tradition of great YC companies like Brex and Deel and Rippling that have sold into, uh, their batchmates-

Aaron Epstein 24:36 Mm-hmm

Sherwood Callaway 24:37 ... uh, really successfully. Uh, and I, I think that Sazabi can do the same thing, so I wanna use that as an opportunity.

Aaron Epstein 24:42 Yeah.

Sherwood Callaway 24:42 Um, then on a personal level, my first YC experience was, uh, was remote. Uh, we were one of the COVID batches, so it's a really different experience. I mean, I'm here in this space today that I've actually... I, I've never been in this building before, uh, which I think is like e- um, points to the fact that I have not had, like, the full experience and, uh, it's been a dream for a long time. So I want, I wanna, uh, I wanna stand on the stage at Demo Day and, and actually have that experience and, and make real connections with my batchmates, um, that happen here in person.

Aaron Epstein 25:15 What's something you wish you knew at the beginning of this journey? You've, you've been through the arc of starting one and-

Sherwood Callaway 25:21 Yes, which beginning are we talking?

Aaron Epstein 25:22 ... exiting and, exiting it and starting a new one-

Sherwood Callaway 25:25 Yeah

Aaron Epstein 25:25 ... and, um, maybe if there was something that, uh, you could go back and tell yourself at the beginning of-

Sherwood Callaway 25:30 Hmm

Aaron Epstein 25:30 ... uh, the Opkit journey, your first time going around during YC. What, what is the advice you would give yourself?

Sherwood Callaway 25:36 I wouldn't wanna give any advice that would actually change the outcome, 'cause I think I'm building the, the perfect company today.

Aaron Epstein 25:40 Mm-hmm.

Sherwood Callaway 25:41 Um, and then I think all of the years of, like, eating glass with Opkit, like, that's, like, my Glenn Gle- Glengarry Glen Ross period, right? Where I'm just, like, knocking on providers' doors, and they, like, don't wanna take my meeting. And, um, we built a lot of grit and learned a lot about, like, how to build a startup under tough circumstances. Uh, so I'm really grateful for those lessons and would not give them away. Something that I might tell myself, or maybe just remind myself, is that startups are... You know, a lot of people perceive startups to be short and fast, but actually we are in a long-term game. It's like a long horizon game.

Aaron Epstein 26:20 Mm-hmm.

Sherwood Callaway 26:20 And the lessons I learned from Opkit, the relationships I built, uh, with investors, with customers, with other engineers, like, I still keep up with my batchmates. Um, the... You know, I, I, I've... Many of my batchmates are starting other companies that I could sell into or, um, that, like, I could potentially recruit them. Like, these are relationships that w- we, you build over a long period of time, and, uh, they compound. And so it's, it's good to remember, to remember that, be mindful of it, uh, have integrity, be good to everyone you work with, 'cause you never know. Like, five years from now you might be asking them for money. Um, so yeah, I, I would... I think that's a good, uh, good tip

Sherwood Callaway 27:00 probably.

Aaron Epstein 27:00 For anybody that's watching, talk about, um, hiring-

Sherwood Callaway 27:04 Yeah

Aaron Epstein 27:04 ... and the ideal engineers-

Sherwood Callaway 27:07 Yes

Aaron Epstein 27:07 ... to come and work at Sazabi.

Sherwood Callaway 27:08 We are definitely hiring. Uh, we're hiring a lot of different roles. I would say that the ideal candidates are, uh, they're high agency, they learn quickly, they, they love tools. That doesn't need to be dev tools specifically, but just tools. Like, they kind of geek out on the, the saw, so to speak.

Aaron Epstein 27:27 Mm-hmm.

Sherwood Callaway 27:27 Like, the thing that makes you, helps you do the thing. Former founders are great. Uh, people... I mean, former founders just, like, have to figure things out.

Aaron Epstein 27:34 Mm-hmm.

Sherwood Callaway 27:34 You know, it's... And, and, uh, they know what it's like to just do hard work, to, um, to put in long hours or to do things that are un- unsexy. Uh, they're also very flexible, and I think right now with AI, we never, we don't really know what our jobs are gonna look like or what the, what the org structure and the different departments are gonna look like two years from now. So we'll, we need people who are flexible and self-starters. Uh, on a more practical level, like, we're looking to basically hire software engineers in a lot of different roles. So, um, data and databases and infrastructure are hugely important to Sazabi. Like, we build our own log ingestion and storage system. It's one of the things that sets us apart from a lot of other observability tools, and so, uh, we need to have best-in-class infrastructure and, uh, expertise at, at the database level or the storage level. Um, we need full stack product engineers, of course. Uh, and we also need, like, front end and, uh, design engineers. So people who... You know, I like to say that we're building the Linear of observability. Like, it's gonna look and feel amazing, um, because, you know, if you look at Datadog, it, it does not feel that way.

Aaron Epstein 28:37 [laughs]

Sherwood Callaway 28:37 Um, so I wanna, I want to hire engineers who have, who, like, appreciate aesthetics and design and-

Aaron Epstein 28:43 Mm-hmm

Sherwood Callaway 28:43 ... um, and attention to detail. And so that, that's where the front end, uh, and design comes into play.

Aaron Epstein 28:48 Amazing.

Sherwood Callaway 28:49 Yeah.

Aaron Epstein 28:50 Well, Sherwood, thank you so much for joining. Um, super excited to work with you on Sazabi as well. Uh, I think you are gonna build, uh, a massive generational company because you're building something that you know a lot about and that you're passionate about and that you have deep insights in, and I'm so glad you came back home-

Sherwood Callaway 29:07 [laughs] Yeah

Aaron Epstein 29:08 ... to the thing-

Sherwood Callaway 29:09 Yeah, thanks for finding me [laughs]

Aaron Epstein 29:10 ... yeah, that you know, you know, you know so well, and really is the space and the idea and the business that you are the exact right person to be building.

Sherwood Callaway 29:18 Yeah. I hope that, um, I hope that other founders can kind of learn from my, my story and, and find whatever their, like... Uh, I've heard it referred to as a tombstone company.

Aaron Epstein 29:26 Mm-hmm.

Sherwood Callaway 29:26 So I want other developer, or other founders to find their tombstone company-

Aaron Epstein 29:29 Yeah

Sherwood Callaway 29:30 ... um, and to come do it here at YC.

Aaron Epstein 29:31 Yeah. Amazing.

Sherwood Callaway 29:32 So.

Aaron Epstein 29:32 Thanks for joining.

Sherwood Callaway 29:33 Cool. Thank you. [outro music]