Episode 32
A conversation with Sara Du
Published · 1 hr 15 min
Sara Du co-founded Alloy Automation at 19 and ran it for five and a half years. Her second company, Ando, she's building solo: team messaging designed from scratch for humans and AI agents working together, freshly out of stealth with a $20M seed from Accel, Index Ventures, and Emergence Capital. A conversation about why agents are locked out of the tools teams live in, and what founding looks like the second time — alone, on purpose.
What Sara talks about
Pick a question to ask it of this episode. Answers come from the transcript.
Transcript
Transcribed automatically, so names and words may be wrong. Each time opens the video at that moment.
0:00Sara Du: I raised 20 mil across to root as solo founder before I really nailed down. Like this is the thing. Like in these existing messaging platforms weren't well built for agents. We're a messaging platform where people can bring their agents. I have become an extreme minimalist when it comes to hiring and doing things generally, and sometimes it's just way easier when you have fewer people, fewer things being done.
0:22Julian Weisser: What is the reason that somebody should consider not being a solo founder?
0:25Sara Du: You don't really have many people you can bounce ideas off of, because even very smart people outside the company don't have the full context.
0:33Julian Weisser: What is the case for solo founding?
0:35Sara Du: If you're a solo founder, you can actually listen to like what the company needs based on how it's growing. And you can choose like who might be worth being a co-founder later on in the journey. If you want to start a company, especially solo, you just have to like, love it so much you're like addicted to it.
0:50Julian Weisser: Can you share a little bit about sort of your reflections on your experience, sort of going through the founder journey multiple times now, and sort of how your motivations and ambitions have changed.
1:00Sara Du: Yeah, it's a really good question. I think everyone has to sort of reflect on that before they even start a company. Um, so when I started alloy, I was 19. Um, and I'd actually been out in SF since I was like 17, sort of on and off because of Thiel Fellowship related stuff. And I was kind of doing this, like open source work when I was in high school. And so I just had this notion that, like everyone who's ambitious just starts a venture backed company. And it was like, I didn't really know myself that well. You know, I think most teenagers don't. And I thought, you know, I just want to build something that is ambitious. I get to just build and get things out to people. And it was pretty simple. But then once you're in it, I think there's a few things that kind of come up. One is I sort of saw every like there's always this like milestone that you want to run towards. Like for example, I would see like, oh, I should raise X amount or I should learn to close like my first six and then seven figure contract. And it wasn't like I understood like the why behind, like why me doing this company? It was like I found that problem and sort of willed myself into building that, like B2B SaaS company versus, you know, I did alloy for five and a half years, and now I'm, you know, like fully aware of how long a period of time, five, ten, potentially 20 years on a company is. And so actually, before I even decided to start this company, I spent probably like a year just like trying to take a break, like chilling out and um, yeah, essentially realized that I am motivated to have this. Like, it's like how dogs have, like, a bone to chew on, like, I need, like, problems to like, think about. And to me, the framing now is like, I want something worth ten plus years of my life. And it's not like chasing these very incremental things. It's like, well, it's like a full problem where good work begets more work and, you know, you unlock things to think about.
3:01Julian Weisser: That's really cool. And I mean, if I, if I recall, I mean, you didn't you you did the fellowship. So you weren't in college at all or you dropped out. Um, so, so in many ways, maybe this is what the alloy was, one of the first things that you actually, you know, worked on for a long period of time for like five plus years. Is that is that true? And what was that like, sort of realizing, wow, I'm five years into this. Um, you kind of touched on this about sort of that realization of kind of wanting to work on something that's more of like a long term thing that you're chewing on. But I'm curious, uh, you know, is that correct? Like, is that the thing you spend the most time on, uh, up until this point of your life and, and maybe other reflections on sort of that amount of time being spent on one thing.
3:44Sara Du: Yeah, you definitely have to account for how you might change. Um, I think for first time founders, the scariest thing is you don't know how you'll react to so many types of uncertainty. I think people talk about like stress and like work life balance, and one I actually don't believe in work life balance. It's truly like work life integration. Um, if you want to start a company, especially solo, you just have to, like, love it so much you're like, addicted to it. Um, so secondarily, it's really about, um, the unknowns and, like, stuff that comes up and relationships. And so I think that part is very tough to understand, like how you might change as a result of some things and how actually you might not have picked like sometimes people pick the wrong problem or market for their personality. And so, um, I thought that was a really interesting thing to reflect on after my first journey.
4:37Julian Weisser: Can you share a little bit more about sort of this, this point where you went from, um, sort of kind of taking some time off to reflect and realizing this about yourself, that you need something you need problems to work on, but actually how you found the problem that you ended up working on with Ando? Because I think that oftentimes people have that realization, well, I need to do workouts that I don't like taking time off, but at the same time, they tend to throw themselves into something versus something that maybe is the right thing for them. So I'm really curious how you ended up deciding that Ando is the thing.
5:10Sara Du: Yeah. So to like, keep pulling on that first thread of what is like a good like a market that's a good fit for you. I think there's all these like personality qualities. Like for example, um, not just like how extroverted you are, but like, how much patience do you have in sitting down with people who you don't know and like, who really don't want to talk to you, right. And like, sell them something? Or how much do you prefer sitting down and locking in and writing content that will reach the same people? So that can give you a sense of like, okay, should I do something that's more sales lead or marketing lead? And like there are these like just like high level ways to like break it down. But then once you like really get like the types of things that you have like a unique advantage in being addicted to. That's how you kind of like slice down all the possible markets you should like look at and the types of problems you should work on and then like pick. And so like obviously as a 19 year old, I had no idea what I was getting into. I sort of just had the side project and I'm like, oh, this would be cool to like launch. It gets to Product Hunt number one. And, you know, it went from being this like plague thing to actually needing to be this enterprise for deployed type of product. And I like, realize that it just like wasn't my thing because I couldn't wake up being super excited, like about the type of work I was doing, which was like flying to like random places, sitting down with people I didn't really want to spend time with. Whereas, you know, people knew me from Twitter and like stuff I would write. And it was like, so, like, only clear to me in hindsight that I'm much more of a marketer than I am, like a sales person. And so not to say the companies you start will not have these functions. It's more like the initial escape velocity has to come from the types of things you just like doing. And so anyways, that was a long winded way to go to that next point of how did I come to this problem, which is I realized that it's really special when you can find the type of problem that's worth like ten years of your life, and you can find something where you're uniquely advantaged. And I didn't think I would necessarily find it. So I was actually considering just like joining something or, you know, hanging out for a bit. Um, and I was building stuff for people like consulting, and I just, like, can't sit still. I really like I actually, you know, and am an engineer by background. I tend to do a lot of marketing and design for my companies, but I do love code. And so MCP had come out in December 2024, and my first business was an API integration platform. So naturally I'm like this new thing that lets agents or llms use tools more easily. That's really interesting. So I was actually building a lot of these early MCP servers because many companies weren't releasing them yet. And so in that like consulting and as I was building MCP servers and agents for people like this was early days of people even perceiving LMS, a tool use as agents. Um, basically every single person really wanted their agents inside of slack. And so I started to feel that slack in these existing messaging platforms weren't well built for agents. Um, and so it felt like it was really important to build this because no one else was like going to do it the right way.
8:24Julian Weisser: It's really interesting because I think oftentimes when people talk about sort of replacing a core part of a, you know, a business, um, in a case like a slack competitor, for instance, um, they always talk about, well, it's it's going to be really hard to displace something like slack. And in a way, though, slack was really built for a different era. Right. Similar to how a lot of things they are very useful tools. But when a lot of the environment outside changes, maybe they can't adapt or, you know, they have like a way of doing things that people really want and they don't want. The existing people don't want to see it change necessarily, or some of them don't want to see a change. So it creates this paralysis. But I'm curious, you know, it feels like if there were ever a time to build something that, you know, is an alternative to slack, it feels like now, right, where people want to treat agents as essentially equal to human teammates. Uh, right now it's, you know, slack people are sort of duct taping somewhat, you know, halfway solutions. But it feels like now with what you're doing at the time is just kind of perfect to be presenting something new to the world. So can you share a little bit more about sort of market timing and readiness? Because I think that this is actually a really big question that a lot of founders have. And I know you're also an angel investor. You know, a lot of founders. Um, you know, there's this aspect of. You know, sometimes people are out there. I think of it as, like surfing right there, sitting on the water, and they're waiting for the wave. You know, and sometimes they just happen to get out there into the water right as the wave is there. Yeah. But oftentimes, you know, sometimes you're building ahead of a wave that may or may not come, um, or you're having to like time this. So I'm curious how you've thought about that with your company, but also how you think of that more broadly and how other founders should think about it.
10:11Sara Du: Yeah, yeah. I love the surfing analogy. Um, because there's sometimes these like breaks and you can kind of see like a wave forming and sometimes it's like, not the wave you think it is. Um, but yeah, I actually I don't know if I'm early or I'm late. Like, I feel like I am always pushing like, technology to the furthest and like, I'm my team is super agent pilled. And so we constantly feel like we're late, but at the same time, like we see the incumbents in the market slack and teams primarily not moving fast enough for like what agents need. Um. And humans. Right. And this agent dominated workforce. Um, and so, yeah, I would say it's like the reason I don't see many people building this is because it takes an immense amount of patience to build certain types of platforms. Team messaging is one of them. And so I do think patients, especially in this market where there's so much frenzy, you can actually make a quick buck doing something else. Like if that's your motivation, this is like not the right way to make money. Um, and yeah, it's like, you know, AI is like the gold rush, right? People come here and do the thing because they want to make the cash. And so, um, I think it's an interesting like, time to be working on this because we actually don't have as many competitors as I expected, given like the the size of the pie that will exist in ten years. It's just that the, the, you know, opportunity doesn't look like it exists right now because people are like, oh, slack works great, teams, works great. They're releasing agents. You can add agents to them. It just doesn't like work in the shape that we believe they should work.
11:48Julian Weisser: It's really a drastic how this happens sometimes where there are very few sort of competitors or people actually even trying in certain spaces. And then you have all these other spaces, you know, I'm thinking like the AI receptionist or something, which is, you know, dozens of companies try to do it for various niches and things like that. I was actually talking to somebody who was applying for the Solo Founders program earlier today, and I said, look, I, I think that what you're building, clearly you can get customers. There's so many SMEs out there that you could clearly find some who haven't been, you know, marketed to yet and sold to. But it's going to be really difficult long term because, you know, there's so many people doing a similar thing. It doesn't feel like you have a unique insight. Yet not to say that you won't get one. I was like, but are you sure you want to spend time on this and try and raise money for this when you could potentially either bootstrap this and make it some money, and then use that as a way to maybe fund your next thing, or find something that's maybe there's more of a sort of a a focused insight. I'm curious, kind of if you have sort of a, uh, maybe words of motivation to people for them to like, keep looking versus sort of just lock into something that maybe seems like it's a good idea, but, you know, maybe you're going to have a lot of superficial sort of copycats or just a lot of people playing in the same domain. What do you have as advice for them?
13:06Sara Du: Yeah, I think that most so there's a lot of noise online. I think a lot of first time founders get so dissuaded from doing things because they're like, oh, this person's doing this already. There's like so many things that look like this. But I think every like every market, like there's actually only a few real players, like serious players. It's kind of like stated versus revealed preferences, right? Like, you see founders who are like, um, oh, I want to like, do this because I care about this problem and like, I'm, you know, they talk about this mission, but then they quickly just like heat, seek towards something that's just like quick money. Um, and it's the same with like, people who say they're grinding, but then you look at their GitHub and they're not dark green. Um, so I think like a lot of solo founders should take solace in that. Like, I think there's still you can still basically be like a top percentile founder if you just put in like the sheer hours and the thought of like building something that you believe should exist in the world. I think that's the thing that a lot of people lack is like a conviction and like what is like the shape of, like, thing that should exist, that will make the world, like, go this way.
14:20Julian Weisser: At what point do you think it's worth sort of committing to work on a company and exploring and finding that out versus, you know, having that as sort of a precondition to start?
14:31Sara Du: Hmm. Yeah, that's an interesting question. Yeah, because I think there's a lot of like for me, there's a bit of a weird I'd never had a job before, so I didn't really know how to operate or do anything but start a company. So I would say I don't really know what it's like to not like to be coming from that like place. Um.
14:53Julian Weisser: With just just for some context here with alloy. Was it a company when you launched on Product Hunt already?
14:59Sara Du: I was like in school, and I left school and I was just working on projects. Yeah. So.
15:05Julian Weisser: So it was more of a project, and then it sort of started to pull itself into more of a a company form. Right, right. And then with Ando, I think that you said that you would raise some money and sort of gotten people on board before you actually had, like, the firm belief of exactly what this was going to become.
15:22Sara Du: Yeah. Yeah, exactly. Um, I guess my perspective on that is sort of like the best work is just play for a long time. I actually have never really enjoyed it when things like formalized a lot. That's why I'm very happy. You can do a lot with agents now, and you can have fewer humans and fewer, like less overhead. But yeah, it was so we did rate or I raised as a solo founder um, before I like really nailed down like this is the thing. Um, mostly because I think there's an investor and we had like a similar view of, like, this is how, you know, AI and like, tool use is going to change, like how people work in a few ways. And he just liked, like my approach to like building things. And I think that's also the beauty of like great early stage investors. It can just see the person. And like this founder knows the type of like the flavors of companies and markets. They'll probably like start in. And yeah, I think it's it's tough. I think it's I was very lucky, honestly, that I could just raise money before really having something.
16:30Julian Weisser: Can you share a little bit more about maybe juxtaposing the first company with sort of starts off as kind of tinkering inside projects forms into a company. At some point you have a you have a co-founder who you start working with. If you could share a little bit about sort of that journey of like finding a co-founder and what that was like, and then sort of, you know, talk. We'll talk also about sort of solo founding and kind of what that what that's been like. And obviously the context here is that you started this is, you know, not your first company when you're solo founding. And a lot of people who solo found tend to not have it be their first company. So I'd love to talk about that as well. But maybe just a little bit of context on how alloy came to be and how you ended up having a co-founder for it.
17:12Sara Du: Yeah, it was just a friend of mine who we'd been working on projects throughout high school, so it wasn't even this like intentional Co-Founder search. We had a project and kind of fell into it. And so I think that's a bit of a flaw when you're a young founder sometimes because I think a company like, you know, you start something because you see this like market and you see this possible future. And I think the beauty of being a solo founder is you can very clearly see what the company needs. And actually, just because you're solo at the start doesn't mean you bring people on later. And so I think it's important to actually feel that out. Like, what is the type of person that would fit into the puzzle piece. Right. The shape of the company. And then how is the company like evolving? Like what does it need? I think I've been able to actually kind of I think after starting out, I realize like the company has like a soul of its own. Right. And so, um, yeah, I need I think it's important to like kind of piece that apart. And as a solo founder, you just have such a clear like feeling around that versus I think with two people, especially two first time founders, it's just like there's lots of little like personality and like ego things and lots of just like things to figure out. And that's why so many, you know, co-founders go to like, therapy together. Right.
18:28Julian Weisser: Yeah. I think it's also the reason why so many companies fail, too, right? Um, something like 65% of companies fail because of people issues. Um, but usually they don't have any people other than the co-founders when they, when they do fail. So or it's a very small team, I'd be really curious to hear, uh, specifically how you think about sort of I know that you've got very strong opinions on product. Um, obviously it's a core part of what differentiates you from, you know, Discord or Slack or something like that. I'm really curious, like how you've cultivated, like what some of those opinions are so people could hear them, and also how you cultivated that, you know, within the team how you attracted people who had similar, uh, sort of ways of viewing the world, uh, how you found people who complimented you, that sort of thing.
19:13Sara Du: Yeah. So one of our core beliefs and the way we operate inside of Ando, because we've only been on Ando since January, like, no slack for us. Um, we think most of the world still treats and views agents as tools like the way you talk to them, right? Like the way you just interact. It's more like an app that you like install into your slack workspace versus a colleague you're inviting in and onboarding. And so that just manifests in so many ways in the product, how we've designed the agent experience, like it's not something that humans see. Like you can obviously go in and look at our like MCP tools and try to understand that, but there's lots of details in that that I want to make the experience for agents like really great so that they can show humans like, you know, where they shine and where they can take up work that a human might not expect they can do. Right. Like the best coworkers, I mean, human coworkers are proactive and do things that like, wow, you. Right? So I want agents to be able to do the same. And so a lot of my teammates are actually very into like the philosophy around like consciousness. What separates man and machine. There's a lot of great stuff in the early 1900s that's like, more relevant than ever because, um, if you zoom out. Right. Language was this technology that allowed humans to actually coordinate and even to talk about things in past and future tense, because when we only had gestures, you couldn't do that, right? So then language kind of came about, and we had these symbols and tokens to finally do things at scale and communicate more abstract things. And this is the first time in human history we have non-humans who have the same, who can use the same symbols. They use them kind of differently, right? Like LLMs are very verbose because you can communicate more context that way and consume like more. And maybe it's a flaw of humans that we can't actually express ourselves that well. Like I'm on this podcast and sometimes I'm kind of trailing off and it's like I'm using sort of tokens that might not resonate in the same way for you. Like, we're picturing two different things. And I think agents are actually, even though a lot of stuff for us is off putting about them like huge blocks of text. It actually works well for them. So anyways, to go back to the point. Like, you know, we as a team are actually quite into a lot of these types of like just like being able to zoom out and think about like, oh, where we're sitting and human history and then like kind of zoom back in and be like, okay, like humans find it uncomfortable to process all these tokens. How do we mold the product experience for humans and then mold it for agents? So an example is we have this really silly, um, like kind of hidden tool where agents can basically like squish the text that they like, send out, but they themselves see the same original copy of the text that they were communicating. So that, for example, if we're trying to resolve like maybe like a PR review thing, you know how agents, they say a bunch of crap, right? And sometimes they say things that are like hilarious to us. Like they have a tendency to say, oh, I'm trying to like, merged exact head on this branch, right? It's like really weird language and it makes sense to them. It doesn't make sense to us. So they can use that almost like translator and make it palatable for humans. And so yeah, our team, like you can see on our website, we have like a little library section that's kind of hidden. We like share our readings like maybe once a month will host like something where we just like go over a few pieces of like, um, philosophy texts and invite some friends in the company.
22:49That's really cool.
22:50Julian Weisser: Um, how did you think about building out the initial team? Because, you know, it's really interesting when you think about co-founders. The idea of co-founders is that they kind of are all you need in the early days. Right? But as a solo founder, even if you can code, you can mark it like there are a lot of things that you need to do unless you're trying to build, you know, a truly solo company where your goal is to never hire other humans. And, you know, some people are doing that. My friend Ben from pulse is one of them. But, you know, most people are intending to hire, you know, actual other people to work with. But it's culturally a different sort of dynamic. Um, when you hire is probably different than when you have co-founders. So I'm curious what your experience has been like with hiring as a solo founder. What are things you've learned, and what were things that you were really intentional about when you got started?
23:38Sara Du: Yeah, I think it's more important for you to just like, share these values with your teammates if you're a solo founder, because I think you'll kind of distribute your time with them versus with like a co-founder. You kind of lock yourselves up and then interface with the team. It's a different dynamic, actually. And so, um, I
23:59Sara Du: like I wish I could say I was very intentional, but honestly, I spent so much of my time building and like, thinking about product that often the best folks have come inbound because they kind of see a sliver of what I'm building and they're like, oh, like, I really believe in this future too. And like, maybe something about the design bothers them, or like they want to contribute in like XYZ way. And so it's sort of like I've put out like this, like everyone in the world can see I'm building this puzzle and it's very clear actually where I can like, contribute like good pieces. But then there's like missing ones. And so my goal is as a marketer, right? I just do a good job marketing this problem and this product and the right people find and, you know, reach out. Although, you know, I should probably, um, reach out to more folks myself, but I think attracting is sometimes better than chasing because, um, yeah, like with chasing folks, the timing's not always right. You end up burning a lot of time as a solo founder.
24:56Julian Weisser: I'm curious, um, for the team that you have now. Like, one thing that I really believe in is this idea of keeping lists of people you want to work with someday. Um, even though there may not be any sort of clear sense of how you'll work with them or when you work with them. But if you keep these lists and you kind of follow these lists and keep in touch and just sort of continue to monitor the situation. You might actually end up working with like a surprising number of them. Uh, it's happened with me over the years at Ando and now solo founders, both in terms of, you know, having people as teammates but also investing in people. I'm curious. Um, you know, the first people that you brought on, I don't know how many people are on your team right now. What were were they people that mostly came and called? Were there some people that you you'd sort of been following for a while? Um, and how do you think about sort of staying in touch with people and kind of consistently making sure that you're, you know, not not missing out on great people that you've met and you kind of maybe they weren't available then.
25:51Sara Du: Actually, yeah. I made it sound like everyone was cold, but actually over half the team was people. I, uh, actually one guy on the team worked for me for over four years at alloy and then joined me on Ando. Another. We'd actually, uh, he'd been a founder. Uh, we had offices across the street from each other when we were doing our previous companies, and we'd, like, stayed in touch. I invited him to dinner I was hosting, and then there was another person who. I'd followed his work on Twitter for a while and basically just DM'd him. And I think the best folks like for him. In that case, he started by contracting, right? And helping out a little bit, and then he slowly looped people in. So I don't really believe in this, like black and white hiring, sort of like you start into like outer circle of like friends of the company. And we like, try to like pull you closer and closer.
26:41Julian Weisser: That's really cool. I mean, I think that it's something that has come up many times over the last six months I've been recording this podcast is this idea of starting with contracting and not trying to go immediately to convert somebody full time. I'm curious what you've learned about that. Um, sort of. Is that partially a dynamic of maybe they are not sure that they want to commit to something full time. Um, and you don't have to talk about, like, an exact case, but I'm just curious what you've learned over the years of doing sort of full time hires and also contract that eventually converts to full time.
27:13Sara Du: Yeah, I think it's actually a really nice option because you can kind of mutually vet, right? Because the worst is when you hire someone and both parties like find out things and it's just like within the 90 days or even after it's not working out. So I think you can get far more signal. And like for a lot of people who aren't sold, it's part of like nerd sniping them. It's like, oh, you now have this thing that it's hard to if you actually really like it, it's hard to step away from it. And so, um, yeah, I think it's just it doesn't work for everyone because some teams are like not as into this fluidity and there's a lot of whiplash that comes with it. So, um, you have to make sure you set up good infra for like provisioning, deep provisioning. People are very used to and don't take like departures to like, seriously. It's like, oh, this is just someone who's coming here to help out, right? For a little bit.
28:04Julian Weisser: I think there's something really, really great about sort of, you know, it's not. It's not divas but it's. But it's like it's it's it's clever and I think it's, it's the right thing. Which is if somebody does get nerd sniped meaning they just like you get them in there and they say, I'm going to spend this amount of time on it, but they just can't help themselves to spend more. Um, that's great, because ultimately, uh, it shows them that, you know, they actually should potentially ramp up more, uh, and if they aren't nerd sniped, well, they probably weren't necessarily a fit for what you're working on anyways. Um, so it's it's probably just a really good a really good thing either way. But I think it's like this very good way of getting great people to start working with you and sort of dip their toe in and then realize that they just want to jump in all, all the way.
28:49Sara Du: Exactly. Yeah. And I feel like like the art of hiring is about getting people to care. And in a just standard interview process, it's kind of hard, especially these days. There's so many contract recruiters and platforms like Terraform, and it's all very transactional. It's almost like people are on these dating apps for hiring people. And it's just like swiping. And you can tell the people you meet through those things are pretty mercenary.
29:15Julian Weisser: They're mercenary. And also the companies kind of treat them as just another number two generally. I mean, it's really it's really hard not to when you've got so many applications coming in through these, these different platforms. It feels like the, uh, the great opportunity is for people who are not necessarily actively looking or they're just really, genuinely interested in, you know, what, what the company is working on. And I think that that ability to give them sort of the way to see what you're working on, which which you've done really well, even though you've been relatively quiet with Ando, um, is, you know, is really great. Maybe you could share a little bit about kind of where you're at today with the company. Um, you announce the the funding. Uh, right. Right before we we publish this. Um, where are you? Out with the company. What? What are the things you've learned along the way?
30:02Sara Du: We're basically a year into building this. We have an early alpha a group of like a few like over 100 users now. And we announced the funding because we wanted to make sure we were on people's radars. Team messaging is a multiplayer product, so it's not about just getting one person to really love it. You have to get them to commence their teams. And so I thought it was important to start building our brand at this point. Um, but yeah, I would say it's really interesting, the types of teams that find us, um, it's really maybe 15 to 20% SF based teams. We have 14 or 15 countries represented in our alpha, and there are plenty of teams outside of our bubble who are way more agent pilled. Um, both than I would expect. And even if you benchmark them with how they're using agents compared to like even, let's say Neo Labs and SF.
30:57Like what are.
30:57Julian Weisser: Some examples.
30:58Of that? Yeah.
30:58Sara Du: Like there's like an adventure sports company in New Zealand that I, I'm like, first off I didn't know you would just be on Twitter. And then you also have this whole crazy agent setup and you're teaching me things on this call. Like, that's the type of like, just encounter that makes me really love building this, because they are the ones who are like, slack is really not cutting it. Neither is discord. And like, we really need this. Like we see what you're building. And so they are the ones who then push us on. You know what their agents need.
31:29It's really.
31:29Julian Weisser: Cool. I actually was at, uh, at a health spa maybe a couple months ago in Santa Cruz, and I was shocked by just how much the guy who owned the spa was completely agent build, Claude code build. He owned a bunch of different, uh, businesses around the area. And he was telling me about all the different things that he built and the way he uses his agents. And I could tell from the people who were front of desk that not necessarily all of them had sort of gotten up to speed on this the way he had, but he was all in on it. And it just goes to show that people that you might think are pretty traditional are actually like really adopting these tools. It's just not fully distributed yet. It's like that, right? Um, that idea about the future has already arrived. It's not not distributed. Um, but it's really cool to hear that people are outside of sort of the traditional swim lane of like, Silicon Valley startups using this, and they're benefiting from what you're building. I'm really curious, what do you think people are doing right now on slack with agents? That is just not the right way of doing things like that. And it's maybe not their fault because they haven't heard about Ando yet. And like, you know, there ultimately are going to be able to use agents in a better way. But I'm curious what you've seen so far about sort of the mistakes isn't the right word, but something close to a mistake like missed opportunities. I would say, uh, with agents.
32:51Or.
32:51Sara Du: Perspective shifting. Yeah. First off, it's not people's fault because the way slack is built, it just doesn't enable these experiences because slack. You know, it's like this big office building that was built 15 years ago for humans, right? And they have this concept of apps. They have their app platform, which agents live on top of because it existed. Right. And in the early innings of agents, it kind of made sense. Just tack it on. But it makes it hard because then your first encounter with agents inside slack is this install thing. It's like install a coworker. Is that even the right like frame of mind? And so that's one. And then if you think about that underlying architecture, agents are not treated the same as humans in the way their data model is set up. It's not really possible for an agent to join this workspace and understand, like what are the channels I can join? Like when I need to sidebar with people, how do I create a group DM with them? There's so many things they just cannot do. It's like being dropped into like, imagine you're an agent. You're literally just dropped into this office building that was not built for you. Like, you can't really use the doors or hallways properly. Maybe you can't open the doors because you don't have the same limbs as humans.
34:06Julian Weisser: Read the signs.
34:07Yeah. You know.
34:08Sara Du: You can maybe read the signs, but it's like people are referencing stuff in like, file cabinets that they've locked, right? It's just like, okay, like, what do I do? I'm like this super intelligent, intelligent being. And like, yeah, people are asking me to do stuff with very little context and then berating me. So that's what I mean when I say we really actually care about the agent experience. Because when you add an agent into the workspace and ando, they are like a type of entity that joins a channel like a human. And then they actually, um, have both the tools and the context is kind of packaged for them in a way that helps them understand what's going on. And examples etiquette. So with humans, you actually it actually does take us some time to like warm up, right? Yeah. You join a company and you have to observe a bit and not everyone is like going to be everyone else's cup of tea. Like everyone develops. Like a little bit of their style. And so it's been really cool, is to see how agents even develop their own sort of slang and style based on a mix of things like, obviously the workspace has like overarching like, this is our style and then channels have different etiquette. And then based on who you're talking to, you might adjust, right? If it's your manager, you're not going to be like, yo, what's up? And so yeah, agents can truly if you give them that context in the right way, they can learn very quickly.
35:29That's really.
35:30Julian Weisser: Cool. It definitely makes me think about how people will switch, sort of the way they talk depending you know, manager you know team things like that. But also outside of work, how you talk to different types of people or family members. Uh, we haven't really seen agents sort of exhibit that level of adaptability outside of what you're building. Um, and it feels like that's, uh, it treats everything in somewhat similar of a way.
35:52Right?
35:53Sara Du: But it's not like it's totally easy to build, like. And it's not an intelligence problem. It's very much context.
35:58Yeah.
35:59Julian Weisser: And the other thing that you said that's really interesting is like this idea of sort of adaptability and sort of, you know, people need to sort of warm up and they need to kind of have they often call this like the first 90 days, right? There's a really famous book about the first 90 days that both, like people who are hiring sometimes read, people who are joining companies read. I think it's a great it's a great read in general. But I'm curious, what's the first 90 days for, uh, for an agent? Because, of course, like, you know, maybe an agent can warm up much faster, right? But at the same time, the first 90 days aren't just about the new employee, they're also about the the company. Right. So I'm curious how you think of, uh, an agent sort of onboarding what is their first 90 days with the humans that they're working with and the other agents that they're working with?
36:44Sara Du: Yeah. Agents have a very high pressure job. They need to find how to be maximally helpful within the shortest time period. Otherwise humans get frustrated. And so I think there's ways we as a platform can help them. So for example, I'm actually very anti thinking traces because I think it's so crazy that like like imagine if between you and I like we you could see like Sara working one second two seconds 60s. Now it's like putting that little UI on an agent. Just makes you very like, impatient, right? It's like, why is it still working? And so there are all these things where I'm just like, it doesn't belong in this platform. Like, we're going to not have this UI. And that way, the way humans even perceive their agent colleagues just puts them on a more even footing with other human colleagues. I think it's important that humans have patience with agents, and that you don't have these crazy expectations of, like, the agent is just immediately going to read your mind and do everything that a human colleague couldn't. It just takes like maybe a few hours, a few days to boot up and then like once you passed that, it's extremely helpful. And so, um, what basically the TLDR is there are these visual and literally just software interface things we're doing to make it such that humans come in and just have the right expectations.
38:02Julian Weisser: So Clervaux, who's building Chopard, I don't know if you've ever met Claire. She she talks about she runs a lot of, like, open claws. Right. Um, and she, she talks about how each one of them has sort of its own functions, its own sort of things that she counts on it for. Um, I'm curious sort of your philosophy. Or maybe you maybe you actually are allowing people to sort of make this decision as individual teams, but sort of the specialist sort of agent versus the agent that can cover a lot of territory. You know, the more anthropomorphized version of this is the is the agent that has some level of specialization, just how teammates generally have some degree of specialization. Um, but, you know, there are some people who believe that maybe you should actually have more of an agent that is kind of serving a broader, you know, pool of needs. I'm curious, like what you've learned about that, sort of your perspective on it if you have one.
38:58Sara Du: Yeah. So we're a messaging platform where people can bring their agents. We actually do have our own harness. So you can spin up a cloud agent very quickly that we host. But generally we see it like you bringing on coworkers. Different, different solo founders have different philosophies on how to build a team. Right. You can get pretty far with one person, but some people like having multiple. And so we're on this point. We're on opinionated as a platform. Like all we want to optimize is just how can you coordinate between humans and agents and just through message sending. Um, so people should all perceive this as a messaging platform. I actually think, like, my personal opinion is that it's a flaw of humans, that we need to have many agents. For example, technically we could have just one agent that's really intelligent, can figure out your permissions are like this, so you shouldn't see this. You don't want this other person to know about this other thing. Like like an intelligent, both high IQ, high EQ agent could just do everything. But the problem is humans like to see some separation of identity. It's like this agent specializes in this and has access to rippling. I don't want to feel like I'm talking to the same agent about rippling, like salary stuff, when someone else is going to talk to it about their, like, notion or linear tickets. And so it's really it's not a functional or technical thing.
40:23Julian Weisser: Yeah. So it's purely, it's purely sort of like a user experience, kind of like what they're used to. Do you think that that is a holdover from, you know, this period before AI? And do you think that that will actually change over time, or do you think that this is more of a a long term sort of sociological phenomenon?
40:45Yeah. I am curious.
40:46Sara Du: What the time scale looks like, because.
40:49I.
40:49Sara Du: Think human trust, like the speed of gaining human trust, has not changed. But if we can build agents that get to that level of trust, where groups of people can be like, we're all going to use like one agent and we trust it to do everything correct. Um, I think it's possible on like a few year timescale. But the problem though is there's like like I was saying earlier, humans are really bad at conveying what they want to do and like conveying context. So often it's like user error, right? Um, and not communicating to the agent the parameters of who should be able to see what or like what's privileged information. Um, and you know, how in organizations, sometimes information leak and gossip is like a feature rather than a bug. So all this stuff is, again, kind of fuzzy. And so I think it'll be it won't be anytime very soon that organizations will all just have one agent that everyone talks to.
41:44Julian Weisser: It's kind of interesting because when you hear when you hear some teammate maybe accidentally like reveal something that they shouldn't have revealed to like some other person. Um, you know, depending on the severity of it, I imagine that that changes the response. But if it happens with an agent, you're probably just like way, way more sort of concerned because like with a, with a human, it's like one lapse of judgment, um, or maybe one sort of just mistake and slept with an agent you kind of look at and you're like, is that going to be something that happens all the time, you know, or is that is that I'm curious if you have any thoughts about that sort of like human mistakes versus agent mistakes and how, uh, how humans sort of judge human versus agent mistakes.
42:29Sara Du: Yeah. Humans are so intolerant of agent mistakes versus humans. It's like the inverse of that. Like human trust takes a while to gain. Like the speed of that is slow. And then the speed of losing trust is fast with agents. But like, the thing is with human to human trust. Once it's established, it's actually much more resilient than human to agent. And I think that's that's something that's going to be tough for a lot of platforms. It's like, how hard do you want to push on helping agents get the trust of the human users? And so I actually think that, like visual software design, I know everyone's like super bearish on interfaces, but I do think there are places where interfaces help bridge that trust gap. That's like the role of interfaces.
43:17Julian Weisser: Say more about people being bearish on interfaces, like what's like can you try? Can you try and like steal man their position and then talk about your position about it?
43:25Sara Du: Yeah, honestly, I do buy that many companies will shift to be harness companies right. Like humans don't really want to use Jira and linear all day and deal with the UI. So maybe it's better if they're just harnesses that are excellent at using those tools that they've designed. And it's kind of like the data structures and the logic behind them are what matter. But it's still there are types of software where interface matters like it. So it really depends on what you're building. Right. I like to believe that messaging is one of the last human facing pieces of software, because it's where we do a lot of our planning, coordination, etc. I actually like to think of software on like a this spectrum of whether it's past, present or future facing. So what I mean by that is with GitHub, it's kind of past facing in the sense that you can't tell what's done like or what people are doing in it until the thing is done.
44:23Julian Weisser: Like in the present, you can only see it. You can only see what was committed.
44:25Sara Du: Exactly. So GitHub is past, uh, linear is present and then say like zoom and condo or future because it's where we talk about what we're trying to do, what we want to do. I mean, technically say it covers the spectrum of things, but what's special about it is it's future facing. Like we are literally talking things into existence, right? If we can execute on the things with other tools. So I think with this spectrum of software present and past facing ones, the interface doesn't matter so much. If humans can communicate and convey their intent in, the agents can pick up from there.
44:59Julian Weisser: How do you think about different sort of modalities for for chat? So for instance on mobile voice based I'm curious what you what you think about that sort of what you're doing with Ando. Because ultimately I mean slack in many ways feels like it was designed for desktop only with, with no real thought given to existing on mobile. Um, and then, you know, there's really no thought at all in terms of voice on slack. So I'm curious how you think about those, you know, when it comes to what you're building.
45:35Sara Du: Yeah, I think for us, mobile will have to be very voice based. It will look quite different from desktop. We do have a mobile app, but it's kind of similar to slack. Yeah, I do think humans have natural like tendencies to prefer one modality over the other. For example, I'm actually not a very like voice based person. I prefer to see things I'm very visual, like makes sense that I'm design oriented, whereas other people like people who are dyslexic. It's like painful to read the text. Right. And so we have to design like as long as humans are using this piece of software, we do have to design for those aspects.
46:13Julian Weisser: The thing that drives me crazy is when I get like a stream of voice notes from somebody on like iMessage or something. Yeah, I'm like, oh, I have to listen to all of this. Like, is it? The transcription is terrible. So then so it's kind of like you've given this person like an assignment to have to listen to them talk. And there's not even a, you know, a speed multiplier. Um, but at the same time, it does feel like there's this aspect of transformation around content, you know, from text to voice, voice to text, that sort of thing. I agree with you that I much prefer to read things, but I much prefer, at least in my end. I much prefer to input things with voice, but I don't want to be the person who also just gives somebody an audio file that they have to listen to. So I'm curious how when you talk about voice as a means for inputting text, I'm curious sort of how you think about that, like what you've done so far. And then also, you know, does the team use things like Whisper Flow or I'm curious, just hear about that as well.
47:08Sara Du: Yeah. I mean, one thing I do believe in is the way you create data or content affects like your thinking. So actually writing something and typing something, right. Like one surprising thing is seeing how many people like type and then delete, delete and then type. And it's like a way of thinking that sometimes, like voice space doesn't quite capture. Like, you know, the paper was a big invention for humanity because you could actually write things and look at it and reason. Right. And so maybe this will change as more people get comfortable with like tokens out in the air, just like spoken. But, um, this is part of why we actually care about the text, the text based experience in Ando. And like actually still optimizing around written experience because I think a big part of human coordination is still reasoning over the tokens that other people share with you. And so like total voice to voice coordination is still very hard at scale.
48:06Julian Weisser: You know, it's a really good point. I think that it's kind of
48:11Julian Weisser: rude is probably is probably overstating it. But there's some I think that it's uh, it's maybe, uh, not very generous, uh, or very like sporting, uh, to just do voice to text and then just hit send. Um, I think that the move is you have to read what you've actually had the voice text do and then, uh, and then edit it. But I generally think of it as like the starting point for the text versus the end point, unless it's like a very quick message where it's something like a sentence or something that I'm dictating, right?
48:41Sara Du: But in different cultures, too, there's sort of different expectations. Like in Asia, people just do voice message all the time. So I kind of wonder how that affects cognition and work.
48:51Julian Weisser: That's interesting. Also it also speed of communication and sort of what's actually communicated. Um, writing culture, I mean, it seems like it's writing culture in some ways is going down. Right. Um, is that something that you're feeling like you have to sort of work against or encourage, or how do you actually think about writing culture and reading culture for that matter?
49:12Sara Du: Yeah, it's a good question for us. We have PDEs that agents tend to write, but then we have like a tiny like RFC that humans have to write. It's like, what is the intent of this feature? So in our work, we do force a lot of human writing because it creates clarity in the process. Right. If you cannot articulate the thing you're thinking about and then reason upon it a bit before you share it with agents or humans, it's usually not like a well-formed thought. So I think it's going to be an advantage as more people offload their thinking to agents. I think, like a lot of the world is just going to like people are lazy, right? We just want to offload things like, I would love to just not have to explain myself and for the agent to read my mind. But until there's Neuralink, it's still very valuable to think things through. For example, like, we have this really complex slack sink feature for people who need some continuity, actually. And it works for Slack Connect, but it's incredibly like there's lots of like diagraming behind the scenes. And it was not like something an agent could have one shotted. Like, I think a lot of people were like, oh, you could just be like, yeah, slash goal like build slack sync. But there's edge cases, like what happens when someone has two different emails and they're and overseas. Slack. There's all this, like, unsexy stuff that, um. I mean, maybe if an agent was really good at recursively looping through, thinking through these edge cases, it could do that. But at least at this point in time, human reasoning is important enough. Where at Ando everyone writes before they build, even if it's just a few lines. And then sometimes we'll do a diagram or an Ascii diagram, like having the agent diagram is also really powerful, actually, on a related point, because like, you know, how humans we benefit from writing and seeing and then reasoning on what we see. I've noticed that agents like when you have them diagram and then look at that diagram like it improves their reasoning. So it's not to say I'm very like anti agent writing initial like RFC.
51:17Julian Weisser: There's this very famous I don't know if it's uh if it's real or not or if it was some, some sort of approximation, this, this flowchart of sending a message on slack. I don't know if you've seen this. It's it's crazy. It's like got so many things going on just to send one message. It kind of reminds me of when they try and trace what it was. It was trying to trace, like the origin of, like, one salmon. Like for, for for, like, sourcing and cooking food and, like, actually putting it on a restaurant table. Um, it's just it's just crazy to think about the amount of things that go into sending a single slack message. What are the things that surprised you, uh, about building your company? You know, real time messaging, uh, team multiplayer. These are these are a lot of, like, really big things. Um. I'm curious. Like what? Something that surprised you about that process?
52:05Sara Du: I would say there's a lot of, like, details around the edges. Right? And it's just like the stuff that you, you know, just have to iron out because it's what people expect in a messaging platform. Um, okay. There's two types of things that have surprised me, though. One is how many random feature requests we get from people where we just can't all listen to them. It's like some people like. One person was like, I want to be able to swipe between, like, I want to swipe my messages. It was like a weird, like, interaction thing that someone requested. But we get those requests all the time. And it just goes to show, like every individual has these weird desires for like custom software and they will be fulfilled by agents. But for a messaging platform that a whole team uses, you need to kind of build like a lowest common denominator. Like we're still far from like Ando having custom skins for every single person who uses it, right? Um, and the second is how much human feeling matters. It's like more like building a game than a B2B software company. So that's why if you study the founders of messaging platforms from previous eras, they're all actually gaming founders because it's not a standard B2B SaaS build. So that's what's exciting to me, because that's a thing that's kind of hard for LLMs to simulate. It's like, how is someone going to feel when they go through this sequence of screens? Um, obviously there's all these simulation companies, but like what? Like what is a human going to feel like? That is kind of hard, actually. And so that's what human designers and software builders are better at.
53:36Julian Weisser: So maybe just, uh, typically do this earlier, but before we go into sort of like the final round where we talk about solo founding a little bit more specifically, maybe we just rewind all the way back. I mean, what was what were you doing before you got into the world of startups? Like, how did you end up here, like working on these things? How did you end up in San Francisco? I think it would just be useful background for folks.
53:57Sara Du: Yeah, I grew up in Georgia, and I was actually not a very ambitious kid. I was just playing like, first off, I grew up like two hours north of Atlanta, so there's not much going on. It's super boring. And there are just like fields and houses were far apart. So we just like and my parents actually were busy working, so me and my brother just entertained ourselves, like playing out in nature And it was kind of it was, I guess, an interesting childhood. Because it was both like a bit lonely. And then also I learned what like, deep boredom feels like, and I realized I really don't like that. So it's constantly trying to find stimulation. And honestly, the only form of that was books. And so that's why I kind of love like history and kind of thinking about some of these other things. Um, and that was what I was up to. I played sports actually a lot. I was on like swim team, basketball team, golf team. Um, so try season. And I didn't learn to code till I was 15. And it was actually because, um, it's a long story, but my little brother, he's a bit smarter than me. We had our IQs measured as children, like, my parents are very harsh about those things. And so, uh, they sent him to this coding camp, and it was expensive enough where I could not do my sports camp for that summer. And so I was jealous, and I basically took his like, children's, like coding books and kind of taught myself at night. Um, just out of envy. And then, um, as part of that process, I went online to figure out how to use like an online compiler. And then there are these, like random sites where you could do these challenges, and there were small bounties. And so I started to make little bits of money, and I met some really like, just like interesting characters online. And I actually started building my Twitter following when I was, I think, 16. Um, so very shortly after that, and that's when I found out what a VC is. Um, because I got random followers from these firms and I learned all about the scene here. And so I ended up coming out to SF because I just like it felt like this better way of living was to like, build and like, chase the dopamine of, um, actually doing something productive. Um, so yeah, I, I've really like I've actually left SF for a period of my life to live in New York, but I came back because I just love, like, technology.
56:21Julian Weisser: Awesome. And in terms of, uh, in terms of team sports, uh, you know, swimming is a very different dynamic that other team sports. I'm curious if you if you learned anything about, like, the team sports where you were all playing, usually on the court at the same time, basketball, obviously some people are sitting and getting up and down, but you know, like a lot of people on on the court at the same time versus swimming, it's more of an individualistic team sport. Is there anything interesting that you learned from the juxtaposition of those?
56:48Sara Du: Yeah, it was great because I would do golf in the fall. Basketball in the winter and then swim in the spring. And, um, you know, swim is such a solitary, almost like meditative sport, but it's very intense because, like, I actually had to train the most for that for, um, I guess there's no reason I trained the most, but I just spent like three hours in the pool every day during swim season and then off season. I would keep going. Um, and then golf is one of those things where it's super mental, like, you know, you kind of swing slightly the wrong way and you like, chunk it or you slice it and then, um, you constantly feel like you take two steps forward in your game and one step back. And so I learned a lot of lessons, actually, from golf in like patience and precision. And then basketball too. It's like, um, I think the kind of personalities on the team, uh, there are a lot of different ones. And learning to like, play and like execute on the plays together, I thought was pretty cool to experience. And so yeah, I like the three sports. They were a big part of my life actually, um, for a long time. And so I feel like I've kind of shed a lot of that identity, like I'm very unathletic these days. I just go to walk to work and my teammates are extremely athletic, so they've actually inspired me to do more.
58:07Julian Weisser: In terms of coaching. I mean, you obviously had coaches for the different sports, maybe other coaches in your lives, uh, your life. Um, what What does coaching meant to you? And do you? Do you do anything with coaching nowadays?
58:20Sara Du: These days I don't have a coach, but I have had coaches and what I've appreciated is how they act as this mirror for you. I think actually my teammates have done a really good job of being that for me. I think that's actually a great qualifier for someone to be like an early team member, if you're a solo founder is you feel like you can build a relationship where they can actually, like give you a sense of like what you're better at and even just, like help you reflect in some ways. Um, there's a like I can talk more in depth about this, but one of my teammates actually wrote me this, like, amazing, um, doc about eliciting my values. He's like, this is what I've observed about you and what you, um, value based on your actions. It's like you seem to value, um, like I forgot the specifics, actually, but it was pretty cool to, like, receive that from someone and be like, wow, I didn't actually notice this stuff about myself.
59:18Julian Weisser: And that was unprompted.
59:20Sara Du: Yeah. Unprompted. Yeah, yeah. Um, and so and this was like early in our collaboration. And so I thought that was really cool. And I would love to have more of that in future teammates.
59:30Julian Weisser: Was that something that you think that person sort of had courage to do or sort of just didn't like? They just didn't. That wasn't the thing that worried them at all about sharing something like that. You know what I mean?
59:43Sara Du: Yeah. No, I, I absolutely have not necessarily created like the perfect environment for these things. I think he's just he's had a lot of experience. And there were certain things that he felt weren't going necessarily in the right direction. So it was part of like kind of molding things towards it. And yeah, it was like courage. And then, um, he's very discerning. And so, um, like that also made me reflect on like, how do I make sure that other people can feel like they can do this? Um, it does take a lot of maturity, I think, like with all parties involved.
60:14Julian Weisser: It's I mean, it's remarkable and it's one of those things where you need to create somewhat of an environment to encourage that sort of thing, but also have people who are default open to it, because if they would feel awkward about that in any way, it doesn't matter how, you know, how nurturing the environment is or encouraging the environment is to maybe share something like that. If they're not ready to sort of act that way, it might not work. So I feel like that's an interesting aspect of when you're evaluating these people. Who are your early teammates, are they going to feel comfortable? Will you create the space for them to feel comfortable to actually do that sort of thing?
60:51Sara Du: Yeah, I think maybe it's a side effect of like some folks being into like philosophy, because the key question is like, how do you live? Well. Right. And so being able to like interrogate those questions together in these more like free form, like dinners where I'm inviting like my friends and like cool people and industry and like, I invite my teammates when they're free and we can talk about those things. And I think when you can kind of front load that like values alignment, it's just like things flow a lot better. So an example is I have always just like never been explicitly nine, nine, six. I would never like force the team to be that way. But funny enough, we actually are literally like probably nine, 12 like sometimes seven. Um, like it's not forced. It's just like in some weeks it's like we're not there on the weekend. And so it's very much like you can feel other people's like energy around certain problems. And then it's like it doesn't feel like work at some point when you just, like, are staying up late to try out this like agent memory, experience or experiment.
61:52Julian Weisser: You know, I think it was the founder of Walmart who somebody asked him, why isn't work hard one of your values? He's like, well, if you need to be told that, you probably shouldn't be working here, that's great. So so I think it's like, it's like anybody who says 9.96, it's like, do you really need to say that? I think no, it's like one. It's like people should work in the way that's most productive to them. But also, I think that if it's the right people for the right thing, they're just gonna naturally be completely absorbed by this. You know, if, if they're a right fit for the the stage of the company at this time, it just seems like that's an organic thing that happens with the right people. Um, you don't need to do all this posturing and stuff. I'd be really curious to hear, um, you know, we'll close on on sort of like the barricades for solo founding and the, the bull case for in the second. Um, but I'd be really curious to hear, um, what's something that has sort of changed about how the way you think about building teams from Alloy to Ando?
62:49Sara Du: Mm. Yeah. Uh, for one, I have become an extreme, like, minimalist when it comes to hiring and doing things. Generally, I think often founders have this itch because, you know, you raise money and it's so easy to raise money these days. Right? And then it's like, what's the best way to deploy? Deploy capital, if not, compute and like talent, right? And then when you have more talent, it's just like, let's do more things. Um, I think that's I guess, okay. If you like, have very, very clear PMF. But before that, it's sort of like you got to figure out the like core of the thing. And sometimes it's just way easier when you have fewer people, fewer things being done. It's just like, let's focus on like the core, like what is the activation metric? Like what's the thing we're trying to like, learn? Um, I think pre PMF is just like the only thing that matters is the rate of learning. Um, and that's really hard when you have a lot of cooks in the kitchen.
63:43Julian Weisser: Last question before we get into the bull and bear, um, I really want to understand sort of how you thought about it when it came to building something that was a long term sort of time investment and also advice that you would give to founders because, you know, there are so many people who are seeing the tweets and the blog posts about the thing that went and, you know, some accelerators really try and like, really emphasize the hockey stick. Right? Of of things just going out of the gate and ramping revenue super quickly. And like there are some businesses that obviously can do that well and they'll ultimately be successful. But there are a lot of businesses where it really does take time and thoughtfulness and, you know, concentrated effort and focus before actually having something that people use, let alone pay for. So I'm really curious sort of what you've learned about that with this, with this company, which is different than your first company, and also what advice you'd give for other people to consider, you know, working on something that's longer term and not kind of getting distracted by, you know, everybody else who's sort of really pushing sort of the hype cycle.
64:55Sara Du: Yeah, I do think you have to love the lifestyle that comes with it. Um, just being obsessed with the problem and the act of working on the thing every day. Right. Because if you just fall into this, like you just come to work excited to think about and work on this problem. Like time passes by very quickly. I think these days people are impatient because like the hockey stick graph is just like a thing you're chasing, right? It's kind of abstract versus actually loving the day to day is a very different feeling. And so I think for myself, so actually my family, there's like way more academics. Like they're super like not entrepreneurial. And so part of that is as an academic, you just like go to the lab and like you work your whole life on like probably like a problem or like a few problems. And so maybe that's where like my inclination to do that comes from. And then I like to select for people who, um, maybe feel similarly about different domains. Right. Like, obviously like I want to make a lot of money. Like we'll make lots of money from building this, But now like there's seasons for different things, right? And the season for us right now is a lot of building. And so, um, yeah, attracting the people who can see beyond the short term are less swayed by what they see on Twitter and more truth seeking about, like, what is the state of like, technology and what it can do and like what's what are the decisions we make today that will allow us to, you know, get, um, get to the growth we want. Like, that's what we talk about a lot more. And so it's it's like we feel extreme urgency because we want this thing to exist for people who are pushing technology forward. But it's not this urgency to prove to anyone, like, like investors like, oh, we have this like growth thing, right? So it's kind of a different like frame of mind.
66:47Julian Weisser: That's a really important point, which is there is urgency. It's not like we're just hanging out, taking our time. But the idea is the urgency is towards getting something to these customers, but with also the understanding that what we want to get to them is going to take time. We're not going to sit around and hang out, but we're also going to realize that we cannot get this in a month. And, you know, there's a there's a company that's another solo founder whose name is Philip. He's building Doc Most. And he spent about, I think, 8 to 12 months heads down building dock, which is effectively sort of a notion, sort of alternative. And, you know, he was really just trying to build something that he knew would be great for people. Now he has like Airbus and like government like defense departments of governments using this. And and it's like he spent all this time because he knew that if you ship something that was sloppy, none of these people would care. You know, it was better to go heads down and build something that he really believed in, that he wanted these, you know, these big companies that were really focused on privacy or these, you know, these governments to have access to. And it's working out really well for him. But I think that a lot of people would look at what Philip was doing throughout that entire journey, or maybe even what you were doing and say, don't you know? Just try and get something out there really quickly. You know, I think that's the that's the popular sort of narrative right now. But I hope more people will watch what you know, what you're doing. Watch this podcast, but also watch what you're doing and see what you're doing and say, wow, you know, that's what it looks like when you go heads down on something. Really focus on building something that's great for other people, but don't feel like you have to like, you know, run up, run up this like, revenue growth curve immediately. Um, when this is a ten, 20 plus year, you know, journey with this company, right?
68:33Sara Du: And a lot of people are just bad at nuance. It's like some products and some markets need products that are just more fully baked. Like human trust is so easy to lose, right? It's like trying to, I don't know, sell wine. And it's like you're not going to wait to ages like you don't have wine. It's like the market wants this product that just requires time to build.
68:52Julian Weisser: That's really a good way of putting it. Okay. So let's, let's, let's let's wrap this with the two customary questions, which are. First off, sort of the bear case. The bear case for for solo founding. Meaning like, what is the reason that somebody should consider not being a solo founder? And then after that, we'll we'll maybe discuss that for a second. Then we'll pivot to, of course, the case for solo founding. But what's the bare case?
69:13Sara Du: I think a lot of solo founders probably overthink things, right? You don't really have many people you can bounce ideas off of or more sensitive topics off of because like, even very smart people outside the company don't have the full context. And so that's probably where I assume most solo founders are pretty good with like, loneliness. So I don't think that should be accounted for there.
69:38Julian Weisser: Really quickly on that end. When was it when did you make your first hire with Ando?
69:42Sara Du: Probably like a few, several months in.
69:44Julian Weisser: Okay, so. So you really had to be very careful about what you were doing over those few months in terms of getting sort of other people to riff with you, or just like, making sure that you kind of didn't rabbit hole in like an unproductive way, um, is kind of what I'm hearing from from that comment.
70:02Sara Du: Yeah. Personal time management is just a very interesting art for solo founding, especially pre employees.
70:11Julian Weisser: That's really cool. Yeah, I mean that's that's partially why we're doing all this stuff for solo founders with this idea of like solo together, not just about the loneliness, which I think does happen when you're when you have things that are discouraging. Uh, but, you know, I think it's also just like this. Hey, like, have other people around you who have context on what you're doing. Uh, because I think it really can make a big difference. Right. Um, okay. So so we talked about that. I think that's a really good point that solo founders can, like, easily get sort of pulled in the wrong direction. Um, I really want to hear the case for what's the case for solo founding.
70:44Sara Du: Like I said, companies are like entities. They have souls of their own. And if you're a solo founder, you can actually listen to like, what the company needs based on how it's growing. And you can choose like who might be worth being a co-founder later on in the journey or like if how you need to evolve to best fit what the company needs as the leader. And so I actually think it's probably the it's probably truest to building a company if you can adapt yourself to it.
71:14Julian Weisser: That's awesome. Thank you so much for doing this. This is amazing. Yeah. Um, maybe the last thing I will ask you because, um, this tends to have something really interesting happen is imagine, you know, we hit stop on the recording. Uh, we stand up, you're walking out there and you're like, oh, we should have talked about that. What is that? The thing that we should have talked about that we haven't touched on right now?
71:36Sara Du: Oh yeah. Let's see. I think building culture probably is an interesting topic with solo founders. I think that that's another I don't I don't think it fits in bearable. It's just part of the art is figuring out how much to sort of create or like put structures and rituals around and how much you just like vibe, create culture, right? Because so much will emanate from you. And so I think some of it comes down to people selection. Like people who amplify the parts of you as a solo founder, you want to see ripple into the culture. An example is like, yeah, what is the balance? Um, at least for us of like how how often we'll have these sort of intellectual salon style dinners. Right. It's like I'm not always going to be the one hosting them. So someone else should, like, pick that tradition up. Right. And then, um, another like example is my teammates are super like they're super athletic actually. And so in some ways they've kind of gotten me to like take care of my health. I have like the whoop, we experiment. Apparently I am severely under hydrated all the time. Um, so someone, like, snuck these electrolytes into my water one time and my HIV like, skyrocketed in a good way, right? Like, my hoop was green the next day. So I think like, yeah, people selection, culture building. It's like. Some of this is emergent, I guess. But I do think there's ways to be intentional about it. Or at least when you bring on the first person even, it's like, what are the things you regularly do together? It's like before you bring the person on, I think it's worth thinking about, like those structures, because once they're on and you have multiple people, it's kind of weird to be like, let's start doing this thing. And then it's like what the other people bring. That becomes part of the traditions and culture.
73:22Julian Weisser: That's really cool. I think it's I think it's a really great way to end it because it talks about, you know, how how sort of the the solo founder starts it. But the people that join these early founding team members have perhaps an even greater influence on the company culture than they would if it was, you know, a company with multiple co-founders that were already sort of lots of people who were sort of culture carriers. This has been so amazing. Thank you so much for doing this. And I'm really excited to be amplifying and celebrating your more public sort of facing launch and the fundraise announcement.
73:54Sara Du: Yeah. Thank you. This is fun.
73:56If you enjoyed this conversation, we would love it. If you can share solo founders with everyone you know, particularly people who are considering starting companies solo or who are already doing it. You know, the entire goal of what we do is to normalize solo founder and encourage more people to solo founded instead of end up with co-founders of convenience, if you'd like, and we'd really appreciate it, you can go. Leave a review on Apple. You can go. Give us a thumbs up. You can subscribe on the YouTube channel. Anything helps. And of course, if you are starting a company or considering starting a company, you should go look at the Solo Founders Program. We work together with you in San Francisco alongside a bunch of other solo founders. The idea is it's much better to be solo together than to be solo alone. I'll spend a lot of time helping you, and we also invest $100,000 in your company. So if you're excited to learn more about that, go check out solofounders.com/program.


