00:01:34 · Waseem Daher, Aditya Agarwal
[00:01:34] S0: Sure. It was principally about the team and the people that I was excited to work with. Like, in particular, for Ksplice, Ksplice was Jeff Arnold, my cofounder Jeff Arnold's master's thesis at MIT. He had had this experience. What what... First of all, what was Ksplice? Ksplice was software for updating running Linux systems without rebooting. And this was Jeff's master's thesis at MIT. Was born out of pain he had administering servers as a student as part of the Student Computer Club. And he was like, hey, I think we have something here. Let's do a company around it. And my perspective was like, yeah, Okay, cool. Let's do it. I like these folks. Let's do something together. It was much more I want to work with these folks on something rather than, oh, I'm super passionate about the Linux kernel or about group chat or accounting, as the case may be in each of these companies. [00:02:20] S1: And you've often talked about your companies being born from pain. But do you think if you had to rewind, perhaps to ask a provocative question, are you passionate about the Linux kernel? [00:02:34] S0: No. No. I mean, it's cool. Like, the tech was cool. K. But when you're in it, you, like, get so in it. And I think there's one of these things where it's like passion in in part in a very real way, I think, like, follows where you spend your time rather than necessarily, like, being the input. It's, like, sort of the output of [00:02:56] S1: Okay. [00:02:56] S0: You know, what you end up doing.
00:02:56 · Waseem Daher, Aditya Agarwal
[00:02:56] S0: You know, what you end up doing. [00:02:57] S1: Okay. So that's... That is already an emerging, probably, perhaps, the non consensus point of view in Silicon Valley, which is that passion is an emergent property of essentially working on something that with a bunch of folks who are smart, talented, that you enjoy hanging out with, that you have a personal relationship with. And in some ways, that if you build something that is sufficiently interesting and being used by people, then passion is kind of an emergent property. [00:03:22] S0: Yeah, I strongly believe that. And maybe the most provocative way of saying it is I think the mission basically doesn't matter. Like if you are in a room with a bunch of people who you think are awesome and you're getting a chance to do good work every day, that is super motivating, and that's not really a function of the specific domain you're in. Now the place the domain may matter is like, well, it could affect your ability to attract and retain other great people to work with. So it is an input into what you do, but I think your day to day happiness is rarely materially shaped by the mission or the purpose. Like, on your bad days, it's not really gonna matter that you're curing cancer.
00:04:02 · Aditya Agarwal, Waseem Daher
[00:04:02] S1: Yeah. I don't know if I've ever talked about this publicly myself, but I would probably put my Facebook journey in that same bucket. So when I joined Facebook, I was not a user of Facebook. Facebook had launched at Carnegie Mellon. Carnegie Mellon was a forty ninth school to launch, and it had launched after I graduated. And I didn't have access to my cmu dot edu email, so I never signed up for it. But I joined the company because the people were amazing. Mark was kind of a legendary leader even at that point. And it was just fun. You show up. There are hard, interesting problems. And I do think that growth and early momentum played a part in it because you get that feedback loop of, I'm doing something. Somebody seems to like it, so I want to do more of it. And then over time, the mission does build on you. That is an emergent property of, like, essentially, I think, momentum and having fun. So would... Does that sound right? [00:04:59] S0: I definitely agree. I mean, this was one of your phrases from back at Dropbox. You used to talk a lot about winning. Yes. And I do think that just, like, being on the winning team, a team where it feels like it is going well, people, I think, underweight that and they overweight other properties. Right. [00:05:15] S1: And this is also an interesting petri dish because not only have you started three companies, but you have started three companies with the same people. Right? [00:05:24] S0: I would actually argue, I
00:05:24 · Waseem Daher, Aditya Agarwal
[00:05:24] S0: I would actually argue, I [00:05:25] S1: don't think I've ever met that condition before, where, like, third company but with three same cofounders. So do you think it's different between you and Jeff and Jessica? Do you think one of the three of you might be more a priority, like mission driven? [00:05:40] S0: I'm not sure. I think actually probably not. Which well, again, the through line on all of the companies, I think, was a problem that we ourselves personally had. Right. It wasn't just sort of like, you know, we got in a room. We did some whiteboarding. We're like, oh, you know, plumbers in Kansas have this problem. Therefore, like, let's go work on it. They were generally problems that were informed by pain we ourselves had. I think that was sort of consistent among the three of us. [00:06:06] S1: So it sounds like one of the things you're saying is that stop focusing so much on, like, I must love this problem so much and I want to work on it for the rest of my life. Just focus on getting some momentum and finding people to work with. And then let that grow on itself. [00:06:20] S0: Yeah. Ideally, you have both items present. But if you have an amazing idea that you love and you have no momentum, it's going to be awful. And if you have tons of momentum and maybe you're kinda so so on the problem, I think you do end up falling in love with it. [00:06:34] S1: Okay. So you did kSplice. It's actually a really cool piece of technology. I don't know whether it's now for... Is it standard issue for most Linux kernels now?
00:06:34 · Aditya Agarwal, Waseem Daher
[00:06:34] S1: Okay. So you did kSplice. It's actually a really cool piece of technology. I don't know whether it's now for... Is it standard issue for most Linux kernels now? [00:06:45] S0: I think... I mean, it's definitely still part of Oracle's Linux offering, and I think Red Hat has its own version and, like, some of the other providers, I think, maybe have their own version. [00:06:53] S1: Got it. I mean, listen. I mean, I... So were you at Oracle for a short while? [00:06:57] S0: I was at Oracle for a year and a day. [00:06:59] S1: Okay. Great. I was at Oracle for six months and a day, let's say. But you stayed there for a year and a day, and then you left to start Zulip. And, you know, Zulip was actually kind of a phenomenal product, arguably maybe a little ahead of its time. But I'm curious, tell me a little bit more about the early days of leaving and arriving at that idea.
00:07:20 · Waseem Daher
[00:07:20] S0: Sure. And so I think our thinking evolved really for all three of these companies, which is for Ksplice, The process was basically Jeff has this cool technology. We're all Linux nerds. We think there are other Linux nerds in the world that would want this. Let's build a company around it. And that was basically the extent that we rigorously analyzed the problem or the domain. With Zulip, it was like, okay, well, what was the problem we were solving? The problem we were solving is at Ksplice, we were in search of a robust group chat tool for businesses, and this was... Just to set the stage, this was like pre Slack. Like, the state of the art at the time was like IRC or HipChat or Campfire. They didn't have mobile apps, and HipChat was like an Adobe Air app. It was like Flash kinda. The stuff did not really work very well. And so we said, Okay. We think that we could build something better that has a mobile app, that has good search, is a nice robustly built modern web app. And also, we sort of had another take on it, which is we thought that if you... Embedded threading as kind of a first class notion in the application, that you would get something that worked well both synchronously and asynchronously. And then we thought, well, okay. Fine. How big could this be? Well, actually, you know, every business does have this problem. And so I think we were right about the market size in the sense that if you look at the adoption and the valuation and the growth of Slack, like, the time and shape were basically right. So I think we were correct in the thesis. I had only have some notes for us on how we executed, but that's that's a little bit about how we got there. It's... Again, it was a problem we had when we were operating k splice, which is how do we... You know, what should we be doing for chat?