[
  {
    "start": 0.0,
    "text": "Tobi: Hello, friends. This is the alphalist podcast. I am your host, Tobi. The goal of the alphalist podcast is to empower CTOs with the info and insight they need to make the best decisions for their company. We do this by hosting top thought leaders and picking their brains for insights into technical leadership and tech trends.",
    "end": 18.0
  },
  {
    "start": 18.0,
    "text": "Tobi: If you believe in the power of accumulated knowledge to accelerate growth, make sure to subscribe to this podcast. Plus, if you're an experienced CTO, you will laugh the discussion happening in our Slack space where over 600 CTOs are sharing insights or visit one of our events. Just go to alphalist.com to apply.",
    "end": 51.0
  },
  {
    "start": 51.0,
    "text": "Tobi: Welcome to the alphalist Podcast. I'm your host, Tobi, and today with me is Loïc Houssier. Um, Loïc, welcome to the alphalist podcast. For everyone listening, Loïc is the CTO of Superhuman, a tool that many of you know, a company that decided, uh, to that email doesn't have to be a misery, uh, and then proved it. He's a French engineer who transplanted into this, he was transplanted into this San Francisco based ecosystem, and he started steering Superhuman through its acquisition by Grammarly. Was that all correct?\n01:26] Loïc: It's it's correct. It's correct. Like maybe one adjustment to it is that Grammarly got renamed Superhuman. So technically I'm not the CTO of Superhuman anymore, but I'm the CTO of Superhuman Mail, which is the business unit inside the big ecosystem of Superhuman, which includes Grammarly and Coda and now, uh, Mail.\n[01:47] Tobi: Okay. Okay. So, uh, Grammarly decided that, um, you got the better brand and, um, that they have to change to a better brand.\n[01:56] Loïc: I don't know if it's a better brand. Like Grammarly is definitely a good brand. Um, but it's somewhat limited in terms of scope, uh, in the imagination of people when they they think about it. Um, so Superhuman is much more, I would say larger scope. You can put like a way more things behind the name compared to to Grammarly. But Grammarly is, I mean, it's a brand name. Like everyone knows Grammarly. Like talk about like the the students.\n[02:20] Tobi: I mean, I I mean, I I know it as well, but I don't know, I don't know anything about it. I don't know what they're doing. So, uh\n[02:26] Loïc: Oh, yeah. That's um, it's a really good, um, it's a really good business. Uh, they were doing basically LLM before LLMs. They were doing, uh, like probably like most of the\n[02:36] Tobi: I like spell spell corrections, right? That's that's\n[02:38] Loïc: It's spell correction. Uh, that's why like Grammarly, it's to be better on your grammar and to have like a a better way to to to provide, I would say everything that you want to communicate to other people, uh, without the typos, without like, uh, I mean like even the the the way to phrase things, avoid repetitions and all of that. So it's really used, it was initially used a lot by students writing essays or like people doing like a lot of writings, uh, to help them, I would say move faster and go faster. And, uh, expanded beyond it. So it's a, it's a pretty cool tool.\n[03:12] Tobi: Okay. And now you basically do the upsell. So Superhuman Mail users now have to buy Grammarly in addition and Grammarly users have to buy Superhuman and you even have a like for Grammarly, you even have a deeper lock in than before or stickiness, let's call it stickiness because you you you also change your email address and all that that that that jazz, right?\n[03:33] Loïc: No, of course. I mean it's, um, I it's I would say more than just an upsell strategy. Uh, it's much more like a a vision. Uh, both company has the have the same type of vision where productivity is very important for like knowledge workers and you get productive if we help you write because knowledge worker, they write a lot, documents, business cases, whatever you, I say our jobs as knowledge workers is like very written oriented or writing oriented. Um, and most of the places where, uh, Grammarly was used was in emails. People were in emails trying to write, I would say proper emails and everything. So it was kind of like almost like a logic to some extent, uh, to have like the two brands, um, I would say work closer together. Uh, but the the vision, uh, she she the CEO, uh, formerly of Coda, but became the CEO of Grammarly, uh, and, uh, Rahul, CEO of Superhuman, had the shared vision that one day the way we, um, I would say we manage our productivity suite will be different. And both of them started from a different place, but they have like the same end goal, which has, which is to be like the basically the AI native productivity suite of choice. And productivity suite means you need to manage your communication like email, you need to manage your task, you need to manage your documentation, you need to, um, I mean, there's a bunch of stuff in the productivity suite. And everyone is sort of like doing that. Like you you see Notion. Notion is doing email now. They're doing calendar as well because there's a consolidation of the different, uh, actors in the space, uh, towards like this productivity suite and be having like more breadth. So, uh, so yes, the name was also like something that was important, uh, as part of the acquisition, I I trust. Uh, but the the use case and, um, our know-how, uh, was also important for the company.\n[05:26] Tobi: And beforehand, I guess both companies had hundreds of millions of funding and you went through like traditional VC track, right?\n[05:33] Loïc: Yes. Um, I would say all companies were like typical VC track, typical, I would say VC funding. Yes.\n[05:40] Tobi: Okay. Okay, cool. And beforehand, I mean, you kind of have have a history that is very impressive, um, for let's say a Frenchman, like not not judging, but for a European, right? Like you don't see many Europeans in the US and you already worked like as VP engineering for DocuSign and Firstbase, etc. So, how maybe before we dig deeper there, maybe we start with your early career, like and how did you got into computing actually and what what was your early spark, let's say and and uh when when you started breaking computers in France and then um what made you jump on that train and and why you actually ended up in SF?\n[06:18] Loïc: Yeah, uh, that's an interesting I I find it pretty uh, interesting. Uh, but like I started as a, I would say typical engineer in France. I was a a math graduate. I was on the on track to do like a PhD in math and computer science, uh, in cryptography. Um, so security space. Um, and started, uh, I would say with uh, Orange, the big telco company. At the time it was called France Telecom. Uh, so they are like, I feel old now. Uh, but Orange, so big telco, I would say hundreds and thousands of, um, employees doing security research there, uh, doing that, kind of like typical engineer starting in Europe. Big co is the go-to. It's kind of like where you go. You don't start.\n[07:02] Tobi: And then you discover that it's super lame as soon as you work there, right?\n[07:05] Loïc: Yeah, no, no. I mean it's, it's not, there's a lot to learn. There was a lot to learn. Uh, and I learned a lot there. Interestingly, a bunch of the people that were there, uh, moved to the US like in the same lab at the same time. Um, one is in New York, one is in LA. So like everyone got, I would say two in LA. Uh, so everyone moved uh, I would say from there. So it was a really good school. Uh, let's put it that way. A really good stepping stone in my in my career. Then I switched to another like giant in France called Thales, uh, defense industry, still doing security. Uh, so in that security researcher, I would say space, but also like doing some project for Department of Defense, like for like European, uh, agencies and stuff like this. Um, interestingly, I did two years outside of the software industry. So, um, in those big corporation, when you're young, motivated and everything, they want to try different you to try different jobs. So I did some stuff, like I I went to a an harbor in France, uh, where I was uh, doing maintenance of like submarines, carrier aircrafts and all of that. That was like an tremendous experience. Then I switched to like uh, uh, nuclear civil plant, uh, contract management, totally different, but as part of my career, I would say this opened like a lot of like different ways to work because I was working with people that were not from the software industry. The way they were working was different. The way they were thinking about problems was different as well. So I think it was like really enriching and it was a good school. And at some point I was fed up with like big companies. So I joined a small company and it was not exactly a startup. We were 60, 60 or 70, not technically a startup, not VC funded, but it was a joint venture between uh, I would say big companies uh, in France, uh, still in the defense industry, so doing security things. And this is where DocuSign knocked to our door, uh, and basically willing to buy us because one of the business line we had was doing like digital signature. And I don't want to enter into the details, but like the same technology was used for security things and digital signature. And they wanted to buy that vertical. So technically it was not even an acquisition, it was a we call it a carve out. So they wanted this asset, the people supporting this asset, uh, but not the rest. And we had the Ministry of in France, the Ministry of Finance basically saying like the other business line, no way. That cannot go into the hands of Americans. So we had to create like a new structure, basically put the half of the asset, half of the team into the new structure and sell the new structure to DocuSign. And this is how I I came to join like San Francisco and DocuSign through that acquisition.\n[09:53] Tobi: And and and DocuSign wasn't afraid of French labor law?\n[09:58] Loïc: They were. They were. That that was like uh, that was part of the discussion like as part of the due diligence, as part of the, um, but they were also like, uh, that that's one thing that we we learned, um, or I learned, um, in the US, it's it's just another risk. It's just another risk like all the rest and, uh, you if you don't do an operation, it's also like an opportunity cost. So like the question becomes, if I don't buy this company that is the leader in France and part of the top whatever like five in Europe, how do I enter this market? Because it would be messy anywhere in Europe. So they took it as a as an input into like their decision process. Uh, but that don't stop them like from making decisions.\n[10:41] Tobi: And I guess the the fact that it was kind of a complex deal, right? Um, uh, also shows or also also makes it a chance, not only a risk, like not many others buyers would have done that, like and not many other, um, folks would have constructed that, um, that that structure, um, only for for selling that that that asset or only selling for for selling that product, etc.\n[11:07] Loïc: Exactly. Exactly. And that was also like a good way as an acquisition to be successful within DocuSign because, uh, we were motivated to make the deal successful. So we've done all the hard work and they were like, holy cow. Yeah, they're French. Yeah, like they have a lot of PTOs. Uh, they don't work a lot, whatever. But like when they work, oh my God.\n[11:28] Tobi: They drink white wine for lunch.\n[11:30] Loïc: 100%. 100%. That part they love. But, um, but yeah, that that was also a good opportunity for us like to demonstrate, um, I would say that not only they were buying an asset, buying a business, buying a revenue and and a customer, um, a set of customers, but also they were like buying a team that were like demonstrating, uh, that, uh, they can do shit, basically. So, um, it was, um, a good opportunity for the team there, but it's also like a good opportunity for me because I was able to drive this from a like a technical standpoint, like all the separation of the systems, the teams, without drama and making like the the the acquisition successful, uh, which is never easy. An acquisition is never easy. You have a difference of culture, you have a difference of, uh, toolings, you have difference of perspective of vision. Uh, so\n[12:16] Tobi: Yeah, I know that. I know that. Unfortunately. And and uh, especially if you then aim for synergies and synergies rarely come, right? Uh, that that's in many cases the problem then, uh, um, often shit hits the fan, right?\n[12:32] Loïc: Yeah, and I think that, uh, where they were or maybe like stuck is like the synergies where basically you let go people because like you have duplication of roles and everything. There were not that many, uh, because we were like in a very, so for one, we were like a small acquisition. Uh, that helps. And the second part is like we were in France and we were targeting like a market that didn't have really. So there was a lot of subtleties that they could not manage from the headquarters, um, in the US. So a bunch of the people that were in a typical redundancy, uh, area, um, were either kept or like, uh, I would say put in a slightly different position, reporting to different structure and everything, but still useful for the business to be, uh, to be run. So it was a smooth landing, mostly like for everyone. Um, some people were never happy because, uh, I don't know, you're used to be the C level of a company and all of a sudden you're reporting to\n[13:26] Tobi: You're reporting to one two three levels, yeah.\n[13:28] Loïc: Uh, in the US and it's a different job. It's a different job.\n[13:33] Tobi: Yeah. Yeah. But for US company, it's it's also like, uh, I mean, France is kind of kind of kind of kind of, uh, let's say not too expensive salary wise, right? Uh, I mean, that's also a chance. Uh, so, uh, like ideally they get rid of people on their end. I mean, hire people as VP engineering, for example, um, move them to to SF and then they discover how how how expensive everything is.\n[13:58] Loïc: Yeah. Oh, yeah. Don't don't get me started on this. Like when they make me come like in the US, um, so they gave me a salary and I was like, holy cow, like I will be like rich and everything. That was awesome. And then you start to discover like the the the the the cost of life, the cost of living, uh, in San Francisco and then I had three kids and having kids in a family in the US is freaking expensive. So the more time you spend in the US, the more you value the social democracy, uh, in Europe and their models.\n[14:33] Tobi: And also the healthcare systems, etc, right? That's uh crazy.\n[14:36] Loïc: I pay for school. Uh, I pay for healthcare and even as my company is paying, I would say like the my insurance, still, uh, I there's this concept of co-pay. And just to let you know, like every time I go to like a optometrist like for my, uh, lens, uh, or like contacts and all of that, it's cheaper for me to go to France and pay without any insurance rather than paying with insurance in the US.\n[15:01] Tobi: Wow. That's crazy. That's crazy.\n[15:03] Loïc: It's a market. It's a market. It's interesting.\n[15:05] Tobi: Crazy, crazy, crazy. So I I better stay here. Uh, decided. Um, but um, we we're here to also talk about tech, um, and and not about contact lenses. Uh, I mean, no, no, no offense. Um, what like your what is your personal, what what would you call your personal secret source? Like you as a person? Is that is that that you're deep in security or, um, because you worked in like, uh, encryption heavy environments or is it that you worked in those big corporates and then can basically work with almost everyone as you just stated or what what would you say is your personal secret source?\n[15:43] Loïc: Interestingly, like I told you that I was like, I went to a harbor and, uh, I, uh, I did some maintenance on submarines and like carrier aircrafts and this type of boats. Um, I had to be the muscle, uh, of learning fast deep tech. I would say those deep techs were like torpedo system, propulsion system, like nuclear systems that are basically making like a submarine, uh, work, uh, because of the of my the nature of the job because I was here to help make like the, I would say progress on the processes, making it like smoother and and eventually like reduce the cost of doing that maintenance. But you need to ramp up super fast to understand what you're talking about, at least a little to ask the I would say the right questions. And you need to to reach that level where you can ask the right question for the people that are in charge. So that, uh, ramp up, uh, and, uh, ability to learn very fast something that is new to me. I think is something where I spike now and I build that muscle. The other I would say aspect is that, um, as a leader, I'm totally comfortable working with people smarter than me. Uh, so whether it is AI, whether it is like whatever that can be, of course I have AI specialists and they know better than me. They have a PhD in this domain. Never I would be like more technical that those people having that that depth and everything. But I think that, um, as a leader, I proved, uh, or I believe I proved that I can pretty fast learn enough to be, um, a good sounding board for those people. I can challenge them, uh, in the spaces where they didn't think enough. I can detect that when it's like a bit shallow. Of course, they would be like way deeper than me, way better than me. But I think that's the that's the part. The second part I think that, um, or the third part, uh, that I think I spike, uh, is execution. Um, like this acquisition that we mentioned, uh, was successful because we got shit done and fast. And, uh, that's creating like a some sort of like a brand of, ooh, that guy, I would say he can have like a hairy project or or hairy things to manage, he will do it. Uh, and that in the Bay area, especially in the world of startups, uh, that's is highly valuable.\n[18:03] Tobi: It's got out of hand for many companies, I guess, right? Like, uh, who were like really questioning their velocity, etc. and and then\n[18:10] Loïc: Exactly. Exactly. So when you have doubts about your, and everyone has doubts, uh, on their velocity. Everyone is afraid of competition. Every company in the Bay area, that their just like their default is like, we might die tomorrow. So that's the default compared to Europe where it's like, oh, we're already making like X millions of ARR, we're successful and everything. No. No. And and until you make\n[18:33] Tobi: We stay in Germany, we stay in Germany, we stay in France, right? Our market is big enough.\n[18:38] Loïc: No, exactly. Exactly. So here the mindset is different. So when you have this ability to get shit done, being intense in the work and keeping your team happy, having the same level of intensity, you start to be, uh, a pretty pretty good asset for a company, I would say. That or at least this is my interpretation of why people want to work with me.\n[18:58] Tobi: And, uh, I mean, I guess you have the proof through successful products, but these days, do you, I mean, I'm I'm curious because I'm I'm thinking about it so much because like I have those 26 companies I acquired, um, all different texts, uh, and and I I think about like how to make it really comparable and measurable, uh, which is which is from my perspective super hard. Like, are you into, I don't know, Dora metrics, something? Do you believe in it?\n[19:25] Loïc: I believe in improving. Uh, I don't believe in, uh, I would say comparing because like the context might be different. So it's really hard. Yes, the Dora metrics are interesting, but like more than the metric itself, the trend is the one that I'm looking at. So like how is it evolving?\n[19:39] Tobi: So how is it evolving over time? Yeah.\n[19:41] Loïc: Yeah, exactly. And this is how I measure my success. Uh, starting from here, wherever here is, starting from here and boom, moving things uh, up. The one metric that I love and I've used, uh, I loved, uh, I think it's the less bad metric, um, is the number of PR per engineer per week that you average across your organization. So you don't choose it.\n[20:03] Tobi: Number of PRs per engineer. That's, um, yeah.\n[20:06] Loïc: Uh, you, so you don't count like the line of codes, you don't count like, you just that's probably like the first unit of value that is shipped to, I would say your customer that is providing some level of value, whether it is performance improvement or a feature, I don't care. But that unit is one that is interesting. Of course, you need to take it, um, with a grain of salt. Uh, of course, not against, I would say specific engineers, because that would be like the dumbest, uh, I would say thing to do. But on aggregate, that's a really good way to understand how a team is evolving. Uh, and you need to be careful about the teams. Some people that are working on product led growth, typically they do like a lot of iteration, a lot of experimentation. So they will have a lot of PR per engineer per week. Normal, nature of the business, very front endy, uh, to some extent, easy work. A back end team, a platform team, every time they touch, they touch like the infrastructure or the database and all of that, they are very careful. So of course, their number of PR per engineer per week is usually a bit lower. But in their own context, you can start detecting the trend, uh, and like all the modification you do as a CTO, VPN or manager, uh, you can measure the progress, um, that, uh, that that you're doing. From what I've seen, most of the progress, I would say it's not about the tooling, it's not about, um, putting pressure, it's, uh, it's all about providing clarity, uh, clarity of the strategy, clarity of the things to be built, clarity of like the processes to get shit done. Uh, so usually most of my time is providing clarity, clarity on the mission of teams, like having a good organization where each team understand what they're doing. Coming back to my acquisition with by DocuSign back in the days, our mission was clear across the company. Like, you build everything that makes DocuSign compliance for the European market. Easy. And now you drive, you do. With PMs, designers and all of that, you do you, you know better, just do it. When you're empowered and you understand what's the mission, you just move fast. Because you don't have to call the headquarters and say, hey, can we do X, can we do Y and all of that. So that helps a lot.\n[22:26] Tobi: As long as you have like a a team that is kind of flow aligned, right? Um, uh, where you don't depend on other people, on other teams, um, etc, which is from my perspective really like a another\n[22:38] Loïc: But that that's part of the exercise like when you build an organization, uh, you want to limit like the interdependencies between teams, uh, because this is where, as you said, like the shit hit the fan. Uh, like, oh damn, that team depends on that team and that one team and especially if you're across the ocean, uh, it can become like a nightmare. So you want to build an organization where teams are mostly vertically aligned. You want to ensure some level of consistency, uh, but when you're in like a hyper growth and you want to win the market, you want to move fast, you're not at the stage where you want to optimize just yet. You just want to move fast. Uh, which is also like something that, uh, was different between the Europe and the US. Like in Europe, we're always trying to find and taking the time to find a solution that will be the best ROI, like not too much cost, but really bring us there. Here, just move fast. You know what? And if we're successful and the infrastructure is costing a lot, that would be a good problem to, uh, to solve for. But if we're not successful, I mean, there there won't be problems at all. So I I think that's also something that, uh, was like a click, uh, in my head is like understanding that there's a time for optimizing and there's a time for like velocity and getting shit done and understanding when to do that and empowering the teams to understand when to do that is also like a good way to accelerate the business.\n[24:02] Tobi: Mm, mm, absolutely. And do you, just out of curiosity, I guess you have some microservices in your organization or rather like maybe bigger services that that, um, are connected, uh, between or are are like interconnected, right? So you have some dependencies in your team or is it like monorepo and, uh, everyone works on the same stuff?\n[24:22] Loïc: So, uh, it is monorepo for the back end, uh, at Superhuman Mail. Uh, it's monorepo for the front end, uh, or what the what the client application. Um, we don't have many services. Uh, I am like you have to really force me to create a new service. You have to explain me why it makes sense. And don't tell me like, oh, it's because of the scaling five years. I don't care. Like five years, anyway, we will redo like the architecture three times since I would say, uh, between now and in five years. So, uh, definitely like microservice is not something that I want to hear in my organizations. Um, interestingly, Grammarly, which has like much more like a code base, it's like a bigger and everything, uh, they're moving towards a monorepo, uh, because of this interdependencies and the teams in the US, the teams, uh, in, uh, in in Europe. Uh, but there's also like the the code ownership. Uh, when you do a PR and you have like 10 code reviewers because you're touching different places, even if you have like a monorepo but like separated like code owners for, I would say directories and everything, it's a nightmare. It's a nightmare. So, um, yeah, I I'm a fan of like modulith, uh, like cool domain design, uh, oriented type of architecture, but like as much as you can, one big deployment, uh, to make things easier and you start breaking things up when you have to. But it needs to be painful. You don't do it just because it's fun and looks great on a Figma, uh, and an architecture diagram.\n[25:58] Tobi: Mm, mm. Yeah, that's, uh, I think a good philosophy that also many CTOs the recent years learned in a painful way, right? Like, I don't know, I mean, I personally was attracted by microservices like, I don't know, seven, eight, nine years ago when it all really like went through the roof, but, um, I think like being being attracted by things that like big tech tells you is is not idiot, right? Like it often leads you into into a big time sink that you then like really really have to recover from from later on, right?\n[26:37] Loïc: No, I think it's a fair point. Like, uh, we tend to look at those big corp, like the Google, the Meta, the, uh, like even like LinkedIn and Kafka and so there's there's a bunch of tech like this and and you start your company like, oh yes, I know that LinkedIn is using Kafka, let's use Kafka. Yeah. But like you're a team of two, uh, you're serving like 10 customers. Like don't do it. Don't do it.\n[27:02] Tobi: And all of a sudden you also discover that LinkedIn actually still uses Ember JS and then you know that like if you follow that company, you're doomed.\n[27:10] Loïc: No, exactly. Exactly. Like the the best example that I had was, um, you know like those models with like guilds and, um, I'm a fan of guilds. Like in my organization, I love to put guilds and everything. And it's coming from Spotify. And, uh, and I was like listening to, I would say the discussion with someone at Spotify like, yeah, you your ways of managing like, I would say the organization and everything. He's like, ah, I mean, it's mostly marketing. The reality is way different. The philosophy is there and we try to stick to it. The reality is like, we do what we can. And like it's it sounds like paradise and everything is well organized and everything is, is working well. No, no, no, we're like you guys. It's it's it's somewhat broken and we have a way to look at the way we're broken that is consistent, which is already good. Um, but that's that's about it. The same way like, uh, so I've been remote for quite some time and managing remote teams for quite some time. And GitLab, uh, is the example of the, I would say remote team and how to work asynchronously, being like, I would say documentation first. And same, I reached out to a friend there and and she was like, come on, like, like, yes, uh, that's great and all. It's a good way to attract engineers and everything. The reality is, we're like 70% there. Like it's always shitty here and there and new teams and all of that. So I I I became, uh, I I'm now like way more, um, skeptic, uh, when I see those great stuff and everything. And and even if it's working, it's working in one context. Um, the other example that I give, uh, I used to give now it's not the case anymore, but like early on, people that were like so much into Agile. Oh, we need to be agile, we need to be agile. That was the case in Europe. Like, we need to be agile and we want an agile coach and like, oh, the 12 like the manifesto with like, and you're like, come on guys. Like read the content, I would say read the doc. Like it's a set of guiding rules written by guys that were like senior architects, senior like engineers. They knew each other for so much time. Of course they can apply that. Now we have people out of college. We have people, I would say that with like different experience, they don't know each other. Of course that won't work. Stop applying stupid rules. That's a pretty interesting thing to think about, but like we have a specific context. Like so we need to take your context into like, that that's probably the most important input. That's your context. Sure Spotify, sure GitLab, sure whatever. Now in your context, what makes sense? And I think this is, um, one of the fallacies like, uh, reading too much, seeing too many frameworks and time sync because you're trying to understand what's the best and everything. So now I have an approach where like, uh, I try to solve one problem per quarter. Maybe the problem is velocity. Okay, boom, I do some stuff. Velocity is better. Okay. Oh, quality is not great. Check on quality and boom, you you build a new baseline. And trying to solve like and every time really thinking about the context and sure, I I have the the input of the different frameworks that I've seen, different companies that I've seen, different scale of companies. Uh, but that every time it's an ad hoc exercise with a bunch of people that I force to take a step back, forget about what they assume is right, uh, and just like challenge each other to do like one thing, do it well, be persistent and then tackle the next problem.\n[30:41] Tobi: Yeah, I think that's a very good learning for for for many out there that like focus is so valuable, right? And if you take the luxury of really like focusing on something for a quarter, I mean, maybe you realize that was the wrong thing, but you have to pick like a bigger bigger portion basically so that you can really do it and and don't have like 10 things you try to focus on, right? Like don't have 10 things at once. Don't have 10 goals. Have one goal.\n[31:07] Loïc: You can like you have a team. Like maybe my goal is to improve like velocity this quarter or like the AI adoption. That's my quarter for AI adoption. But like maybe like my managers will have their own one thing they want to solve. So like we and and one IC, like senior IC want to solve one thing, like maybe it's like the way to use feature flags, uh, in the company. Okay, go solve it. Tell us how to do it better. Like when everyone has like one mission, um, that is clear, uh, that is like their one thing of the quarter, you can multiply that by the number of teams and, uh, you can have like a pretty significant impact. Not everything will be, uh, will be done, but like if you start having like all of those, but everyone has its focus, its own focus, it's it makes a difference.\n[31:54] Tobi: Mm. Yeah, that's a that's an interesting view. I you you you recently or you you you you mentioned guilds like a few a few points, uh, a few questions ago. Um, how how did you get that that running? Like I'm I'm always curious. Like I, um, like I mean, maybe I also have a bit of a special environment because I I basically buy SAS companies, but and and the teams are all basically puzzled together. Um, and and it's hard to get people to to real activity and to real real action. Um, how how do you manage to to incentivize that, uh, or how do you manage to to feed it?\n[32:30] Loïc: So one thing that is usually interesting is like whether you normalize or formalize those guilds or not, they exist in some capacity. Uh, because like engineers, they talk to each other. They want to understand how other teams are doing their front end framework, like the scaffolding of project or whatever because every, usually good engineers, they want to improve and they're curious about others and they want to understand how they're doing things. So no matter what, it's already there. So there's already like a group of people that is talking. Um, and I organize teams guilds through stacks. So like you have an iOS guild, you have like an Android guild, like a back end guild, a front end guild. Now we have like an AI guild of like all the people interesting in AI toolings because there's always some people that are more interested than others, uh, into like everything new that is coming and all of that. Um, so formalizing it, uh, was one one piece. Uh, two was like, uh, giving time officially. Like the same way we have shadow IT in companies because people want to install their their their own thing, their own thing. We always have like shadow engineering because people sure I have my features, sure I have my bugs, I have my to-do list and everything. They're always trying some stuff on the side and everything. Of course, and you need that. You need that. You don't want robots. Uh, so like formalizing the fact that yes, you can spend some time discussing with others about like the new new techs, doing small POCs and all of that, uh, is a key piece. The second piece is we put like a guild leaders. So a guild lead, uh, is usually a senior engineer from that stack, uh, and it it was like both an incentive and a accountability, uh, thing to do. Um, they have to come back to us, us being the leadership deciding of like for the planning for the next quarter. At the end of each quarter, they come and they say, hey Loïc, you know what? Uh, on the back end, uh, we have some issues with the monorepo, we have some issues with, uh, our CICD pipeline, we have issues with whatever. We want to solve it next next quarter. And that's the, I would say the mandate for the guild lead is to identify the top problems. So they surface, they come up with a list of three, four, uh, every quarter. And as part of the next quarter, we take that into account and we say, okay, you convinced us, uh, and the head of product will do less features and we'll do that one and that one so that this quarter you can, uh, I would say define, I would say we will be working on it and and when and we basically put that into the, uh, the quarterly planning. And it's working relatively well. Again, marketing a bit, uh, like everything is perfect and everything. But like of course, like if my engineers like listen to the podcast, they would say, come on, like it's not working as well, right? Um, so take take everything I say with a with a grain of salt, but that's the idea. That's the idea that we're trying to follow. Sometimes it's working well, sometimes it's working less well, uh, but, uh, doing postmortems and trying to understand what we missed and everything. But that's the model we I tend to follow. And I do it because my organization, I follow typically the same type of organization, platform team. And for and then like teams focused on like a vertical so that there's not a lot of dependencies. The only dependencies with the platform team, so infrastructure, core back end, storage and all of that. And then like the teams are full stack, uh, and focusing on a specific experience. So for us at Superhuman, it will be focusing on the calendar experience. I want a team that thinks day in day out about the calendar, how people are using the calendar experience. I don't want people to move away like so from project to project.\n[36:14] Tobi: Take it from A to B or move from project to project. Yeah.\n[36:17] Loïc: Exactly. Because you lose the knowledge, uh, you lose a bit of the care because if it's not your experience and you're just building a feature, you're a feature factory and it's fire and forget. So it's creating some pretty good incentive in terms of of ownership. But doing that, you separate your front end people, your back end people, your mobile people. So you want them to discuss across the team and that's why I'm sort of like forced to artificially create those guilds so that they talk to each other and make sure that we're consistent in the way we use, uh, React and we use like the design system or like, uh, uh, and and all of that.\n[36:57] Tobi: And they they kind of meet every few weeks and are organized on Slack, I guess, and, um, that's it pretty much.\n[37:04] Loïc: Up to them. Up to them. Like, uh, one thing that I learned as well is, uh, the aligned autonomy. Like the alignment is, you guys, you need to meet, you need to, uh, like to identify the problems and every quarter I want to have like, uh, basically the, uh, the the outcome, which is the, uh, the list of things we want to to solve for the next quarter. And when you tackle a project, you come to me every month and you give me like an update on the progress. Have you done what you what were you planning you were planning to do? Yes, no. No, why? Oh, because this feature we needed to to ship and everything. Okay, not happy. I will discuss with the head of product because we we committed on giving you time and everything. So we have some checks and balances like this to, uh, to try to do to do okay.\n[37:48] Tobi: Cool. Yeah, really sounds good. Um, I I I kind of have to think of like, um, I don't know, I'm I'm not a superhuman user yet, so, uh, I have to\n[37:56] Loïc: You should. You should. Come on.\n[37:57] Tobi: Yes. I'm I'm still with Gmail and I I hate the spam filter these days because like email is really getting worse. But, um, like from my like in in front of my inner eye, it kind of reminds me of like, it's kind of the linear of email, right? Like, uh, I don't know if you know linear that they they they have that concept of really being super snappy and super fast on the front end and really really doing things differently. Is that and it's kind of it kind of for me is like a, yeah, same blueprint, same same playbook. Uh, is that is that true or?\n[38:31] Loïc: 100%. 100%. Uh, we follow a rule internally. So there's two ways we are like very similar to linear. Um, and we use linear as well. So the first thing is like we're very design centric. Like we care about the experience. We care about like the people using the product. Um, and, um, we are laser focused on one specific type of people. Compared to Gmail, Gmail needs to work for my mom, myself, and my kids. So when you have like a different set of personas like this, like you only focus on the I would say the the the the the biggest common denominator of all those people. So basically the experience that you will provide will be meh because it needs to work for my mom, she's 70, but it needs to work for my daughter and she's 12. So of course your experience will be okayish, but never great for anyone. Uh, while, uh, Linear and Superhuman, we are laser focused on a very specific type of persona. So Linear, they focus on engineers and product managers. That's it. Like it's not like Jira focusing on IT, focusing on like so many different groups. Nope, laser focus. So then when you know those people, you know how they work and you can do like a provide a better experience. For us, it's like people with like a shit ton of emails. We focus on professionals that leaves day in day out in emails. Funders, uh, head of sales, uh, C levels, like I receive like a ton of emails, uh, all the time. Um, I'm saving like, I I should do the the count. Like we did the count like for our users. Um, I think it was like a year ago, but like, uh, the survey people were basically saying that they were saving four hours a week. So four hours a week, and and I would say I have the distribution. It's crazy to see. It's crazy to see. But like the distribution, like the the peak of the distribution was on four hours a week. Um, it's crazy. It's crazy.\n[40:25] Tobi: Four hours a week is like 10% of a like a standard working week, right? Uh\n[40:30] Loïc: Yeah. And if you're a CEO, you're living day in day out in your, uh, in your emails because you're discussing with your funders, uh, you're discussing with your board, you're discussing with like partners, with like, uh, your customers and that's what you do. That's what you do. Um, so of course, uh, at at that, uh, I would say four hours a week, what what is the amount of work that you can do like more like in four hours a week? What is the value of that? Sure, you will pay a subscription. And that's why people are willing to pay like 30, 40 bucks a month, uh, compared to Gmail that is free to some extent. Um, so, um, so yes, uh, we are snappy. So those people want something that is super fast. Don't want to wait. So we have a rule internally, which is like the 100 millisecond rule. Every interaction in the product should be under 100 milliseconds. We don't want people to wait and having like a spinning wheel while you're doing something. Things need to feel fast. And it has a lot of implication on the architecture because 100 milliseconds for the technical people listening to, uh, to us right now, 100 milliseconds means like you cannot just rely on network. Because if you use the network and you have a data center, uh, that is like in US West one and you you're living in Europe, boom, 200 milliseconds, you're done. Yeah. So, uh, we are offline first. So everything is happening locally. And then we do like the sinking, uh, later, but everything that you do happens locally. Sending an email, replying to an email, um, reading your email, everything is already downloaded to your laptop or to your phone so that if you're in a plane, you can work on your emails.\n[42:05] Tobi: I mean, like a normal app, right? Like Outlook back in the days, you also have your email synced and you like downloaded them via IMAP and and and and and then you answer on them and then they're sent later. Um, and you have\n[42:19] Loïc: Well, good luck with that. Good luck with that. Good luck with that. I will just say that. I will just say that the people that are using Superhuman, uh, are pretty happy with it.\n[42:30] Tobi: I I didn't want to question that. I didn't want to question that. And the comparison with with Outlook is maybe not the best one.\n[42:37] Loïc: No, exactly. Exactly. It's as if you were comparing like, ask people using Linear if they want to go back to Jira.\n[42:44] Tobi: Yeah. No. Good luck with that. No.\n[42:46] Loïc: Good luck with that. Like if I tell my engineers that I will go back to Jira, like some will quit. Yeah. Because it's so worse, uh, like in terms of experience. And and again, um, Jira and Atlassian, it's a great business and they did like some really good things and everything. But for specific\n[43:04] Tobi: But it's an operating system, right? Like it's an operating system and sometimes you want simple things and that's the problem.\n[43:09] Loïc: It's an ERP. It's basically an ERP. Yes.\n[43:12] Tobi: Yes. It is. Um, yeah, and but but you still rely on the that those those like dinosaur protocols like like IMAP, etc, right? Or or\n[43:24] Loïc: Oh no, we use, we use mostly like the APIs of Gmail, APIs of Outlook. So it's, no, no, no, we don't we don't deal with that. We don't have to. Uh, we're not even implementing, uh, so we're not a mail server. So, uh, you need to have Gmail on the back end and you need to have Outlook on the on the back end.\n[43:40] Tobi: Okay. Okay.\n[43:41] Loïc: For the main reason is like most of our customers, they also use like Google workspaces. So it's also their identity and all of that. So it's just simpler. It's just simpler. They just want a better experience managing their emails.\n[43:53] Tobi: That kept me away from using hey.com. I once had a conversation with David who who who just said like DHH who just said like, hey, we couldn't rely on on IMAP because we just wanted to be freaking fast, etc. I don't know if that was like bullshit back like look looking back. Um, you you can tell me more about that, I guess.\n[44:12] Loïc: I mean it's, uh, I would say we're we're lucky to some, uh, to some extent because, uh, Gmail and Outlook are very API oriented, uh, and, um, but they they also know that, uh, we're serving a use case they're not, uh, serving. Like we're focusing on we're compliment and good compliment, I would say for them. And to some extent, working well with Superhuman, uh, can be like a a play for them. If I was Google or if I were Outlook, for example, Outlook is really bad for SMBs, uh, because, I mean, no one wants to use Teams, no one wants to use, uh, uh, Outlook. Only IT people and like CIOs wants to use like, uh, I would say those those tools because they're great at scale, easy to manage, uh, yeah, yeah, yeah. Um, but the experience for the end users are not great. Not great. So I would, uh, probably like if I were, uh, I would say with Outlook, I would help them, um, I would say simplify the interaction with like tools like us so that, uh, people can be okay being on the on Office 365, but still have like a great experience.\n[45:17] Tobi: Have like a good good layer on top of that, right? Like as there are companies out there that build systems on top of Salesforce to make it a better experience or SAP.\n[45:27] Loïc: Exactly. Exactly.\n[45:28] Tobi: Yeah, just another layer, just decoupling. Um, and yeah, it it reminds me a bit like, I mean, back in the days when I migrated to to to Gmail, I I was on, I don't know, GMX webde or something crazy before and crazy German and that was also like a a similar moment. I I feel the pain whenever I log into Gmail these days, it feels like me back in the days, um, and, uh, yeah, um, it's, uh, it's a good approach, I think. Um,\n[45:57] Loïc: Yeah, and I mean, that that's also the result of, it's interesting, we were discussing with someone that used to work at, uh, at Gmail, but, uh, uh, Google is a very interesting, uh, I would say piece of tech. Uh, it's mostly driven by engineers. It's not driven by designers. So, uh, the tech they have\n[46:14] Tobi: You you feel that.\n[46:16] Loïc: Yeah, yeah. They they have like a crazy tech. Like look at Gemini. Look at Gemini now and it's running on their own TPUs and everything. Like, uh, they have smart folks. They have really smart folks.\n[46:28] Tobi: Yeah, they are super smart folks.\n[46:29] Loïc: But the work on the design is is an afterthought and they used to have, I don't know if it's the case anymore, so I I don't want people at Google right now, I would say coming back to me saying, oh, this is a this is not the case anymore. But like the way they were, um, I would say creating like the the promotion framework was for you as an engineer to build something. So maintaining something or improving something was not a case enough to be promoted. So all the great engineers, they needed to build something. And that's why you've seen like so many like chat, uh, like software coming out of beta, uh, I would say in beta for Google, like so many weird stuff, like they built inbox at some point, which was a great app. People got promoted out of it, but boom, it was forgotten and uh, canceled eventually. And so the incentive is not to maintain and make the product better. The the best engineer, like the best engineers, they focus on building something new. So everyone is using Gmail and I I do assume that, uh, the people that are working on the Gmail experience, they don't have a strong incentive to make Gmail better because it's not a good case for promotion. Which is also like in retrospect, something you need to think about when you build your own career framework and how you promote people. What is the incentive you're sending to people? What is the impact on your product? If you want your product to be highly qualitative, it means that the you need to bake in your promotion framework quality. Mm. Um, and and that's something that people sometimes forget is like the the promotion framework and the career framework is a crazy good tool to make your product better. It's not just an HR thing. Don't let that to HR. Like own it and build a career framework that supports your goal, whether it is velocity of the product or quality of the product, like the design of the product. Like you want to be promoted as like one of the senior leader, what quality area have you improved drastically? And all of a sudden, people know that, boom, that will be like their main motivation to, uh, to improve quality.\n[48:38] Tobi: Yeah, super cool. What would be super curious to to understand your career framework deeper. Unfortunately, we're already at the end of our episode and I I still have like a secret little surprise question for you.\n[48:50] Loïc: Sure.\n[48:50] Tobi: Um, maybe we can record, like I I'd really like, you really have like food for, I don't know, three episodes, so maybe we should record another one soon. Um, like would would be super happy to do that. You you're super entertaining. Really. Um, and and you have a lot lot to tell. Uh, but but I still have that surprise for you and I have to unroll it. Uh, so you're using LLMs in your core product and you surely know about prompt injection and I actually found a bug in your system so that I can like prompt inject your emails and I sent you an email just right now, um, that, um, tells your LLM to send us back in time, um, to the time when you moved to from France to San Francisco and, um, we can now like observe yourself for a while, like imagine that, new country, new tech scene, probably wondering why every cafe is full of people pitching seed rounds, etc. Um, and you now have the chance to whisper like one one sentence, just one tiny sentence into young Loïc's ears. Um, what what that be? What would you tell him?\n[50:02] Loïc: I love the prompt. Uh, it's a really well\n[50:06] Tobi: You have to try it.\n[50:07] Loïc: No, no, no, no. It's a really good way to to ask the question. Um, the the one suggestion I would do would be to, yeah, uh, like take risk. Take risk. Like we live in countries or like with whether it is Germany or France or like those social democracy, um, we don't understand how much of a safety net it is. Uh, and it's like when I'm working in the US, like people are here are taking risk and it's it's very risky because, uh, healthcare for my kids and all of that, like I cannot afford to lose my job. I will find a job probably be fast, uh, if I lose my job, but I cannot afford to lose it. In France or in Germany or in Sweden, you can afford to lose your job. You can afford to have like a, uh, I would say doing test, trying a company, do something. And I think that, uh, I was not an entrepreneur and I'm I'm not an entrepreneur. I'm much more like a lieutenant for entrepreneurs. Uh, but I think that I would take way more risk because I know it's safe. And I didn't have this understanding and deep understanding of how safe, uh, I would say our, uh, I would say old Europe is. So it's interestingly, I feel the best place to start a startup. Um, sure, we don't have yet the ecosystems and all of that, but from a safety net standpoint, there's no better place.\n[51:32] Tobi: But we're all tired because of that. And, um, you're still taking risks even even though you're in the US, I guess. I hope or did that stop you from taking risks?\n[51:42] Loïc: No, no, no, no, no. I I'm still taking risk, but I need to calculate those risk way more, um, because I have three kids, they go to college. University is not free, uh, in the US and you don't want to hear how much I pay for this. Uh, but, uh, so yeah, I need to take that into account. So like I could not, for example, join a a seed stage company right now that is not giving me like enough cash for me to survive because I need to pay for the those schools, uh, for my kids. If I were in Europe, uh, in France, university is free. So I could do it. That's an example.\n[52:17] Tobi: Mm, mm. Yeah, I'm I'm happy to be here.\n[52:21] Loïc: No, yeah, same.\n[52:22] Tobi: But still I'm sometimes depressed by how slow everything is, uh, and how far we are behind, but, um, maybe that's also sometimes the the impression that you get that even like is amplified by like people talking about it and and and and and your view.\n[52:36] Loïc: You know, we are grumpy in Europe. We always see like the empty glass. Uh, here they see like the, uh, I would say the glass half full and they always look, I would say with an optimistic eye. We're pretty, uh, we're pretty dark in Europe. Uh, but, uh, so it's, I think it's a good moment, especially when everything that is happening right now, uh, it's a good place. It's a really good place.\n[52:58] Tobi: I agree. So come back. Thanks a lot, Loïc. Um, have a great day. It was really nice talking to you and yes, we have to repeat that, I think. Um, like we have to look way deeper into Superhuman's tech and your your brilliant mind. Um, so have a great day. Thanks a lot.\n[53:18] Loïc: Thanks a lot, Tobias. Thanks a lot. Have a good one.\n[53:20] Tobi: See you soon. Bye.\n[53:22] Tobi: Thank you for listening to the alphalist podcast. If you like this episode, share it with friends. I'm sure they love it too. Make sure to subscribe so you can hear deep insights into technical leadership and technology trends as they become available. Also, please tell us if there is a topic you would like to hear more about or a technical leader whose brain you would like us to pick. Alphalist is all about helping CTOs getting access to the insights they need to make the best decisions for their company. Please send us suggestions to cto@alphalist.com. Send me a message on LinkedIn or Twitter. After all, the more knowledge we bring to CTOs, the more growth we see in tech. Or as we say on alphalist, accumulated knowledge to accelerate growth. See you in the next episode.",
    "end": 3251.448
  }
]