How to Hire Great Engineers: Hiten Shah’s Proven Method

I think culture is bastardized, so to speak, because nobody really knows what that means. People don't realize a culture evolves and it doesn't really get developed. At Nira, which we ran for quite some time, my co-founder and I, like five-plus years, and we pivoted and all that stuff, we never had a single offsite. We found something fundamental that we believe in, that we think is the right way to think about a company, and that is environment fit first. Measure twice, cut once culture, which basically meant that when we hired engineers, we would go through this very specific process. If it was a front-end developer, it was a to-do list app. If it was a back-end developer, it was a link shortener. And they got a two-page brief on what we wanted to build. And all we asked for is their plan. Then we would get on a 90-minute call and give feedback as if they were on the team. And then we would rescore after that, because some people take feedback extremely well and fix everything. Then if that all checks out, then we'd move to the team meeting and the culture warm and fuzzy, because if they can't fit in the environment and the way we work, then the culture, if we like them, it doesn't matter. You can go and apply this in a thousand-person company, actually.

This is WHO We Hire. I'm Luiz Cent, your host, and today we're joined by Hiten Shah. Hiten has been an entrepreneur for over 20 years. He founded Crazy Egg, KISSmetrics, and Nira, which sold to Dropbox, and now he's all in on Crazy Egg. I'm really excited to talk to him about hiring, building teams, and how it's all changed in the age of AI. Hiten, welcome to WHO We Hire.

Thanks for having me.

So, when you were four, your dad told you to become an entrepreneur. Very untraditional. You went and did just that. And a lot about being an entrepreneur is building teams and getting people to follow you and join you on your mission. How has that changed over the last 20 years? And having worked at a thousand-plus-person company now through acquisition, what have you learned that you can now apply to Crazy Egg as you're building out the teams today?

Yeah, you know, at Dropbox, I had the pleasure of working with a lot of people and not managing anyone. They made me a product manager and really wanted me to focus on starting new things, working on new products and features and shipping them. And that's kind of a DNA that's very hard to get at a larger company. So when they acquire people like me, they either make us manage folks... I had a co-founder who was able to manage folks and all that. We both were. And in my case, they wanted me to work on product. And it was my first job in my life, and Dropbox at the time was somewhere between 2,000 and 3,000 people when I joined. I think the thing I learned the most is that what executives say and don't say gets read, so to speak, between the lines, and not between the lines, sometimes literally, in all different ways, by the people that are actually in there doing what I call the dirt. And it's not that executives aren't doing the dirt. They're doing a different kind of dirt, but they tend to have big machines to move the dirt. While those of us on the ground who are trying to build a product, work with engineers and designers and, you know, marketing, sales, all the people, compliance people at a larger company, we have to move the dirt usually just with our hands, right? AI can help and all that. And what I mean by dirt is, you're taking some information or some asset and moving it somewhere else, convincing people a bunch of stuff needs to be true or whatever, and then you're going to your team and trying to execute on it and making sure everyone's on the same page and aligned. But then leadership could come in and just say something, and then everyone's scrambling, wondering what happened, right? The classic example is layoffs. The second you start having layoffs, or even if you let go of one person on a team, or one person disappears, so to speak, everyone's freaking out about their jobs. So it's awesome that when you started, you started saying mission and this and that. And I'm like, what's changed is self-preservation has turned into the way people survive in companies. It used to be you survive because there's a mission. Now you survive because of self-preservation. And I think that evolution was from lots of us being technologists in tech to lots of people who are in tech who are not technologists. They're not here for the technology. They're here because it's an industry and they want a job and they want to make money and they want to make more money, right? And I'm not saying at all that that's bad, but you asked me how it evolved. Well, we used to geek out together. Now it's not geeking out together, right? It's not like a cottage industry or a bunch of people blogging or whatever. It's an entire production, and arguably a Hollywood or investment banking type and style of operation is necessary. So when you say mission, I'm like, yeah, screw the mission. If you're leaning on that to lead people, that's not going to work alone. And it's great if you have one, and you really have one, like SpaceX AI. Yeah, they have one. They're going to take us to Mars. Like, hell yeah. I get that mission. But literally, we're in a world where technology is everywhere. So a mission like Google's wouldn't work today. "Organizing the world's information." Yeah, it's already organized. Well, what more you got? So I would say those things are the target for what a mission is, and whether people will come on for a mission alone has completely moved. Even the labs have this problem. They can say, "Oh, we're doing frontier AI." Okay, congrats. You're one of maybe five to ten companies that can do that, but then you're also telling us that AI is going to take all our jobs. So now how does that impact the people that are working at your damn company?

Yeah, I've never heard anybody speak about it like that. Most of the folks that come on the show say, "You need a mission. You need alignment around it. You need to push this." But it's definitely a great take, that the mission alone is not enough, especially if it's not taking us to Mars, right? How do you...

That's because this is an industry that makes people rich, and it's not just the founders that get rich. So a lot of startup missions, like during zero interest rates, were basically, "We're going to make you rich, your stock's going to be worth something." I never played that game because I never felt comfortable promising people that.

You never had the stock options and the...

No, no, we had all that, but we never promised people that that's how they're going to get rich, right? We promised people, you're going to work in an environment that you're going to like, but if you don't like it, then you don't need to work here either. So we really focused on what environment we're creating for the people there. Now, I go to an extreme. At Nira, which we ran for quite some time, my co-founder and I, like five-plus years, and we pivoted and all that stuff, we never had a single offsite.

You never had a single offsite. You never once...

Never once got everyone together in one place as human beings, IRL. If you look around me, I have no problem with human beings.

Yeah.

I want to hang out with the human beings I work with. I would love to. And it's a common recommendation, right? Do an offsite every year. But you know what happens with offsites? People save certain types of work for an offsite. The classic one is strategy work. "Oh, we're going to have the offsite. Let's do the strategy meetings." Wait, hold on. Strategy is constant now. So another thing that changed is we used to be able to do this work in chunks and get away with it. But when all these signals are coming and the world's changing and every company's launching stuff, and AI and code and this and that, you can't wait three months, six months, a year, whatever, to have some strategy meeting. And that's just a classic example of what people save for that. And again, I'm not against people meeting up. I'm not against camaraderie at work, any of that. But the inconvenience of trying to fly however many people somewhere and have them meet up, it's an inconvenient thing for their lives.

Oh, we're trying to plan our first one.

Tell me about it. [laughter] And if people want that, then at the companies that I run, that's not what you're going to get. And that's what I mean by environment. This is the environment we created. And again, if someone sits there and says, "I want that. Convince me why I don't need that," I would sit there and be like, if you're someone that gets energy and really feels like you build relationships because you're IRL, more power to you. I totally understand that. I can do the same. But that's not how we run our companies.

I love it. Yeah. And you talked about environment there, and I heard you say on another talent podcast that it starts with vibe, then environment, and then culture. Can you talk to me more about that?

Yeah, absolutely. So the reason I say it starts with vibe: one, it's a cool word to use these days, so I'm going to use it. I think culture is bastardized, so to speak, because nobody really knows what that means. And people don't realize the culture evolves and it doesn't really get developed. You don't develop culture, but you can develop the environment, and the environment leads to the culture, because you're incentivizing certain things and not others, right? If you have offsites and you're a remote-first company, you're incentivizing something when you have offsites. If you don't have offsites, you're incentivizing something too, which is personal space and implicit self-preservation, because I'm not controlling your time. The other silly part of this for me is basically that you go to school, you're told when to go, what classes to go to, at most schools when to eat. I mean, everything. When you go to work, why are we doing the same thing? We're not at school. And when you do some of these things, that's what it turns into. So to me, going back to vibe, my vibe is we get on calls, we talk, we get stuff done. My vibe is actually don't even get on calls unless we need to. I try to cut as many meetings as possible, but I never have to cut them, because the companies and the environment I've created don't incentivize meetings at all. Zero. If we have no meetings and we can get stuff done, everyone's happy. Everyone in the company is happy.

So then your calendar is only the necessary ones now.

I don't know. I could start counting, but it's not important. My recurring ones, yes, I have meetings. I love meetings, but if meetings don't get stuff done, I hate meetings, right? And I love meetings and I like people and I want to get stuff done with people. This is not "I don't like meetings." This is, hey, are meetings helping us do what we're trying to do or not? That's why, yeah, I have recurring meetings. Every single one is productive, right? And if it's not, I cut it, because no one wants to be in meetings. I don't know anyone that wants to be in meetings all day, unless they're in sales and all the meetings are leading to something productive. I'm down. And I'm also someone that could do 25 meetings in a day. Not a problem. Zero problems with that. I've done it before, right? But these are customer development calls or sales calls or a regular recurring meeting in between all that, whatever. That's cool. So here's the thing. This is why I'm saying all this. There's people that are going to be watching this, some people listening to it in their ears, right? Have you figured out my vibe?

We're getting there.

Right? [laughter] So I've even talked about the environment, right? So let's say you're in an interview at one of my companies and you happen to be talking to me. It's exactly what I'd tell you. This is what we believe, this is it. But it's my vibe, my co-founder's vibe, whatever that is, that's what it is, right? And that creates the environment, because it's my preferences. It's not yours. And yes, if you align with my preferences as a founder, leader, whatever, great. And if you think about even large companies, it works exactly the same. In fact, let me explain. You can go and apply this in a thousand-person company. No problem. Because why do people say, "Oh, you should be on that team, not that team"? The leader on the team is making the vibe. So what I learned at Dropbox, because I happened to work with someone I've known for a very long time, probably 20 years at this point, and he came on and he became my boss...

Is it Morgan?

Yeah, Morgan Brown. And right away, I saw him doing all the moves to basically get his vibe across and get people to understand it. And I'll give an example. He has a Morgan Brown manual. A lot of people have these user manuals. Some of them are really bad and some of them are just exceptional. His was straight-up exceptional.

What are some things that he did differently?

It was just honest to him. One of the lines was, "I don't believe in politics." It might not have said it exactly like that, I don't know how he said it, but he said it. And so here's what was awesome. My point was, he made the manual. He came in on a team, which was our entire product and growth team, so lots of people trying to do new things, and the manual and those lines he had in it allowed me, who is a fanboy of Morgan, to go to other people on the team like, "No, no, no, this is what he means," or "Hey, he said this." I started handing people his manual, because I don't know if they had it. "Read this." Then, because it was the early days of ChatGPT and all that, because this is a couple of years ago now, so 2023, 2024, I made a Morgan bot. So when someone's like, "Oh, give me feedback," and I know that he's actually the ultimate vibe owner here, I'm like, "Oh, I threw it to Morgan bot. Here's the feedback. Forget my feedback. Let's go with his." He's setting the vibe, right? And I bought it 100%. So he set that when he came in, and I watched him do it as he came in. And even bringing all the issues to the surface: we had this massive meeting at an offsite talking about the issues, with people standing around doing it, not even sitting down. It was just a fascinating thing to watch, because he was bringing his vibe, which means he was bringing his energy and how he thinks about the world, his principles, all this stuff. I count that as vibe, and the reason I say vibe is people feel it more than they read it. And that's vibe. That vibe created the environment. That vibe enabled me to help with the environment, because I knew his vibe and I bought in, right? So people are buying into a leader's vibe on a team, and then they're buying into the environment the leader created, and then all together they're creating a culture. That's kind of my breakdown.

I love it. Yeah. There's some folks who want the 996 in person in San Francisco, and they may need that. There's some that, you know, love you but never want to see you in person, because they're taking care of their kids or they live on a remote island, for their own personal reasons. And you're able to attract the best talent for your team by creating that environment for them. I love that take, because I also think culture is like, "Hey, it's these values we have. It's this mission." We talk about it once a year, but we never really promote people on it or fire people based off of it or hire people based off of it. So with that, how are you creating your teams today? There's AI that is supposedly taking all of our jobs and making developers ten times more effective. How are you looking at it in terms of revenue per employee? How fast do you need to move? Who do you want to attract in your new vibe here at Crazy Egg? And what are some of your goals?

Look, I haven't changed anything, because I always started with the environment. So if I change the environment, that doesn't count as change to me, because the change I'm talking about is how do we hire? We haven't changed anything. What does that mean? Well, if the environment changed, are we changing how we hire? Well, if we already hired based on environment, it doesn't matter. We're aligning who we hire against an environment. So now, do we need to evolve the environment because of AI? Absolutely. Do we change our processes because of AI? In our case, no. Because we found something fundamental that we believe in, that we think is the right way to think about a company, and that is environment fit first.

So back in the day, when it was a back-end or front-end engineer that you were hiring...

We had a measure, we still do, a measure twice, cut once culture, which basically meant that when we hired engineers, we would go through this very specific process. We would give them a two-page brief. It included some gotchas and stuff, like vacation time and stuff like that. And so if it was a front-end developer, it was a to-do list app. If it was a back-end developer, it was a link shortener. And they got a two-page brief on what we wanted to build. And all we asked for is their plan. Because on our engineering team, unless you have 90% confidence in your plan, you don't get to write production code. So you could write code before that to get 90% confidence in the plan. Do whatever you want, but it doesn't ship to production. That's not the point. So you don't write production code until you have 90% confidence in your plan, and your manager, not hiring manager, your manager, understands that. So what we would do is we'd screen someone, usually not talk to them, just screen them, sometimes maybe depending on the role. Then we would give them one of these two, when back end and front end were a separate thing, which they still are technically, but we don't need to get into that. I think it's a really good way to think about the example because it's two different types of engineers, basically. And we would score their plan when they delivered it. Then we would get on a 90-minute call and give feedback as if they were on the team, from the person who would probably give you the feedback if you were on the team, on your damn plan, which is your work, right? It's the start of your work. And then we would rescore after that, because some people take feedback extremely well and fix everything. So we just rescore how they take the feedback, and so you can see how much we did there. Then if that all checks out, then we'd move to the team meeting and the culture warm and fuzzy crap, because if they can't fit in the environment and the way we work, then the culture, if we like them, it doesn't matter, because that means that they won't be productive in the environment anyway.

So you're taking a little piece of the environment out and testing around that.

And finding the right little piece. So let's say we're hiring product people. We were a research-first company, which means we would do user research on all these things. So all of our tests, depending on your role, coming in on product, would be oriented around making you do mock research or whatever. And mock meaning, go do the research, or here's a bunch of research, go analyze it, what would you do based on this research, things like that. And that was the same process as we used on engineering in terms of tactically what we did: score them, interview them, feedback, and then rescore.

I love it, and I love your answer so much there, because I watched one of your podcasts with Claire Vo on how much changed in the last year with AI, and you shared some valuable lessons on how to think about it that will stay relevant. But the interface changed, the models changed, what we're using, everything's changed in a year. But the process for hiring and finding and attracting great people, that's human fundamentals, I think, and a lot doesn't change there. So to me, just the combination is really exciting, and to hear how you think about it and share your learnings from it. And I read somewhere that you also follow the WHO method. So this sounds like a part of the WHO method, where you're doing an assessment.

We invented the method before we learned the WHO method. I think the WHO method uses similar things, but it's not quite environment fit. And for anyone listening, the WHO method is a way to hire executives. That's the main thing. And it's a very involved method. But the key part of the method is actually to find people who have done it before, and literally don't show them the job description, the internal one, until the last meeting. And you've done a good job if at that last meeting they feel like that job description was written for them. So that's how the WHO method works, and there's a lot of buy-in that you get when you do it. But what you're really doing is probing the person very deeply around how they've done what you need them to do for you in the past. And so it's very different from environment, because it's oriented around understanding the person's capabilities and things like that. So when it comes to senior execs, the stuff I'm talking about doesn't really work with execs, but the WHO method does work with execs. The reason the stuff I'm talking about doesn't work with execs is my point earlier: they're making the environment on their little team or big team or whatever. So it's much trickier to be like, "Hey, here's the environment we have. Can you work in it?" That's not the goal, right? The goal is more, this is the company, have you done this crap before? Because if you've done it before, let's go talk to you about doing it before and whether we honestly believe you know how to do it. That's all you're trying to figure out. But if you're already self-selecting someone who's done it two or three times, you're already bringing in the right candidates. So the WHO method has all these nuances and actually has very little to do with environment. It has more to do with environment creators, and finding the right environment creators for what objectives you have, right? So that's why I'm saying I would split those two up. They're very different. But we came up with the environment stuff based on studying how an engineer I've worked with for 20 years actually does his own work, and he used to manage a team. So even to this day, he manages his coding time in 15-minute chunks.

Okay.

He plans out code in 15-minute chunks. Again, most engineers cannot, will not do it. But the planning part is what we realized is how we like to run engineering, because we were always measure twice, cut once. Period. Even today with AI, same. In fact, with AI, that's the way, which people don't realize. But even the environment fit stuff, I talked about the planning stuff, this is all how you use AI. But we didn't have to change anything. We just accelerated everything, because this is already how we work. Most companies out there do not work like this. They don't plan their crap before they build it. But AI needs you to plan it extremely well, because it's going to go build it. So that's why not much has changed on most things. And then the WHO method, I would say, is more for hiring executives. Big fan of it, though. Big, big, big fan. I try to tell everyone this is the only way to hire execs. [laughter]

But you don't incorporate parts of it into hiring engineers?

No. Environment fit trumps that method, unless it's execs.

Wow.

I don't really care if someone did it three or four times before, in most cases, when I'm hiring an engineer. I just care that they can work in the environment we built, because that's more important.

Because they're going to have that coaching or mentorship there, or you can teach them.

It doesn't even matter. We don't care about the aptitude. That gets sussed out over time. What we really care about is, do they know how to think? Because if someone knows how to think, you're right, we can teach them. But we don't even need to teach them. That's the funny thing. That's why when you say, "Oh, you can teach them." No, no, no. If they know how to think, we're good. And this is just testing if they know how to think. Think about it, right? Any human being that does not plan before they start building something with an outcome doesn't really know how to think, because they never realize that. And when I dig in and meet people who can do that, I'm like, "Wait, you're really good at this. Are you doing this piece of it?" And they're like, "No." And then I ask, "Wait, do you do this? Do you look at this thing? Do you look at that thing before you even start?" They're like, "Yeah." That's planning, buddy. So even people who think they don't do it, but know how to critically think, are doing some piece of it implicitly in their brain or with some process. They just don't describe it as planning. Maybe they're allergic to it. Maybe it comes natural to them. Maybe they learned it. But that's the kind of thing.

Are you looking for curiosity there?

Maybe. Maybe. I'm just looking for people with literal raw learning power. I don't care about anything else. If curiosity is the way we label that, cool. But it's raw learning power. One thing I like to say about my co-founder from Nira is she learned faster than anyone I've ever met before her. And that's why I didn't worry about her learning tech, because when I met her, she didn't know anything about tech in terms of building software. But by the end of our journey, even within the first year or two, we had built a bunch of stuff. She could run product. I wouldn't care. And that's a very precious piece for me. Within a year or two, I'd say easily 18 months, if not tighter, of us building stuff, she was on it. Even today, I could show her stuff and she'd tear it apart. But this is a human being that worked in the wine and spirits industry as an executive. She was on the line when wine bottles would fall, and they had to make sure that they didn't fall, the bottles or the liquor bottles or whatever. She did a bunch of joint ventures with a bunch of celebrities and things like that and made all that stuff. But now she's in tech, right? And within, like I said, a year or 18 months, she had already figured out whatever she needed to. Of course, I was helping her. We had a team, in the sense of helping her learn. So to me, sure, curiosity, all those things are important, but someone's speed from not knowing something to knowing it well enough for the job they need to do, that's what I'm really looking for.

What are some things you think really accelerated her learning, other than, you know, osmosis through you and the team?

It was just her. She just had this thing. Some people just have this thing of, I don't want to be wrong. And instead of lying or faking it, because a lot of people do that... Like, you mentioned some people you worked for earlier, when we were talking before this. Some of them lean on fake it till you make it more than they lean on I'm going to figure it out. There's fake it till you make it, and there's I'm going to figure it out. Eventually you have to get to I'm going to figure it out, if you're a person of substance, right? But a lot of people lean on fake it till you make it hard. Other folks are more like, I've got to know this stuff or I don't believe I can say it.

Yeah.

So her energy went into knowing it as fast as possible. So she would ask us questions. She wasn't afraid of that. She never said, "Oh, this is a dumb question." She'd just ask it, because she knows there's no dumb question. It doesn't matter what environment you're in, there should be no dumb questions. But again, that's why she was my co-founder, because we created environments where people felt comfortable asking dumb questions. Because how do we learn if we don't ask questions? So how can we say that there's a dumb question, right? Maybe there are no dumb questions, only dumb people. I'm just kidding. But you know what I mean.

You have co-founders that you haven't met in person. Talk to me about that. How have you found them?

Just organic. You just bump into people, right? You start working with them. It's online. I've been online for a very long time, right? Through Twitter, through LinkedIn, through all of it. I've had a computer since I was like six. So, yeah. I mean, I think if you're perpetually online, you bump into enough people, and you know how to do that, you build that trust. These days, if you talk to the kids, they're usually in Discord, right? Hanging out, talking to each other. They love Discord because it has voice chat and screen sharing. And even if they're not gaming, a lot of them go in there starting with some random games. Even the casual games have Discords now. And obviously AI is changing a lot of that, but that's how you learn how to communicate online.

Now you have text and all that.

Yeah, I use all of them. I use wherever the communities need to be for whatever I'm trying to do. I was deep in crypto and was a degen in that sense. Deep as in not a business in it, but as a trader and all that, too.

Talking through those, and community, right? So we're seeing HubSpot and these big brands build huge communities and have conferences. Are you seeing that in the future with companies, or are you seeing more small, nimble companies get formed and more entrepreneurs than ever in the next 10, 20 years?

I have no idea. I think usually technology opens up more businesses, not less. And this is just a weird technology where we're confused. I don't meet a single person that's not confused today. [laughter] Sometimes they have a lot of confidence. They want to fake it till they make it. But once you get underneath it, they're like, "Yeah, I'm confused. I don't know." In the past, I would know three to five years out. If I do this, this, and this, this will likely happen. Today, no one can tell you that. No one. I don't know anyone that is like, "Oh, three to five years, here's where I'm going to be. Here's what I'm going to do." Of course, if you're going after a profession, you're in college, you've got an 8-year, 10-year, 12-year, whatever journey, sure, more power to you, you can do that, right? But if you're trying to build a business or work in a business of any kind, I think it's very tough to have any prediction. Even the medical profession, even though I just said someone can have a plan, the medical profession is changing. So in the next 10 to 12 years, or whatever the journey is for someone who's in college trying to be a doctor, what's going to change? Well, a ton's going to change about the technology they have to use, if nothing else.

Oh, absolutely. We're helping a client right now hire developers that are building AI software for surgeons, right? And it's insane. But I think you're going to make some strategic bets. I mean, you made some with crypto. So with AI as well, it sounds like you're all in. I got to watch the one with Claire and how you're using GPT. How has that changed over the last year, and how you're using AI? Did you create a bot of yourself, and then your team goes and talks to your bot, or...

I think that's dumb. I have 18 machines sitting here.

18 machines.

I have 18 machines. I have two DGX Sparks right there. I have a cluster of three Mac Studios there. I have four right here, two over there, a bunch of Mac minis, a couple of other random things you've never heard of that are AI machines meant for home use or enterprise.

Do they have names? Are they embedded in your team? Talk to me about...

They don't have names. I don't do email or calendar productivity with them. I think that's one of the weakest use cases for AI. Although it makes sense, right? "Monitor my email." I'm not saying it doesn't make sense, but I never gravitated towards that. So when OpenClaw came out, I threw that on these machines. Now that we have Hermes, I threw that on these machines, trying to do actual workloads, trying to use local models. I use frontier models with them as well. But you ask me what it is? Well, when I was six, I got a computer. Now I'm 45. I've got computers. [laughter] What changed? Right? In the middle, I didn't have 18 computers. I always had like five to ten, because I'd mess around. Back in the day, it was actually 80, 90% PCs. Microsoft used to love them. But then post Windows 95 and this and that, it got really bad. And then I'm like, "Oh, these Macs are cool." So then I went more Mac, and you couldn't see me with a PC running around. But now I'm playing with Omarchy from DHH, the Ruby on Rails creator. And it's amazing. It's a new operating system. And I think any operating system vendor, like Apple and Microsoft, should be like, "Oh crap," because it's an agent operating system and it's insane. But again...

Right, the founder of Ruby on Rails has built an operating system.

He's built it, and it's called Omarchy. They might say it differently, but I say Omarchy. O-M-A-R-C-H-Y. And you can install it on certain Macs and you can install it on PCs, and it's an entire operating system that's just designed to let you do things with AI and be more malleable, but everything works, and it's really clean. He just announced, even today, a bunch more funding from SpaceX AI, so he can keep building it. So I think there's a rebel alliance even on operating systems now because of AI. And I think it's because of AI, because he was able to build it because of AI. But he also came out a couple of weeks ago and said engineers should not be writing code anymore, or something like that. And it caused a little bit of controversy. But I think he's not wrong. Most of the engineers I know don't write the code. They don't read the code. If they need to, they will, but they don't anymore. So that changes things. So when you ask me something like what I think you're going after, and you used the word curiosity: I learn by doing. A lot of us learn by doing. I also learn by thinking, don't get me wrong. I enjoy that as well. And so in this case, I wanted to understand AI. I can now talk to researchers about AI. A year ago, I couldn't, because I didn't understand what they were dealing with and how to do training or reinforcement learning or pre-training or any of that stuff. I have machines. I tried to do all this stuff. I also don't need to read or write any code, and I never did. Well, I did a little bit, but not a lot. It was never my profession or what I enjoyed, and I never learned it in school. But now I can go hack together whatever, conversing with an agent I have running on these machines. I'm the top user of this product called Dark Bloom, because it allows you to put your machines into a pool and it gives you some money. It's not a lot of money, but it's the easiest thing I found, and the closest thing I found, where I can put my machines in a pool. It runs through OpenRouter, and my machines are doing inference for people I don't know, and I'm getting paid for it, right? I'm the top user because I have five of my machines, I used to have more, sometimes less, running Dark Bloom when they're not doing other stuff. So you ask, and yeah, it's curiosity. I know you asked about the word and I dismissed it earlier, but in my case, it's curiosity. But the curiosity started when I was very young, even before I had a computer, because my favorite book was The Way Things Work. It's an awesome book.

I need to put it on the list. So, is it harder for you to bring people into this vibe and environment and ultimately help shape your culture now? In a sense, isn't it easier to just work with the machines?

No, I don't care about that. I think we still need to work with humans. It's funny. We're in October now, and people are understanding one thing I already understood back in January or February because of these machines and OpenClaw, which is that you can put an AI in Slack. You can put that AI in a channel and you can start talking to it, and you can put humans in the channel and they will learn AI. I was doing that since January, February of this year. Now people are figuring it out and building this and saying, "Oh, you AI-pill a company by putting the AI in Slack." Yeah, no kidding. I used to do pair prompting, where I would teach people AI by showing them how I do AI in a Slack channel with them. Now that's a norm, right? So what gets me excited is the fact that I'm figuring this stuff out before anyone else that I see talking about it, and I have receipts. I have articles where I wrote about this pair prompting and stuff like that, and now it's being caught on. Great. So for me, my satisfaction is the fact that I understood where the world is going before other people did. Whether I capitalize on it or not is a whole different conversation, but I get my jollies off because intellectually I can say that and it's true.

So do you like hiring people that challenge you intellectually as well?

I don't consider it a challenge. That's not the point of interaction. It's not to challenge each other. It's to make each other better. Some people love using that word. It's like, great, challenge me. Go for it. Let's debate. That's not my vibe. It's more like, let's talk about this. Let's go align on the common goal. Let's go brainstorm. Not challenge, not argue. Even a debate is not interesting to me. But if we have a common goal, then let's discuss the ways we're going to get there. And so that's the way I think about it, more than hiring people that challenge you or whatever. But I think a lot of people need that, because they think being a founder is being sure of yourself. I even see founders of public companies acting that way, and I'm like, you're silly. You're just silly. You don't know anything either. Why? What? Don't claim that.

Do you think that's because they have to focus on short-term results and deliver on stock prices and have this facade, or... I see you shaking your head.

So I think a lot of people in that boat are one-hit wonders. Something worked, and they're just trying to chase, trying to find the next thing that worked, and it's been 10, 15, 20 years, but the one thing that worked just works so damn well, right? And that was timing and luck, but it was probably mostly luck, because timing is luck in a lot of ways. So I'm shaking my head because I'm like, nah. These people I'm talking about, they don't know any better. And my clearest sign, and people can argue with me about this, is when I don't hear someone admitting they were wrong. I just want to hear you admit you're wrong. And a lot of these people running these companies, one, never admit they're wrong, but two, if you don't admit you're wrong, you're not going to get anyone telling you you're wrong. And that's the part that I try to identify, whether a company will be successful or not, right? A good example: just the other day, SpaceX came out and basically said, with Grok bot, which is their agentic AI agent thingy, "We're going to route it to the best model," and if the best model is Opus 5.5 from those other folks, so be it. So it's not just that they were wrong, but that's what I mean. Elon doesn't have to admit he's wrong. He just has to make moves that imply that he understands he was wrong, or they were wrong, or they could do better. Whatever way you want to describe it. That's what I mean. So I don't mean someone has to admit it. I just have to see the moves where, oh, you were clearly wrong and you corrected it. So I'm actually not looking for wrong. I'm looking for the correction. When you see an apology on Twitter, X, whatever you want, that's not what I mean.

Yeah. Yeah. You want to see actions correct the wrong things, even if they don't admit they're wrong.

And I'm good. I'm not trying to get everyone to say, "Oh, I'm wrong. I'm wrong all the time." But if you correct them, then obviously you're willing to admit to yourself that there was a problem, you were wrong, and you fixed it. And this is an extreme one, right? And the genius of it, though, is by doing this, he's essentially putting the demands on his own team and saying, "What percentage is being routed outside of our models? Oh crap, it's that percentage. Cool." Now he has a metric.

Yeah.

This is genius. It's straight genius. Again, I don't know if that's why, but if I put a genius hat on, I say, "What would a genius do?" A genius would do something like that. He's probably got like ten other moves or steps, because it's Elon. But oh my god, this is a leadership lesson.

Love it. Love it, Hiten. And so, as we're wrapping up, where can folks go and find you? And who should go and find Crazy Egg?

Yeah. Crazy Egg's mostly built for people who are looking to optimize their website. That's what we've always done. So if you've got a website, check out Crazy Egg. We have lots of stuff and are much more cost-effective than any other solution we've found, although people don't know that, because we're self-funded. We don't have the same criteria as other folks, and we've been around 20 years. Then Twitter. I tweet a lot. So @hnshah on Twitter. The reason my name is that is because the first car that my dad bought in America, the license plate was HN SHAH, and N is my middle name. So that's why that's my Twitter handle instead of Hiten. But if you don't like Twitter or don't want to go there, or want to find me elsewhere, I own hiten.com. So just go there and sign up for my AI stuff, where I teach lessons every Friday. So even tomorrow, I'm doing an AI lesson. Today's Thursday. Not that this will be out on Thursday, but yeah.

Amazing. This will be out next week, and we'll put links to your Friday lessons there. And thank you so much for joining the show and sharing your learnings and stories, Hiten. I personally learned a ton, and I know our audience will too.

Thanks for having me, and thanks to everyone who's listened to me today about all this stuff.

Love it.