He Raised $26M Betting on Probabilities, Not Guesses | Koah Labs, Nic Baird

1. 0:00 – 1:50

   Intro

   1. SPSpeaker0:00

      Technology is moving at absolutely incredible speeds. Relative to 2024 to now, people were still very cautious about sharing information with an agent. That was the general sentiment in the market at the time. But nowadays, people are like, "Open Claw sure has access to my entire computer. No problem," right? [laughs] We don't know how the market's gonna play out exactly. We don't know what all the different players are gonna do. We don't have all the information about whether or not a product is gonna be successful. But you have to make decisions in a market that is uncertain. But what we can do is we can give probabilities to a thing. We can say it's very likely that the market will do this thing, right? In sailing, you have these decisions all the time. If you are very confident that the wind is gonna shift five degrees to the left over the next five minutes, right, you wanna go towards that wind shift, so you wanna go towards the left side of the course. But let's imagine that the entire fleet, everyone else, is going the other way. Are they being idiots? You kinda have to ask yourself, "Do they have information that I don't have?" If I'm wrong, the risk is so high because so many people are going this direction, so let me mitigate that risk. While I'm highly confident this thing is gonna happen, I'm gonna make a different decision than I would if I was just purely trusting my instinct because the math supports, the stats supports that I go over here. And so I think that is a great proxy for how really the best decisions are made for your own career, for your own life, for a business, is you have to look at the information that you have available, understand that you don't have all the information, and you have to make a statistical bet. You have to say, "I think it's most likely this thing will happen." And then depending on how confident you are, that's how much you invest in that thing. I'm Nic Baird. I'm the CEO of Koah Labs. Koah is basically the ad monetization platform for AI tools. We've raised $26 million to date. You know, we're reaching now single-digit millions of users, um, and growing very, very quickly. [instrumental music plays]

2. 1:50 – 5:23

   How a World-Class Sailor Decides Under Uncertainty

   1. SPSpeaker1:54

      A lot of people say that you shouldn't start a company, it's a bad economic decision, you know, all these things. But I like to think of things in terms of variability versus risk, right? And starting a company is a high variable thing, but it's... I actually think it's a low-risk scenario. Why? Because you learn so much being a founder of a company. You will become a more valuable asset to a company later and be able to make more money later, whether you fail or succeed. Sailing was by far the most important priority and where I was spending almost all of my time. Of course, sailing is many things. Like, sailing is also cruising through the Caribbean and, you know, having a cocktail on the back of a nice big yacht. But the w- the type of sailing that I do is closer to NASCAR or Formula One: fast, high-performance racing. Uh, growing up with my dad doing that as a profession is by the time that I got to college, I, I was able to go and compete at the Youth World Championships in two different, uh, classes in two different years and finished third and fourth respectively, and had won three national championships by the time that I got to college. [keyboard clacking] Chess is theoretically a solved game. You can play perfectly. In poker, there are too many confounding variables, and you're operating under too much uncertain information. It means that you could play, quote-unquote, "perfectly" in poker and still lose. Because there's uncertainty, there's a variable of randomness, there's a variable of chance. In sailing, you have these decisions all the time. When you are sailing in a, you know, a sea breeze environment we can call it, so what happens is the land will heat up faster than water. And so when that happens, you basically get a sea breeze. And so what we see that happen in San Francisco Bay Area all the time, for example, the San Joaquin Valley is, you know, 100 degrees, and the Pacific Ocean temperature is 60 degrees. And so that temperature differential is partly what creates such a powerful sea breeze and why it's so windy in San Francisco in the summertime. And so when you know those things, you can start to predict. We know that as the land heats up, typically when there reaches this temperature differential under these pressure circumstances, that's when we expect that sea breeze to come in, and we usually expect that it comes in from this direction, right? So you can position yourself well to be prepared for that to happen. That is a great proxy for how really the best decisions are made for your own career, for your own life, for a business, and you have to make a statistical bet. You have to say, "I think it's most likely this thing will happen." And then depending on how confident you are, that's how much you invest in that thing. So you're saying, "I'm 80% confident that it's gonna rain today," right? It's not really expensive for you to carry around an umbrella, so you're gonna bring an umbrella. I'm betting it's gonna rain, and if it doesn't, the decision that I've made to bring an umbrella is not that harmful. Whereas if it's a 10% chance of rain, you'd be like, "Ah, yeah, I'm gonna weigh off the trade-off of like, yeah, 10%. Uh, maybe I get a little wet, but yeah, I don't really wanna carry an umbrella around. Oh, that's inconvenient," right? So that's like a very low-stakes example of doing this. The high-stakes examples of doing this is I was actually saying no to sailing opportunities and professional opportunities to go get paid to race because I was too curious about what was hap- happening in startup land. The ambition in me was saying, "Great, I could become the best sailor in the world. Is that as important as becoming the 100th best business person in the world? No, it is not." If there was a 1% chance that you could make a trillion dollars, 1% of a trillion dollars is still $10 billion. Thumbs up, right? [laughs] I'll go after that. Whereas if you say, "I have 100% chance of making a million dollars," people would say, "Well, go to the million dollars," but the expected value is actually much better if you go after the low chance but high reward. And so that's the way that I like to think about a lot of decisions in my life and also within the business and, of course, within sailing, too, which is where that article came from.

3. 5:23 – 8:34

   To Become a Builder, Be Around Builders

   1. SPSpeaker5:27

      And after all that, I thought, "I've gotta get back to where the builders are, the people who are creating things, and, and find a very talent-dense community." So I sought out in-person San Francisco-based opportunity. And originally, I thought I was gonna go work for a startup like a Series B or a Series C startup. But I was lucky enough to discover South Park Commons, um, which is a, a community of founders, basically, and a venture capital firm that backs them. The answer to when I wanted to build stuff by myself has been pretty much since I was a child. The question was, why didn't I? And you can be critical about a lot of things, you know, lack of agency and things like that. Um, but I think the reality, just to be, like, totally honest, is I had a lot of doubt in, in whether or not I could. And so actually, my experience at SPC and being able to have exposure to people who were great builders and having them teach me how to do that, how to run that process, how to build things on my own, was the first moment where I realized that I actually was capable of building things. On my own. I knew somewhere in my brain that I was capable of it, but I hadn't actually understood it or seen that path until I was able to meet the people at SVC that were really, really amazing at being able to make things come to life. I usually think about this as kinda two different forces. You have the driving force behind building, right? Curiosity, excitement, like wanting to do something for somebody else, wanting to solve problems, right? And then there is the sort of pulling back force of sort of doubt and, like, friction and sort of, like, difficulty of building the thing. Usually, what I like to try to think about is, like, how can you either increase that sort of driving force, which is hard to do, but it's possible. If you are, for example, really in love with somebody and you don't know how to cook, but they love home-cooked meals, you're gonna learn how to cook because you want to make that person happy, right? And that is something where maybe there w- that motivation to learn how to cook was never there before. So this is the classic advice of, like, find customers who are asking you for things. If you have a group of people that you're building for and those people are asking you for things and are excited about what you're doing, that's an incredibly motivating force, and that will basically pull ideas and pull things out of you as a, as a person because you won't be sitting around having time to think about, "Could I build this? Could I not?" You're just trying to solve the problem and getting there as quick as possible. On the other side is how do you reduce the friction to building? Some of this is mental and sort of getting the patterns and sort of getting the encouragement from the community and from others to say, "Hey, you can do it." And some of it is also tooling and things like that. And so I think now, of course, more than ever, uh, the friction is getting less and less to actually go out and build things, and that's a great place to be. Mike and Herick and I all met at South Park Commons together. They were the fastest builders I had ever seen. Every month, they had shipped, like, a new thing, and I was like, "Wow, I'm always curious to learn what you guys are working on." And eventually, I was around long enough that they just asked me to jump in at the next hackathon and kinda help them think of something that was a little bit more ambitious, and we did that, had a lot of fun, and decided to start a company together. So what I would say is focus on the motivating factors. That's a really important part of it and the environment that you're in. And then, of course, for me, and this is gonna be a theme, you know, throughout all of the decisions I've made in the past and the things that I've focused on is being around the right people is incredibly important. Being around people who are also builders and are also people who are putting things out into the world, who are also curiosity-driven and who are encouraging of that type of behavior is, is one of the most important things

4. 8:34 – 11:36

   Subscriptions Killed AI Apps, Ads will Save Them

   1. SPSpeaker8:34

      you can do. Well, I have to give most of the credit to my co-founder, Mike, who actually had the idea for the sort of AdSense for AI a year before we ended up actually building the product. And the reason why it wasn't built a year before is because the market was way too early to need this. Everyone in the industry was pushing towards subscription because that was the precedent that ChatGPT was setting. This was 2024 TC Disrupt, and some consumer VCs were basically preaching to the market that the way that AI was going to monetize primarily was gonna be subscription. The general sentiment amongst at least most consumer VCs was ads are dead, and there are articles you can read about this. Ads are dead. The whole ad model of the internet is going away. We're all gonna pay for things now because it's so useful and people know exactly what they want, right? My personality is one where I just love to battle. And so when I heard that, I was like, "Oh, fuck yeah," like, 'cause I don't mind taking the contrarian opinion, right? But now if you have 10,000 users with an AI app, especially if they're, if you're doing a good job and they're engaged and excited, uh, you're spending thousands of dollars a month, right? Because the inference cost tokens are very expensive. This is why so many people are pushing for subscriptions, 'cause subscriptions are able to cover that cost. But it's actually not a scalable solution. For example, a good friend of ours was building a, a tool basically for parents to tell stories to their kids so that their kids could kind of like visualize what was going on. So they were taking AI and allowing the parent to sort of describe different stories, and then the AI would sort of, like, generate the images or the videos around, like, what that story was. The engagement was incredible, and this person was getting fan mail being, "Wow, I, like, for the first time, I'm really connecting with my kids, and this is so fun, and they love it, and they're having so much," you know? But, like, he was losing, you know, tens of thousands of dollars a month. As, you know, a bootstrapped founder, it's not sustainable. And despite the fact that he had a subscription in place, it wasn't something that people were actually willing to or able to pay for. And, and ultimately, he ended up shutting down the business and getting a job, you know, at some B2B SaaS company, which made us really sad. We were built to solve the problem of how do we allow the developer of the AI experience, the generative experience, to create a sustainable revenue source so that they can grow as large as they want to without, you know, risking basically that losing money period of time where, like, inference costs are, you know, higher than their revenues are, right? At the same time, like, that's not 100% of my energy. There are still-- It did increase the number of moments of doubt, right? But I think that the main, the main thing there in terms of the doubt is that the doubt gets erased when you have customers that are asking for the thing. And so back to the motivating factor of what's pulling these ideas out of you, what's encouraging you to build. When you have customers that are telling you every day, like, "Please give this to us. We really wanna see somebody innovate here. Like, we're willing to pay you to develop this product," you realize that while somebody who is seeing the market as a whole and seeing the trends over here is probably seeing something that's accurate today, when you get to talk to the customers who are the ones who are driving what the products will look like in six to twelve months, you actually have a more accurate picture of where things will be in twelve months from now. And so that was, uh, the thing that allowed us to

5. 11:36 – 12:54

   Can Ads Actually Be a Good Experience?

   1. SPSpeaker11:36

      kind of erase most of the moments of doubt that came, came up from that. People think that ads by default suck and are terrible for the user experience and everything about them is bad. We actually ran a test last week where we basically showed fewer ads. We capped the frequency at which we were showing ads and we said, "What happens if we only show, like, you know, one-tenth the number of ads that we were showing before?" But we also measured what about the user base? And what we actually saw is when we capped the number of ads we were showing, the user engagement for organically on the app went down. People churned more often when we stopped showing ads. It's an indication that the, the ads that we're serving are very high quality and actually encourage the user to spend more time on the app that they're actually in, which could indicate that the, the ads themselves are a great experience, right? So and an example of this would be sailing in San Francisco Bay. San Francisco Bay, it's windy. There's, you know, wind shifts and things like that, but there's also a lot of current, and it's a very... If you know the local sort of current patterns, you might have information that no one else has that might make what you're doing look really stupid until it pays off. And then people go, "How did that person know that?" Right? And you knew it because you've seen it before. You've been- you have experience there. You've tried that thing. You, you're a local in the space. And so again, back to what we were s- hearing and seeing from customers, we knew something that the market did not at that time, and we were confident

6. 12:54 – 15:15

   The Bottleneck is Now Humans, Not Tech

   1. SPSpeaker12:54

      in that information. Agentic professional social network. We actually spent, uh, several months building that product before pivoting to Koah as it is today. The idea was that in the same way that actors or athletes have agents that go out into the world and sort of represent them and try to get them opportunities and things like that, why can't everyone have an agent for themselves to go out and try to proactively find the best opportunities for them, and said, "Hey, there are certain people that are within your network that you could talk to that would really unlock certain things that you're trying to figure out," right? And so we actually spent several months building that product before pivoting to Koah as it is today, realizing that while even we were getting some strong engagement, we had no idea how to monetize. Interestingly, at the time, what we learned was that people were still very cautious about sharing that type of information with an agent. That concept of sharing access to email or access to personal information or the ability to go out and, like, talk to people on behalf of a person, the industry wasn't ready for that type of agentic behavior yet. And so we ended up seeing a lot of skepticism, and therefore not enough usage for us to commit to that product long term. And so as a result, we, we, we pivoted away from it. Consumer behavior and sort of user adoption is lagging the technology quite a bit. The future is here. It's just not equally distributed. So for example, would you let an agent buy a car? Probably not. But would you let an agent buy a pair of socks? Maybe, 'cause, like, you know what? The consequences are pretty low, right? And so what does that mean? What it means is that the trust is only at a very low level of consequence, right? But we think that over time, both user behavior and the technology will increase to the point where we actually will start feeling like an agent could manage my investment portfolio. That's something we could imagine happening in the future because eventually we think agents might be able to prove themselves to be better at doing that than humans. Early adopters do it. We have to see them prove it out. We have to see them build case studies. We have to see the rest of the world kind of, you know, skeptically try it for the first time, realize it works, and then adopt it themselves. And while the technology is moving at absolutely incredible speeds, human brains still take a couple of years to kind of get used to a thing. I think all we can do and where we are as a business is position ourselves to be ready to react to the market when it's there. [outro music]

Episode duration: 15:16