Transcriptly

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

EO15:16Added Aug 31, 2026

Nic Baird was on track to become one of the best sailors in the world. He walked away to make a statistical bet on startups, and built Koah Labs, the ad mone...

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format

  1. Intro

  2. 00:00technology is moving at absolutely
  3. 00:02incredible speeds
  4. 00:04relative to 2024 to now. People were
  5. 00:07still very cautious about sharing
  6. 00:09information with an agent. That was the
  7. 00:11general sentiment in the market at the
  8. 00:12time. But nowadays, people are like,
  9. 00:14"Open claw sure has access to my entire
  10. 00:16computer. No problem." Right? We don't
  11. 00:18know how the market is going to play out
  12. 00:19exactly. We don't know what all the
  13. 00:21different players are going to do. We
  14. 00:22don't have all the information about
  15. 00:23whether or not a product is going to be
  16. 00:24successful. But you have to make
  17. 00:26decisions in a market that is uncertain.
  18. 00:28But what we can do is we can give
  19. 00:30probabilities to a thing. We can say
  20. 00:31it's very likely that the market will do
  21. 00:33this thing. Right? In sailing, you have
  22. 00:35these decisions all the time. If you are
  23. 00:38very confident that the wind is going to
  24. 00:40shift 5° to the left over the next 5
  25. 00:42minutes, right? You want to go towards
  26. 00:44that windshift. So you want to go
  27. 00:46towards the left side of the course. But
  28. 00:48let's imagine that the entire fleet,
  29. 00:50everyone else is going the other way.
  30. 00:52Are they being idiots? You kind of have
  31. 00:54to ask yourself, do they have
  32. 00:55information that I don't have? If I'm
  33. 00:57wrong, the risk is so high because so
  34. 00:59many people are going this direction.
  35. 01:01So, let me mitigate that risk. While I'm
  36. 01:03highly confident this thing is going to
  37. 01:05happen, I'm going to make a different
  38. 01:06decision than I would if I was just
  39. 01:08purely trusting my instinct because the
  40. 01:10math supports the stats supports that I
  41. 01:11go over here. And so, I think that is a
  42. 01:13great proxy for how really the best
  43. 01:15decisions are made for your own career,
  44. 01:18for your own life, or business is you
  45. 01:20have to look at the information that you
  46. 01:22have available, understand that you
  47. 01:23don't have all the information, and you
  48. 01:24have to make a statistical bet. You have
  49. 01:26to say, I think it's most likely this
  50. 01:27thing will happen. And then depending on
  51. 01:29how confident you are, that's how much
  52. 01:31you invest in that thing. I'm Nick
  53. 01:32Barrett. I'm the CEO of KA Labs. KA is
  54. 01:35basically the ad monetization platform
  55. 01:37for AI tools. We've raised $26 million
  56. 01:40to date. You know, we're reaching now
  57. 01:41singledigit millions of users um and
  58. 01:43growing very, very quickly.

  59. How a World-Class Sailor Decides Under Uncertainty

  60. 01:54A lot of people say that you shouldn't
  61. 01:56start a company. It's a bad economic
  62. 01:57decision. You know, all these things,
  63. 01:58but I like to think of things in terms
  64. 02:00of variability versus risk, right? And
  65. 02:02starting a company is a high variable
  66. 02:04thing, but it's I actually think it's a
  67. 02:06lowrisk scenario. Why? Because you learn
  68. 02:08so much being a founder of a company.
  69. 02:10You will become a more valuable asset to
  70. 02:13a company later and be able to make more
  71. 02:15money later, whether you fail or
  72. 02:16succeed. Sailing was by far the most
  73. 02:19important priority and where I was
  74. 02:20spending almost all of my time. Of
  75. 02:21course, sailing is many things. Like
  76. 02:23sailing is also cruising through the
  77. 02:24Caribbean and, you know, having a
  78. 02:25cocktail on the back of a nice big
  79. 02:27yacht. But the the type of sailing that
  80. 02:28I do is closer to NASCAR, Formula 1,
  81. 02:31fast, high performance racing. uh
  82. 02:34growing up with my dad doing that as a
  83. 02:35profession is by the time that I got to
  84. 02:37college, I I was able to go and compete
  85. 02:39at the youthful championships in two
  86. 02:41different uh classes in two different
  87. 02:42years and finished third and fourth
  88. 02:44respectively and had won three national
  89. 02:46championships by the time that I got to
  90. 02:48college.
  91. 02:54Chess is theoretically a solved game.
  92. 02:57You can play perfectly. In poker, there
  93. 03:01are too many confounding variables and
  94. 03:02you're operating under too much
  95. 03:03uncertain information. It means that you
  96. 03:06could play quote unquote perfectly in
  97. 03:08poker and still lose because there's
  98. 03:09uncertainty. There's a variable of
  99. 03:11randomness, there's a variable of
  100. 03:12chance. In sailing, you have these
  101. 03:14decisions all the time. When you are
  102. 03:17sailing in a, you know, a seabbze
  103. 03:19environment, we can call it. So, what
  104. 03:20happens is the land will heat up faster
  105. 03:23than water. And so, when that happens,
  106. 03:25you basically get a seab breeze. And so
  107. 03:27what we see that happen in San Francisco
  108. 03:28Bay area all the time. For example, the
  109. 03:30Sanwakin Valley is, you know, 100° and
  110. 03:33the Pacific Ocean temperature is 60°.
  111. 03:35And so that temperature differential is
  112. 03:37partly what creates such a powerful sea
  113. 03:40breeze and why it's so windy in San
  114. 03:41Francisco in the summertime. And so when
  115. 03:43you know those things, you can start to
  116. 03:44predict. We know that as the land heats
  117. 03:47up, typically when there reaches this
  118. 03:48temperature differential under these
  119. 03:50pressure circumstances, that's when we
  120. 03:52expect that seabbze to come in. And we
  121. 03:54usually expect that it comes in from
  122. 03:55this direction. Right? So you can
  123. 03:56position yourself well to be prepared
  124. 03:58for that to happen. That is a great
  125. 04:00proxy for how really the best decisions
  126. 04:02are made for your own career, for your
  127. 04:05own life, for a business. And you have
  128. 04:06to make a statistical bet. You have to
  129. 04:08say, I think it's most likely this thing
  130. 04:09will happen. And then depending on how
  131. 04:11confident you are, that's how much you
  132. 04:13invest in that thing. So you're saying,
  133. 04:15I'm 80% confident that it's going to
  134. 04:17rain today, right? It's not really
  135. 04:19expensive for you to carry around an
  136. 04:21umbrella. So you're going to bring an
  137. 04:22umbrella. I'm betting it's going to
  138. 04:24rain. And if it doesn't, the decision
  139. 04:25that I've made to bring an umbrella is
  140. 04:26not that harmful. Whereas, if it's a 10%
  141. 04:28chance of rain, be like, uh, yeah, I'm
  142. 04:29going to weigh off the trade-off of
  143. 04:31like, yeah, 10%, uh, maybe I get a
  144. 04:33little wet, but yeah, I don't really
  145. 04:34want to carry an umbrella around.
  146. 04:35Although it's inconvenient, right? So,
  147. 04:36that's like a very low stakes example of
  148. 04:38doing this. The high stakes examples of
  149. 04:40doing this is I was actually saying no
  150. 04:43to sailing opportunities and
  151. 04:44professional opportunities to go get
  152. 04:46paid to race because I was too curious
  153. 04:48about what was happening in startup
  154. 04:50land. The ambition in me was saying,
  155. 04:51"Great, I could become the best sailor
  156. 04:53in the world." Is that as important as
  157. 04:55becoming the hundth best business person
  158. 04:56in the world? No, it is not. If there
  159. 04:58was a 1% chance that you could make a
  160. 05:00trillion dollars, 1% of a trillion
  161. 05:02dollars is still $10 billion. Thumbs up,
  162. 05:05right? I'll go after that. Whereas, if
  163. 05:07you say, "I have a 100% chance of making
  164. 05:08a million dollars." People would say,
  165. 05:10"Well, go to the million dollars." But
  166. 05:11the expected value is actually much
  167. 05:13better if you go after the low chance
  168. 05:14but high reward. And so that's the way
  169. 05:16that I like to think about a lot of
  170. 05:17decisions in my life and also within the
  171. 05:20business and of course within sailing
  172. 05:21too which is where that article came
  173. 05:22from.

  174. To Become a Builder, Be Around Builders

  175. 05:27And after all that I thought I've got to
  176. 05:29get back to where the builders are, the
  177. 05:30people who are creating things and and
  178. 05:32find a very talent dense community. So I
  179. 05:35sought out in person San Francisco based
  180. 05:37opportunity and originally I thought I
  181. 05:38was going to go work for a startup like
  182. 05:40series B or a series C startup but was
  183. 05:42lucky enough to discover South Park
  184. 05:44Commons um which is a a community of
  185. 05:46founders basically in a venture capital
  186. 05:48firm that backs them. The answer to when
  187. 05:50I wanted to build stuff by myself has
  188. 05:52been pretty much since I was a child.
  189. 05:54The question was why didn't I? And you
  190. 05:57can be critical about a lot of things
  191. 05:58from lack of agency and things like
  192. 05:59that. Um, but I think the reality, just
  193. 06:01to be like totally honest, is I had a
  194. 06:03lot of doubt in in whether or not I
  195. 06:05could. And so actually my experience at
  196. 06:07SBC and being able to have exposure to
  197. 06:10people who were great builders and
  198. 06:11having them teach me how to do that, how
  199. 06:13to run that process, how to build things
  200. 06:15on my own was the first moment where I
  201. 06:17realized that I actually was capable of
  202. 06:19building things on my own. I knew
  203. 06:21somewhere in my brain that I was capable
  204. 06:23of it, but I hadn't actually understood
  205. 06:25it or seen that path until I was able to
  206. 06:27meet the people at SBC that were really,
  207. 06:29really amazing at being able to make
  208. 06:30things come to life. I usually think
  209. 06:32about this as kind of two different
  210. 06:33forces. You have the driving force
  211. 06:35behind building, right? Curiosity,
  212. 06:37excitement, like wanting to do something
  213. 06:39for somebody else, wanting to solve
  214. 06:40problems, right? And then there is the
  215. 06:42sort of pulling back force of sort of
  216. 06:45doubt and like friction and sort of like
  217. 06:47difficulty of building the thing.
  218. 06:48Usually what I like to try to think
  219. 06:50about is like how can you either
  220. 06:51increase that sort of driving force
  221. 06:52which is hard to do but it's possible if
  222. 06:54you are for example really in love with
  223. 06:57somebody and you don't know how to cook
  224. 06:58but they love home-cooked meals you're
  225. 07:00going to learn how to cook because you
  226. 07:02want to make that person happy right and
  227. 07:04that is something where maybe there that
  228. 07:06motivation to learn how to cook was
  229. 07:07never there before. So this is the
  230. 07:09classic advice of like find customers
  231. 07:10who are asking you for things. If you
  232. 07:12have a group of people that you're
  233. 07:14building for and those people are asking
  234. 07:15you for things and are excited about
  235. 07:16what you're doing, that's an incredibly
  236. 07:18motivating force and that will basically
  237. 07:21pull ideas and pull things out of you as
  238. 07:23a as a person because you won't be
  239. 07:25sitting around having time to think
  240. 07:27about could I build this, could I not?
  241. 07:28You're just trying to solve the problem
  242. 07:30and getting there as quick as possible.
  243. 07:31On the other side, it's how do you
  244. 07:33reduce the friction into building? Some
  245. 07:34of this is mental and sort of getting
  246. 07:36the patterns and sort of getting the
  247. 07:38encouragement from the community and
  248. 07:39from others to say, "Hey, you can do
  249. 07:41it." And some of it is also tooling and
  250. 07:43things like that. And so, I think now,
  251. 07:44of course, more than ever, uh, the
  252. 07:46friction is getting less and less to
  253. 07:48actually go out and build things. And
  254. 07:49that's a great place to be. Mike and
  255. 07:51Heric and I all met at South Park
  256. 07:53Commons together. They were the fastest
  257. 07:55builders I had ever seen. Every month
  258. 07:57they had shipped like a new thing. And I
  259. 07:59was like, "Wow, I'm always curious to
  260. 08:00learn what you guys are working on." And
  261. 08:02eventually I was around long enough that
  262. 08:03they just asked me to jump in at the
  263. 08:05next hackathon and kind of helped them
  264. 08:06think of something that was a little bit
  265. 08:08more ambitious. And we did that. Had a
  266. 08:09lot of fun and decided to start a
  267. 08:10company together. So what I would say is
  268. 08:12focus on the motivating factors. That's
  269. 08:15a really important part of it in the
  270. 08:16environment that you're in. And then of
  271. 08:17course for me, and this is going to be a
  272. 08:19theme, you know, throughout all of the
  273. 08:20decisions I've made in the past and the
  274. 08:21things that I've focused on is being
  275. 08:23around the right people is incredibly
  276. 08:24important. Being around people who are
  277. 08:26also builders and are also people who
  278. 08:28are putting things out into the world,
  279. 08:29who are also curiositydriven and who are
  280. 08:31encouraging of that type of behavior is
  281. 08:32is one of the most important things you

  282. Subscriptions Killed AI Apps, Ads will Save Them

  283. 08:34can do.
  284. 08:39I have to give most of the credit to my
  285. 08:40co-founder Mike who actually had the
  286. 08:42idea for the sort of AdSense for AI a
  287. 08:45year before we ended up actually
  288. 08:47building the product. And the reason why
  289. 08:49it wasn't built a year before is because
  290. 08:52the market was way too early to need
  291. 08:54this. Everyone in the industry was
  292. 08:55pushing towards subscription because
  293. 08:56that was the precedent that CHBT was
  294. 08:58setting. This was 2024
  295. 09:01TC Disrupt and some consumer VCs were
  296. 09:04basically preaching to the market that
  297. 09:06the way that AI was going to monetize
  298. 09:08primarily was going to be subscription.
  299. 09:10The general sentiment amongst at least
  300. 09:12most consumer VCs was ads are dead. And
  301. 09:15there are articles you can read about
  302. 09:16ads are dead. the whole ad model of the
  303. 09:18internet is going away. We're all going
  304. 09:19to pay for things now because it's so
  305. 09:20useful and people know exactly what they
  306. 09:22want, right? My personality is one where
  307. 09:24I just love to battle. And so when I
  308. 09:26heard that, I was like, "Oh, [ __ ] yeah."
  309. 09:29Like, cuz I don't mind taking the
  310. 09:30contrarian opinion, right? But now, if
  311. 09:32you have 10,000 users with an AI app,
  312. 09:33especially if they're if you're doing a
  313. 09:35good job and they're engaged and
  314. 09:36excited, uh you're spending thousands of
  315. 09:38dollars a month, right? Because the
  316. 09:39inference cost tokens are very
  317. 09:41expensive. This is why so many people
  318. 09:43are pushing for subscriptions cuz
  319. 09:44subscriptions are able to cover that
  320. 09:46cost. But it's actually not a scalable
  321. 09:48solution. For example, a good friend of
  322. 09:49ours was building a a tool basically for
  323. 09:52parents to tell stories to their kids so
  324. 09:55that their kids could kind of like
  325. 09:56visualize what was going on. So they
  326. 09:58were taking AI and allowing the parent
  327. 10:00to sort of describe different stories
  328. 10:01and then the AI would sort of like
  329. 10:03generate the images or the videos around
  330. 10:05like what that story was. the engagement
  331. 10:07was incredible and this person was
  332. 10:08getting fan mail being, wow, I like for
  333. 10:11the first time I'm really connecting
  334. 10:12with my kids and this is so fun and they
  335. 10:14love it and they're having so much, you
  336. 10:15know, but like he was losing, you know,
  337. 10:17tens of thousands of dollars a month. As
  338. 10:20you know, a bootstrapped founder, it's
  339. 10:22not sustainable. And despite the fact
  340. 10:24that he had a subscription in place, it
  341. 10:27wasn't something that people were
  342. 10:28actually willing to or able to pay for.
  343. 10:31and and ultimately he ended up shutting
  344. 10:33down the business and getting a job you
  345. 10:35know at some B2B SAS company which made
  346. 10:37us really sad. We were built to solve
  347. 10:39the problem of how do we allow the
  348. 10:41developer of the AI experience, the
  349. 10:43generative experience to create a
  350. 10:44sustainable revenue source so that they
  351. 10:46can grow as large as they want to
  352. 10:47without, you know, risking basically
  353. 10:49that losing money period of time where
  354. 10:51like inference costs are, you know,
  355. 10:52higher than their revenues are, right?
  356. 10:54At the same time, like that's not 100%
  357. 10:56of my energy. There are still that did
  358. 10:58increase the number of moments of doubt,
  359. 11:00right? But I think that the main the
  360. 11:02main thing there in terms of the doubt
  361. 11:03is that the doubt gets erased when you
  362. 11:05have customers that are asking for the
  363. 11:06thing. And so back to the motivating
  364. 11:07factor of what's pulling these ideas out
  365. 11:09of you, what's encouraging you to build.
  366. 11:11When you have customers that are telling
  367. 11:12you every day like please give this to
  368. 11:14us. We really want to see somebody
  369. 11:15innovate here. Like we're willing to pay
  370. 11:16you to develop this product. You realize
  371. 11:18that while somebody who is seeing the
  372. 11:20market as a whole and seeing the trends
  373. 11:22over here is probably seeing something
  374. 11:24that's accurate today, when you get to
  375. 11:26talk to the customers who are the ones
  376. 11:27who are driving what the products will
  377. 11:28look like in 6 to 12 months, you
  378. 11:30actually have a more accurate picture of
  379. 11:31where things will be in 12 months from
  380. 11:33now. And so that was the thing that
  381. 11:35allowed us to kind of erase most of the

  382. Can Ads Actually Be a Good Experience?

  383. 11:36moments of doubt that came came up from
  384. 11:38them. People think that ads by default
  385. 11:41suck and are terrible for the user
  386. 11:43experience and everything about them is
  387. 11:44bad. We actually ran a test last week
  388. 11:47where we basically showed fewer ads. We
  389. 11:50capped the frequency at which we were
  390. 11:52showing ads and we said what happens if
  391. 11:53we only show like you know onetenth the
  392. 11:55number of ads that we were showing
  393. 11:56before. But we also measured what about
  394. 11:57the user base? And what we actually saw
  395. 11:59is when we capped the number of ads we
  396. 12:01were showing, the user engagement for
  397. 12:03organically on the app went down. People
  398. 12:05churn more often when we stopped showing
  399. 12:07ads. It's an indication that the the ads
  400. 12:09that we're serving are very high quality
  401. 12:11and actually encourage the user to spend
  402. 12:14more time on the app that they're
  403. 12:15actually in, which could indicate that
  404. 12:16the the ads themselves are a great
  405. 12:18experience. Right? So, and an example of
  406. 12:20this would be sailing in San Francisco
  407. 12:21Bay. San Francisco Bay, it's windy.
  408. 12:24There's, you know, wind shifts and
  409. 12:25things like that, but there's also a lot
  410. 12:26of current and it's a very, if you know
  411. 12:28the local sort of current patterns, you
  412. 12:32might have information that no one else
  413. 12:34has that might make what you're doing
  414. 12:36look really stupid until it pays off and
  415. 12:39then people go, "How did that person
  416. 12:40know that?" Right? And you knew it
  417. 12:41because you've seen it before. You've
  418. 12:43been you have experience there. You've
  419. 12:44tried that thing. You you were local in
  420. 12:46the space. And so again, back to what we
  421. 12:48were hearing and seeing from customers,
  422. 12:50we knew something that the market did
  423. 12:51not at that time and we were confident
  424. 12:53in that information,

  425. The Bottleneck is Now Humans, Not Tech

  426. 12:59a gentic professional social network. We
  427. 13:02actually spent uh several months
  428. 13:04building that product before pivoting to
  429. 13:07KOA as it is today. The idea was that in
  430. 13:10the same way that actors or athletes
  431. 13:12have agents that go out into the world
  432. 13:13and sort of represent them and try to
  433. 13:15get them opportunities and things like
  434. 13:16that, why can't everyone have an agent
  435. 13:19for themselves to go out and try to
  436. 13:20proactively find the best opportunities
  437. 13:22for them and said, "Hey, there are
  438. 13:23certain people that are within your
  439. 13:24network that you could talk to that
  440. 13:25would really unlock certain things that
  441. 13:26you're trying to figure out, right?" And
  442. 13:28so we actually spent several months
  443. 13:30building that product before pivoting to
  444. 13:33KA as it is today. Realizing that while
  445. 13:36even we were getting some strong
  446. 13:39engagement, we had no idea how to
  447. 13:40monetize. Interestingly, at the time,
  448. 13:43what we learned was that people were
  449. 13:44still very cautious about sharing that
  450. 13:46type of information with an agent. that
  451. 13:49concept of sharing access to email or
  452. 13:51access to personal information or the
  453. 13:53ability to go out and like talk to
  454. 13:54people on behalf of a person. The
  455. 13:56industry wasn't ready for that type of
  456. 13:58agentic behavior yet. And so we ended up
  457. 14:00seeing a lot of skepticism and therefore
  458. 14:02not enough usage for us to commit to
  459. 14:04that product long term. And so as a
  460. 14:06result, we we we pivoted away from it.
  461. 14:08Consumer behavior and sort of user
  462. 14:10adoption is lagging the technology quite
  463. 14:11a bit. The future is here. It's just not
  464. 14:13equally distributed. So for example,
  465. 14:15would you let an agent buy a car?
  466. 14:17Probably not. But would you let an agent
  467. 14:19buy a pair of socks? Maybe because like
  468. 14:22you know what the consequences are
  469. 14:23pretty low, right? And so what does that
  470. 14:24mean? What it means is that the trust is
  471. 14:27only at a very low level of consequence,
  472. 14:29right? But we think that over time both
  473. 14:31user behavior and the technology will
  474. 14:32increase to the point where we actually
  475. 14:34will start feeling like an agent could
  476. 14:36manage my investment portfolio. That's
  477. 14:38something we could imagine happening in
  478. 14:39the future because eventually we think
  479. 14:40agents might be able to prove themselves
  480. 14:42to be better at doing that than humans.
  481. 14:44Early adopters do it. We have to see
  482. 14:45them prove it out. We have to see them
  483. 14:46build case studies. is you have to see
  484. 14:47the rest of the world kind of, you know,
  485. 14:49skeptically try it for the first time,
  486. 14:50realize it works, and then adopt it
  487. 14:51themselves. And while the technology is
  488. 14:54moving at absolutely incredible speeds,
  489. 14:56human brains still take a couple of
  490. 14:58years to kind of get used to a thing. I
  491. 14:59think all we can do and where we are as
  492. 15:01a business is position ourselves to be
  493. 15:04ready to react to the market when it's
  494. 15:06there.
He Raised $26M Betting on Probabilities, Not Guesses | Koah Labs, Nic Baird — Transcriptly