Transcriptly

"50 AI Agents Running My Company" Is a Lie. Here's How I Build It | Gumloop, Max Brodeur-Urbas

EO16:07Added Aug 31, 2026

"50 AI agents running my company" — that's not automation. That's a slop machine. Max Brodeur-Urbas, Co-founder and CEO of Gumloop, just raised a $50M Series...

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format

  1. Intro

  2. 00:00There's a category of like entrepreneurs
  3. 00:02might be a negative term, but someone
  4. 00:03who thinks that they can just build a
  5. 00:05business by with a single click, you
  6. 00:06know, like it's effortless. Because I
  7. 00:08bought this course, I now have the the
  8. 00:09recipe to making a million dollars a
  9. 00:11year as a side hustle. There's kind of
  10. 00:13an anti pattern with AI right now where
  11. 00:15I have 50 AI agents running my company
  12. 00:17and I have a seuite of AI that tells me
  13. 00:19what to do every day. I [music] think
  14. 00:21that's the wrong approach. That's just
  15. 00:22like you're making a slot machine. Slop,
  16. 00:24not slot. Everyone on Twitter is like,
  17. 00:26"I automated [music] everything and I I
  18. 00:28don't do I work one hour a week and I
  19. 00:29make $10 million on the weekend with my
  20. 00:31[music] SAS app." Like, most of that is
  21. 00:33just marketing. They're lying to you.
  22. 00:34Uh, for the most part, you can sell hope
  23. 00:36really easily. But you're selling this
  24. 00:38vision of like skipping the hard work,
  25. 00:39shortcutting directly to the value,
  26. 00:41which will never happen. It'll never
  27. 00:42work. But for the person who's selling
  28. 00:43you that course, they're [music] going
  29. 00:44to make a ton of money. They they found
  30. 00:45the way to print money. My name is Max.
  31. 00:49I'm the founder and CEO of Gum Loop. Gum
  32. 00:51Loop is an automation platform [music]
  33. 00:52that is automating about 4 million
  34. 00:54workflows every day across huge
  35. 00:56businesses like Instacart, Shopify, Door
  36. 00:58Dash, Gusto.
  37. 01:01We're a team of 15 [music] at this
  38. 01:03point, scaling pretty quickly. What it
  39. 01:04does is it lets people who really
  40. 01:05understand a problem within a business
  41. 01:07like the marketer or the salesperson or
  42. 01:09the ops [music] person automate their
  43. 01:10own work instead of having to spec it
  44. 01:12out and explain it to an engineering
  45. 01:13team to build that [music] solution for
  46. 01:15them.

  47. Rule 1.Throw Yourself Into the Ether

  48. 01:26I went to McGill in Montreal. I studied
  49. 01:29software engineering. Really liked doing
  50. 01:31well in school. So, I put a lot of
  51. 01:32effort into that. Um, studied all the
  52. 01:34time. Tried to get the best grades I
  53. 01:36possibly could. I remember [music] my
  54. 01:37goal all throughout college was just get
  55. 01:38a good job. I wanted to work in big tech
  56. 01:41cuz I thought that would be the the
  57. 01:42right path to follow. Realized pretty
  58. 01:44quickly that I hated it. What most
  59. 01:46people say when when [music] you're
  60. 01:48thinking about starting a company is
  61. 01:49like I'll just go to big tech for a
  62. 01:50little bit. I'll learn a bunch and then
  63. 01:52I'll do my own thing. I think that's uh
  64. 01:54kind of cope in a way like a lot of
  65. 01:56people say that they go and do it and
  66. 01:58then they they [music] end up kind of
  67. 01:59getting golden handcuffs and never
  68. 02:00leaving and they never actually go and
  69. 02:02start their own thing. I don't think
  70. 02:03I've used anything I learned in big tech
  71. 02:05in my startup at all. I think the only
  72. 02:07thing working in big tech actually
  73. 02:08provided is some sort of default respect
  74. 02:11I guess from uh people when they they
  75. 02:13realize you're not just a random person
  76. 02:14like you actually had a company validate
  77. 02:16that you're you're competent in some
  78. 02:17way. So getting that logo on your resume
  79. 02:19is is helpful a little bit. But I don't
  80. 02:21think I learned anything novel when I
  81. 02:23was there. A lot of the things that I do
  82. 02:24now are motivated by the things I didn't
  83. 02:26like at Microsoft. Most of the things I
  84. 02:28do are the exact opposite of how they
  85. 02:30work in big tech. what you can do when
  86. 02:32you're 21, 22, 23 with no
  87. 02:34responsibilities, no obligations to
  88. 02:36anyone. Those are kind of years you can
  89. 02:37never get back. And if you're spending
  90. 02:39it just going to your 9 to5, logging in,
  91. 02:41fixing a ticket, and then logging off,
  92. 02:43you're kind of wasting those really,
  93. 02:44really important years. If we're going
  94. 02:46to do it, why don't we just do it now?
  95. 02:47Figure out what the more reliable,
  96. 02:48responsible thing to do later. I think
  97. 02:50that was the right choice, just like
  98. 02:52throwing ourselves into the ether and
  99. 02:54seeing what worked.
  100. 02:56I actually got deported from the United
  101. 02:58States and banned for 5 years. But it
  102. 03:00was nothing illegal or anything. I quit
  103. 03:02Microsoft and then I moved back to
  104. 03:04Vancouver. And my thought was like, I'm
  105. 03:05just going to live in Vancouver for a
  106. 03:07year, build stuff in my bedroom. And
  107. 03:09then one weekend I was going to visit my
  108. 03:10old roommates in Seattle and they were
  109. 03:12like, the border people are like, where
  110. 03:13are you going? Where are you staying?
  111. 03:14What do you do? So they turned me around
  112. 03:16at the border after a bunch of questions
  113. 03:18cuz I I was suspicious. Basically, the
  114. 03:20suspicion was that I was going to stay
  115. 03:21for longer than I was saying, even
  116. 03:23though I was just going to stay for 2
  117. 03:24days. But that came with a 5-year ban
  118. 03:26from the country. And then that was kind
  119. 03:27of the moment where I realized I had to
  120. 03:29build a company because I I had no
  121. 03:32fallback plan. I decided to just take it
  122. 03:33more seriously, build things that people
  123. 03:35actually want, something that could make
  124. 03:36money. I was pretty terrified, honestly.
  125. 03:39I remember driving back from the border
  126. 03:42to my girlfriend's apartment, just like
  127. 03:45almost in shock. [music] It took me a
  128. 03:46couple days to settle and calm down a
  129. 03:49little bit, but after that, it was just
  130. 03:50focus. I kind of just worked for the
  131. 03:52next 6 months as hard as I could.

  132. Rule 2.Prove Yourself Wrong

  133. 03:57I tried everything like I was trying to
  134. 04:00build anything that seemed remotely
  135. 04:02valuable. Built a video game moderation
  136. 04:05software in VR and then I did general
  137. 04:08trust and safety tooling. I did bot
  138. 04:09detection software um for web traffic, a
  139. 04:12platform like an anti-cam platform. I
  140. 04:14built a ton of stuff and then would make
  141. 04:16an MVP and then try to sell it and see
  142. 04:18if there was interest from the market.
  143. 04:19Over and over I tried that kind of like
  144. 04:20once a week. I had a different idea,
  145. 04:22something I was experimenting with, and
  146. 04:24then I'd learn pretty quickly that it
  147. 04:26was a bad idea. The more I did it, the
  148. 04:28more I got used to proving myself wrong,
  149. 04:29and I I kind of learned the
  150. 04:30counterintuitive fact that in startups,
  151. 04:32you're actually chasing proving yourself
  152. 04:34wrong. That's the best thing that can
  153. 04:35possibly happen cuz you're saving weeks
  154. 04:37or months of time. In the beginning, I
  155. 04:39was building ideas for months and then I
  156. 04:41would hope that someone would prove me
  157. 04:42right and say this was worth building,
  158. 04:44but that's the opposite of what you
  159. 04:45should be doing. I wasted maybe 3 months
  160. 04:47doing that. You should actually be
  161. 04:49hunting for someone to tell you why this
  162. 04:50won't work. If you can't find a reason
  163. 04:52it won't work, then you actually have
  164. 04:53some sort of tangible idea you should
  165. 04:55pursue. If I could do it all over again,
  166. 04:56I would be hunting for the strong reason
  167. 04:59why something won't work instead of
  168. 05:01hoping for the reason that might build
  169. 05:03as much as you can. Like that's what
  170. 05:05provides the most information. I don't
  171. 05:06think I would have ended up building Gum
  172. 05:08Loop if I didn't fail 10 times before
  173. 05:10that. Talking to users as much as you
  174. 05:11can, it's almost a a privilege you have
  175. 05:13to earn because in the beginning you
  176. 05:14won't have any users to talk to. you'll
  177. 05:16be kind of begging people to try your
  178. 05:17product, which is not a fun position to
  179. 05:18be in, but something you have to kind of
  180. 05:20go through. I think putting them at the
  181. 05:21forefront of everything you're building,
  182. 05:23even if uh they're saying things you
  183. 05:24don't want to hear, your product sucks
  184. 05:25or you're not building the right thing
  185. 05:27or it's not solving my problem, that's
  186. 05:28actually the most val valuable feedback
  187. 05:30you can get. When I started working on
  188. 05:32the current version of Gum Loop, that
  189. 05:34was like one of the first days where I
  190. 05:35stayed at the office till midnight
  191. 05:36because I was so excited about what I
  192. 05:38was building. Like I I couldn't get
  193. 05:39enough of it. And that just kind of
  194. 05:40compounds over time. Like if you find
  195. 05:42that thing that will get you to work
  196. 05:43until midnight and not want to close
  197. 05:45your laptop, every day becomes easier.
  198. 05:47You kind of just uh build this crazy
  199. 05:49momentum that you can't get enough of.
  200. 05:51Do you remember AutoGPT? There was this
  201. 05:53super popular open source agent
  202. 05:55framework. It kind of took Twitter and
  203. 05:57the world by storm because it was the
  204. 05:58first time that it felt like AI could do
  205. 06:00something agentically and solve a
  206. 06:02problem on its own. I saw it on Twitter,
  207. 06:04tried it out. It was pretty amazing. I
  208. 06:06didn't actually push the limits and see
  209. 06:08how it would break. But in in the first
  210. 06:10demo, it seemed really cool. So I joined
  211. 06:12the Discord server which was growing
  212. 06:14pretty exponentially at the time and I
  213. 06:15would see all these people asking like
  214. 06:17how what is GitHub? How do I use a
  215. 06:19terminal? How do I install something
  216. 06:20locally? What is a dependency? I
  217. 06:22realized if I just solved that problem
  218. 06:24for them and built this nice little UI,
  219. 06:25then I mean I'm doing something
  220. 06:27interesting. And my thought was I would
  221. 06:28learn how to build a front end so it
  222. 06:30would be a good use of time. I didn't
  223. 06:31think it would become anything special.
  224. 06:33But whenever anyone would ask in the
  225. 06:34Discord server for help setting up their
  226. 06:36local environment, I would send them a
  227. 06:38link to AgentHub, which was the first
  228. 06:39version of it. After building it for a
  229. 06:41couple days, I was like, "Oh, this could
  230. 06:42be like GitHub for agents. If agents are
  231. 06:44useful, I could have the platform that
  232. 06:45you host them on that you can interact
  233. 06:47with them on." That idea kind of
  234. 06:48crumbled after a few days when I
  235. 06:50realized the agents weren't useful,
  236. 06:51which was the the aha moment. I had a a
  237. 06:53platform that people wanted to use, but
  238. 06:55they were frustrated with because the
  239. 06:56agents were so unreliable. So I kind of
  240. 06:58gave them what they were secretly asking
  241. 06:59for which is just reliability,
  242. 07:01predictability. All of their use cases
  243. 07:03were so simple that I thought I could
  244. 07:04just build a framework that lets them
  245. 07:05kind of automate the steps one after
  246. 07:07another. And then that kind of just
  247. 07:09naturally grew into this crazy
  248. 07:10automation platform. It was for
  249. 07:12developers because it was an open source
  250. 07:14project. But the audience that really
  251. 07:16went crazy for it were non-technical.
  252. 07:18they were like business admins at a
  253. 07:20company or ops people or HR people who
  254. 07:22were just enthused about the idea of AI
  255. 07:24actually solving a problem for them and
  256. 07:26and automating their work. So realizing
  257. 07:28that the actual audience was 80%
  258. 07:29non-technical that was when I realized
  259. 07:31like I need to build something that is
  260. 07:33approachable for them that feels fun for
  261. 07:34them to use that doesn't frustrate them
  262. 07:36with all the complexity that was kind of
  263. 07:37the motivation behind the the simplified
  264. 07:39version of autograph. We got into IC 5
  265. 07:41months before the batch started. For
  266. 07:43that 5 months the product was still
  267. 07:44completely free. We knew that we were
  268. 07:45going to turn on pricing at some point.
  269. 07:47During the batch, I think in the first
  270. 07:49week, we turned on pricing. We made the
  271. 07:50product $20 a month because we couldn't
  272. 07:53imagine charging more than Chat GBT. But
  273. 07:54our first paying customer was just this
  274. 07:56guy named Kai who he paid $20. We
  275. 07:59freaked out. It was like the greatest
  276. 08:01moment ever. Uh seeing the Stripe
  277. 08:02notification pop up. And then he's still
  278. 08:05a user.

  279. Rule 3.Real Networks Aren't Built at Cocktail Parties

  280. 08:09I was stuck in Canada during all of YC.
  281. 08:11So, I had all of the pressure of needing
  282. 08:13to build something amazing and live up
  283. 08:15to just the expectations that we had for
  284. 08:17being in YC, but I didn't have any of
  285. 08:19the distractions. So, I was just in a
  286. 08:20small studio apartment in Vancouver just
  287. 08:23in my room coding as fast as I could. I
  288. 08:25think we learned that staying focused
  289. 08:26and and not getting distracted by like
  290. 08:28the networking events and the the random
  291. 08:30parties that people in tech host is
  292. 08:33really powerful. Like the people who are
  293. 08:34actually building something amazing are
  294. 08:36not at those events. Like no, no one is
  295. 08:38really networking if you're bu if you're
  296. 08:39on to something. So, I think I've kept
  297. 08:41that attitude going forward. Like, I
  298. 08:42don't normally go to events for the most
  299. 08:44part. My co-founder never goes to events
  300. 08:46almost to a fault. Most people haven't
  301. 08:47met him because he's just working all
  302. 08:49the time. That's the biggest takeaway
  303. 08:51that if you stay focused and you just
  304. 08:54talk to users [music] as much as you can
  305. 08:55like your network will emerge pretty
  306. 08:57naturally. Fundraising like a lot of
  307. 08:58people think you have to go out and
  308. 08:59network in order to [music] find meet
  309. 09:01these investors and convince them about
  310. 09:02your idea, but if you build something
  311. 09:04exceptional, they will actually come to
  312. 09:06you. It's not actually not that
  313. 09:07complicated. You just have to show them
  314. 09:08that you'll succeed without them and
  315. 09:10then you'll be the one getting the email
  316. 09:11telling them to wait. That was probably
  317. 09:13the biggest realization during YC that
  318. 09:15it's not that complicated. Just build
  319. 09:17something great. Yeah. It's not a
  320. 09:19network is not made at the cocktail
  321. 09:20party.

  322. Rule 4.Great Products Aren't Built in One Click

  323. 09:24There's kind of an anti pattern with AI
  324. 09:26right now where you can go too far in
  325. 09:28one direction and become the type of
  326. 09:29person who's like I have 50 AI agents
  327. 09:32running my company and I have a seuite
  328. 09:33of AI that tells me what to do every
  329. 09:35day. I think that's the wrong approach.
  330. 09:37That's just like you're making a slot
  331. 09:38machine basically. Slop, not slot. I
  332. 09:41think it's important to know what you
  333. 09:43should be using AI for and what you
  334. 09:44shouldn't. Like keeping the human touch
  335. 09:45in the important parts and automating
  336. 09:47the repetitive things is the way to go.
  337. 09:49I always see the best users as being
  338. 09:50super AI enabled, not replacing their
  339. 09:52entire job with AI. But it's it's a
  340. 09:54slippery slope. Like it's everyone on
  341. 09:56Twitter is like, I automated everything
  342. 09:57and I I don't do I work one hour a week
  343. 09:59and I make $10 million on the weekend
  344. 10:01with my SAS app. Like most of that is
  345. 10:03just marketing. they're lying to you.
  346. 10:05Uh, for the most part, like I think
  347. 10:06there's a lot of uh course bros, I like
  348. 10:09to call them, uh, like people who sell
  349. 10:10the dream and all these people on
  350. 10:12Twitter are kind of inspired by the fact
  351. 10:14that I could make 30,000 this weekend if
  352. 10:16I just copy this workflow and and
  353. 10:18comment and then they'll give me the the
  354. 10:20recipe. I guarantee that's never worked.
  355. 10:21They're they're shilling the vision of
  356. 10:23productivity without actually offering
  357. 10:25anything [music] novel. I think that the
  358. 10:27value comes from when you apply AI to
  359. 10:29something you understand deeply. If
  360. 10:30there was some magic solution that would
  361. 10:32make you $30,000 in a weekend, they
  362. 10:34wouldn't be giving it to you on Twitter.
  363. 10:35The most productive people who are
  364. 10:36actually generating the most value, it
  365. 10:38normally comes from taking something you
  366. 10:39really understand, applying AI to it,
  367. 10:41wherever you see fit, and then [music]
  368. 10:42just scaling from there. I think it's
  369. 10:44cuz I I see the same patterns emerging.
  370. 10:47Like whenever there's a hype bubble,
  371. 10:48like with crypto or with NFTTS or with
  372. 10:50AI, there's a section of [music] the
  373. 10:52community that's a little vulnerable. I
  374. 10:54think easily persuaded and convinced
  375. 10:56that something will save them from
  376. 10:58whatever [music] situation they're in.
  377. 10:59you can sell hope really easily. So I
  378. 11:01think there's a lot of people online who
  379. 11:02take advantage of that and make content
  380. 11:04that is just fictitious. There's a
  381. 11:07category of like entrepreneurs like
  382. 11:09entrepreneurs but a want entrepreneur
  383. 11:10might be a negative term but someone who
  384. 11:12thinks that they can just build a
  385. 11:13business by like with a single click you
  386. 11:15know like [music] it's effortless
  387. 11:16because I bought this course I now have
  388. 11:18the the recipe to making a million
  389. 11:20dollars a year on as a side hustle but
  390. 11:22you're selling this vision of like
  391. 11:23skipping the hard work cut shortcutting
  392. 11:25directly to the value which will never
  393. 11:27happen. it'll never work. Uh, but for
  394. 11:29the person who's selling you that
  395. 11:30course, they're going to make a ton of
  396. 11:31money. They they found the way to print
  397. 11:33money. Basically,
  398. 11:35I only automate the things I really
  399. 11:37understand. Like, if you're automating
  400. 11:39something you don't understand, it's
  401. 11:40just going to be a slot machine. If
  402. 11:42you're using AI to do something you
  403. 11:44don't actually understand at all, often
  404. 11:45times you're creating just uncertainty.
  405. 11:47Like, if you're using AI to code and you
  406. 11:49don't know how to code at all, you're
  407. 11:50making malware at the end of the day.
  408. 11:52Like, it'll it'll come back to bite you.
  409. 11:54And vibe coding can only go so far. Same
  410. 11:56thing with automating workflows. Like if
  411. 11:57you're trying to automate something you
  412. 11:59could never do yourself or you don't
  413. 12:00understand, you're going to create
  414. 12:01something that's pretty poor. So I think
  415. 12:04I I apply AI to speed myself up and take
  416. 12:06the things I do understand, do it way
  417. 12:08faster so I can learn more things and
  418. 12:10like grow as a person, but I'm never
  419. 12:12[music] like trying to shortcut
  420. 12:13understanding something or like
  421. 12:15expanding my skill set by like having AI
  422. 12:17just replace me and and do things for
  423. 12:19[music] me. It's possible that the last
  424. 12:21generation of great engineers has been
  425. 12:23born because there was this era of
  426. 12:25actually needing to understand what's
  427. 12:26going on and then getting accelerated by
  428. 12:28AI. But now people can skip the
  429. 12:31understanding part and just accelerate
  430. 12:32with AI. I think there's going to be a
  431. 12:34like a few a much smaller community of
  432. 12:36people who actually use AI as a learning
  433. 12:37tool and they understand the
  434. 12:38fundamentals, use it as like a a teacher
  435. 12:41to really grasp what's going on. But
  436. 12:43it's a slippery slope as well because
  437. 12:45it's so easy just to not want to
  438. 12:46understand why something works cuz it
  439. 12:48just like your website worked or the
  440. 12:49feature did what you wanted but you
  441. 12:51didn't take the time to like really dig
  442. 12:52into why why it worked or why it didn't
  443. 12:54or what knock-on effects this could
  444. 12:56have. So I [music] think there will be a
  445. 12:57a bigger split in the people who are
  446. 12:59truly exceptional versus the people who
  447. 13:01are not. Because if you can actually
  448. 13:03have the determination to like pause,
  449. 13:05try to understand the problem, have AI
  450. 13:07teach you the things you don't
  451. 13:08understand, you'll become exceptional
  452. 13:10even faster [music] than before and then
  453. 13:12the average person will just kind of
  454. 13:13fall to the slop.

  455. Rule 5.Hiring is Like Dating

  456. 13:17Almost everyone we've hired is through
  457. 13:18network. A lot of people we hired were
  458. 13:20actually customers. So like our customer
  459. 13:22from Instacart quit his job and joined
  460. 13:24us. Our customer from Web Flow quit his
  461. 13:25job and joined us. Someone from Shopify
  462. 13:27quit their job and joined us. They
  463. 13:28already had conviction. Like they love
  464. 13:29the platform so much that they decided
  465. 13:31to drop everything and contribute to the
  466. 13:33mission. It's a really like fun way to
  467. 13:34hire because the only thing you really
  468. 13:36have as a startup is the optimism in
  469. 13:38general. Like you need to be excited to
  470. 13:40come to work every day and and be the
  471. 13:41one proving everyone wrong for when they
  472. 13:43say like, "Oh, why won't this company do
  473. 13:44this or why won't [music] this big
  474. 13:46company just steamroll you?" But these
  475. 13:48people are already bought in. Like they
  476. 13:49use the tool every day. So they love
  477. 13:50what you're doing already and they're
  478. 13:52like they see the vision. So, it's a
  479. 13:54very quick transition. You just have to
  480. 13:55figure out the details of how you
  481. 13:56onboard them. It's like dating in a way,
  482. 13:58like you have to be the person that
  483. 14:00someone would want to date. [music] You
  484. 14:01can't just like beg someone to date you,
  485. 14:02just like you can't beg someone to join
  486. 14:04your company. So, you have to build
  487. 14:05something amazing and have traction
  488. 14:06[music] to make the best people on earth
  489. 14:07want to join you. But it all starts with
  490. 14:09just like doing it. You have to start
  491. 14:10somewhere. My co-founder only joined
  492. 14:12because I had a version of the product
  493. 14:13that was working in the early days.
  494. 14:15Like, I was demoing the first version of
  495. 14:17it and he got excited by that. So, he
  496. 14:18joined. You have to build something
  497. 14:19exceptional. We're having a lot of fun
  498. 14:21at Gum Loop. Like every every day
  499. 14:23there's no one like being told to stay
  500. 14:24late. There's no hour you have to show
  501. 14:26up or hour you have to leave. Everyone
  502. 14:27is just kind of all we're all equally
  503. 14:29excited about the mission [music] and
  504. 14:30love what we're doing. And a big filter
  505. 14:33for how I hire people is like, do I
  506. 14:34actually want to spend all my time with
  507. 14:36this person? Could I hang out with them
  508. 14:3724/7? That has just kind of like
  509. 14:39compounded over time. And now we have a
  510. 14:40big group of people that are all really
  511. 14:42well adjusted, really fun to work with,
  512. 14:44ambitious, excited, uh, intelligent. The
  513. 14:47more people you hire like that, the more
  514. 14:48exciting it gets for everyone. So it
  515. 14:50just kind of like the momentum builds
  516. 14:51[music] in this direction and and people
  517. 14:53love what they're doing. I think there's
  518. 14:55a million reasons to not build every
  519. 14:57startup. Like it's it's very easy to
  520. 14:59hear an idea and say like what's the
  521. 15:00moat? Why won't X company do this? Why
  522. 15:02won't Y company do that? I think the
  523. 15:04people who are like really asking
  524. 15:05themselves that and obsessing over those
  525. 15:06questions will never build anything. You
  526. 15:08end up just working for a big player and
  527. 15:09being a pawn in their game. But if you
  528. 15:11try things and try to prove just like
  529. 15:13take a risk and try to prove other
  530. 15:14people wrong, you'll end up somewhere
  531. 15:16that people are like how how did you get
  532. 15:18there? how are you leading this company
  533. 15:19that so many people use? And the answer
  534. 15:21was just like, you tried and it worked.
  535. 15:24And when it didn't work, you tried
  536. 15:25again. I could have on day one explained
  537. 15:27to myself a hundred reasons why Zapier
  538. 15:29would do this better than us or why
  539. 15:30OpenAI and all these big players would
  540. 15:32crush us and then no one would be using
  541. 15:34Gum Loop. We wouldn't have created
  542. 15:35anything new. I think the biggest
  543. 15:37quality [music] that makes a founder
  544. 15:39start founding something is thinking
  545. 15:40that they can do it. Like you will never
  546. 15:42start a company if you don't think
  547. 15:43you're going to be the person to do it.
  548. 15:44So it just takes this blind confidence.
  549. 15:46It's pretty exciting just how much
  550. 15:48someone can do when they think they can
  551. 15:49do it.
"50 AI Agents Running My Company" Is a Lie. Here's How I Build It | Gumloop, Max Brodeur-Urbas — Transcriptly