Transcriptly

What Young Founders Get Wrong About Startups | Articul8, Arun Subramaniyan

EO19:27Added Aug 31, 2026

Meet Arun Subramaniyan, Founder & CEO of Articul8. Before launching Articul8, he led global AI, quantum, and HPC initiatives at AWS, and later served as VP o...

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format

  1. Intro

  2. 00:00So the biggest thing that is different
  3. 00:01of course being in a startup is that you
  4. 00:03don't have any safety net. You really
  5. 00:05need to have a hustle that goes way
  6. 00:07beyond anything you've seen in life. You
  7. 00:09don't have any time to sit and think.
  8. 00:12You just have to deal with it and keep
  9. 00:13rolling. I would say the biggest thing
  10. 00:14I'll do differently is to make sure that
  11. 00:17people who come to the company
  12. 00:20understand the fact that there's no
  13. 00:21safety net. And we've had our fair share
  14. 00:24of turnovers. And even people who were
  15. 00:26in the early early teams, we've had to
  16. 00:28let go. Partly because they may be
  17. 00:30perfect candidates, perfect employees in
  18. 00:32a large company, perfect candidates or
  19. 00:34outstanding candidates in a medium
  20. 00:36company, but a terrible misfit in a
  21. 00:38small company. Like I'll tell you a
  22. 00:40general conversation that would usually
  23. 00:41go. We ask for say going and meeting
  24. 00:44with a customer or something. We're
  25. 00:45expecting a week goes by. You go ask
  26. 00:47them, "Hey, did this get done?" And
  27. 00:48they'll say, "No, I sent an email. We
  28. 00:50still waiting for a response." Now, in a
  29. 00:52large company, that's a perfectly fair
  30. 00:53statement to make. No problem. You'll
  31. 00:55get a response. And typically, say
  32. 00:56you're an Amazon or you're in Google or
  33. 00:58Microsoft, you send a response, people
  34. 00:59respond back. The same people, if you're
  35. 01:02sending an email from a startup, they
  36. 01:04don't respond. Most of the time, they
  37. 01:05don't have time to do that. But second,
  38. 01:07we don't have the time to wait. We send
  39. 01:09an email, then call the person, then
  40. 01:11text the person, and text their wife if
  41. 01:12we have to to make sure we get a
  42. 01:14response. There have been cases where we
  43. 01:16would say, "Look, do you want me to go
  44. 01:18pick up your kids so that I can actually
  45. 01:20get 15 minutes with you?" It's that kind
  46. 01:22of hustle that you rarely almost ever
  47. 01:25need in any other job, but you need it
  48. 01:27here. And for that, you also need
  49. 01:29humility because normally you would say,
  50. 01:31well, I am the CEO. I can't be going and
  51. 01:35doing something. It's exactly the
  52. 01:37opposite. In fact, you have to do
  53. 01:38everything. Meaning, I can work 18 hours
  54. 01:40a day. I can work 7 days a week. I would
  55. 01:43still not get to even 30% of my tasks.
  56. 01:46If you're an early stage startup, there
  57. 01:47is no such thing called work life
  58. 01:49balance. Don't kid yourself. That is
  59. 01:51work. That is life. It's the same thing.
  60. 01:53The real definition of entrepreneurship
  61. 01:55is aiming to do something for which you
  62. 01:58don't have the resources. That's the
  63. 02:00definition of entrepreneurship. I am
  64. 02:02Arun Subraman, founder and CEO of
  65. 02:04Articulate. Articulate is a domain
  66. 02:06specific Genai platform that enables
  67. 02:09enterprises to get value from their deep
  68. 02:11domain expertise and data they've
  69. 02:13collected over decades.

  70. Principle 1: Make a Dent in the Universe

  71. 02:25I grew up in a family where technology
  72. 02:27was a given. Normal quintessentially
  73. 02:29Indian uh upbringing in the 80s and '90s
  74. 02:32assumed that you have to do well in
  75. 02:34school. That was the number one
  76. 02:35priority. I was fascinated by anything
  77. 02:37that flew even before I got to college.
  78. 02:39And very early on I wanted to be an
  79. 02:40aerospace engineer. Um there are very
  80. 02:42few places you could go to study
  81. 02:43aerospace engineering at the time and I
  82. 02:45was lucky enough to get into one of the
  83. 02:47top schools. I went to Anna University
  84. 02:49and there's a school called Matas
  85. 02:50Institute of Technology. My college
  86. 02:52education was just fun for me all the
  87. 02:54way through. So I finished undergrad
  88. 02:56there went off to do my graduate studies
  89. 02:58at Purdue. Many people may not know this
  90. 03:00is called uh the birthplace or cradle of
  91. 03:02astronaut is a school where most
  92. 03:04astronauts in the US actually come from.
  93. 03:06Neil Armstrong actually graduated from
  94. 03:08Purdue. There is a particular incident
  95. 03:10that deeply touched me. I had never
  96. 03:11thought about it that way. This was when
  97. 03:13I was in college. I was attending a
  98. 03:15seminar by somebody and he said
  99. 03:17something that I had never thought
  100. 03:19about. You need to feel ownership and
  101. 03:22responsibility to actually help. And
  102. 03:25what that means is he gave an example
  103. 03:27which is think of the people who first
  104. 03:30designed dams and built dams. Whoever
  105. 03:32they might be, they stopped the flow of
  106. 03:34rivers. At the time that they started
  107. 03:36building dams, it was all an act of God.
  108. 03:39Wherever whichever culture you were,
  109. 03:41they really wanted to make sure that
  110. 03:42floods didn't happen. They didn't people
  111. 03:44didn't die. They didn't suffer. They
  112. 03:46didn't sit and think, "Oh, this is an
  113. 03:48act of God. God wanted floods to come
  114. 03:50in. We can't do anything." They actually
  115. 03:52thought that I will help people not
  116. 03:54suffer and actually go build dams. When
  117. 03:57they first built it, it's an impossible
  118. 03:59act to be doing. Now, it's a natural
  119. 04:01thing. Sure, engineering marvels. We
  120. 04:02build dams all over the place. But it's
  121. 04:04not for the first person to have thought
  122. 04:06about it. Think of the responsibility
  123. 04:08the person took to stop an act of God,
  124. 04:11right? And that's a massive
  125. 04:13responsibility. And not only they they
  126. 04:15thought about it that way, they went and
  127. 04:16built the dam. And if you actually think
  128. 04:18about it from the perspective of
  129. 04:20ownership and responsibility, you don't
  130. 04:22need to have anything to take
  131. 04:23responsibility. You can do your own
  132. 04:26small thing to make a massive impact in
  133. 04:28anybody's life. That made a very
  134. 04:30personal impact for me because it was
  135. 04:32mind-blowing to me as at that time
  136. 04:34probably I was still a teenager to think
  137. 04:36about life like that. Until then I just
  138. 04:38thought like the world just happened
  139. 04:40around me. We didn't think we manifested
  140. 04:42something in the world right that's one
  141. 04:43thing. The second thing that affected me
  142. 04:46very deeply was the concept of paying it
  143. 04:48forward. So when I was in AWS we ran the
  144. 04:51simulations helped scale the simulations
  145. 04:53for running COVID predictions in terms
  146. 04:56of spread of COVID. And at the time of
  147. 04:57course COVID was just starting. The
  148. 04:59shutdowns had never happened anywhere
  149. 05:01else. No state had ever shut anything
  150. 05:02down. And we helped one of the state of
  151. 05:05California divisions scale their
  152. 05:06simulations to predict when the state of
  153. 05:09California is potentially going to run
  154. 05:11out of beds. So what we ended up doing
  155. 05:12was we helped them take their models and
  156. 05:15then scale them on massive amount of
  157. 05:16compute cuz that's what we're really
  158. 05:18good at doing. And it predicted that in
  159. 05:205 days or so the total number of beds in
  160. 05:22California would be exhausted. And uh
  161. 05:24that simulation along with all the other
  162. 05:27data they had was instrumental in the
  163. 05:29state of California announcing the first
  164. 05:31statewide shutdown. And that same
  165. 05:33simulation is being run by many
  166. 05:35countries today. And it was a a deeply
  167. 05:37touching project to be working on
  168. 05:39because I didn't go to Amazon thinking
  169. 05:40that's what I would be doing. At the
  170. 05:42time we didn't realize it but it had a
  171. 05:44huge impact on a lot of things in the
  172. 05:46world. And it is truly humbling that the
  173. 05:48work we did helped in a small way in
  174. 05:50making the world a better place. In
  175. 05:52Steve Jobs's world, it is like make a
  176. 05:54dent in the universe. But it's really
  177. 05:55around saying, look, whatever we find
  178. 05:57ourselves in, whatever little we do,
  179. 05:59make sure that we leave the place in a
  180. 06:01better world than we found it. The group
  181. 06:04of people that we have with us, I would
  182. 06:06say none of them need this job. None of
  183. 06:08them need to be here working God knows
  184. 06:10how many hours they're working daily.
  185. 06:12They're here because they believe in the
  186. 06:13mission. They believe in why we do what
  187. 06:15we do and how we do what we do.

  188. Principle 2: Don't Put Yourself in a Box

  189. 06:21If you ask me how my career got built is
  190. 06:24really being dumb enough not to say no.
  191. 06:27And the reason I use the word is
  192. 06:29whenever a problem was given to me even
  193. 06:31very early days I didn't know that other
  194. 06:34people would never touch that problem.
  195. 06:35It was given to me because I was silly
  196. 06:37enough not to ask why would we do this?
  197. 06:39We basically said oh it is an
  198. 06:41interesting problem. Let me go try
  199. 06:43solving it. I failed more often than I
  200. 06:45succeeded. But the times I succeeded, I
  201. 06:47succeeded in things that everybody else
  202. 06:49said cannot be done. The problems that
  203. 06:51were the highest value, the problems
  204. 06:53that were the most difficult to do and
  205. 06:55failure was okay. So the way it started
  206. 06:58was I was at Intel. I was leading the
  207. 07:00cloud and AI strategy and execution
  208. 07:02teams. My charter was to help Intel sell
  209. 07:04more AI hardware. But we knew that the
  210. 07:07largest workload was large language
  211. 07:09models. Genai is going to happen. This
  212. 07:11is 2022. And this was 6 months before
  213. 07:13Chad GBD was launched. So we had built
  214. 07:15an AI supercomputer and then we were
  215. 07:17trying to build a model for that one.
  216. 07:19And we actually worked with a very large
  217. 07:21company trying to get them to buy Intel
  218. 07:24hardware and the project went all the
  219. 07:26way up to a final approval where
  220. 07:28everything broke loose. There was a
  221. 07:31misunderstanding in terms of who was
  222. 07:33going to fund what portions of it which
  223. 07:34basically ended up being the project
  224. 07:36completely stopping and it almost like
  225. 07:39the rocket analogy is the rocket blew up
  226. 07:41on the launchpad. That was a failure
  227. 07:43everything started with. But because we
  228. 07:45had already done the modeling, we then
  229. 07:47went to a customer and said, "Hey, we
  230. 07:49can do this for you. Would you want to
  231. 07:51do this?" They said, "Oh, okay. We can
  232. 07:53try this. Let's do a search model." We
  233. 07:54actually built a model for them. That
  234. 07:56model ended up being a software product.
  235. 07:58We built the software. We did all that
  236. 08:00to sell hardware. The customer comes
  237. 08:02back and says, "Maybe we'll buy your
  238. 08:04hardware at some point. Can I buy your
  239. 08:05software?" So, in a sense, that was the
  240. 08:07second biggest failure. We tried to sell
  241. 08:09hardware. We couldn't sell the hardware,
  242. 08:11but we could sell the software. And
  243. 08:12within a year, we went from having a
  244. 08:15project that is almost in the bag that
  245. 08:17was already going to become the biggest
  246. 08:19project out there. That project failed.
  247. 08:21Then we went and got another project to
  248. 08:23sell hardware. That project also failed.
  249. 08:25Ended up being Articulate the Company.
  250. 08:27And if you think about if we had stopped
  251. 08:28at the first failure, we would have
  252. 08:30never gotten the second project. If we
  253. 08:31had stopped at the second failure, we
  254. 08:33would have never got an articulate. The
  255. 08:34perspective of looking at it from failed
  256. 08:37projects getting turned around. Doesn't
  257. 08:38matter how many times we fall down, can
  258. 08:40we actually get up

  259. Principle 3: Why 95% of AI Projects Never Reach Production

  260. 08:45say principles we go by in the company
  261. 08:46which is we typically do not do PC's.
  262. 08:49PC's meaning proofs of concept. We only
  263. 08:51do production pilots. You may ask what's
  264. 08:53the difference? The difference is a P is
  265. 08:55something that anybody can do. Honestly,
  266. 08:58even a high school student with a good
  267. 08:59enough knowledge of some of these tools
  268. 09:01can build a PC with a few tens of
  269. 09:03documents. The difference between a PC
  270. 09:04and a pilot is pilot is production
  271. 09:06scale. You have to go after your most
  272. 09:08complex use cases. In fact, we go into a
  273. 09:10customer and say, "Give us your messiest
  274. 09:12of data set. Give us your most complex
  275. 09:14problem and let's actually show you how
  276. 09:17to actually take that to production."
  277. 09:18Right? That's what we mean by a
  278. 09:19production pilot. And the other big
  279. 09:21difference between a production pilot
  280. 09:22and a PC is once you're done with your
  281. 09:25production pilot, turning on the
  282. 09:26production is just a matter of a
  283. 09:28contract signing. All the technical work
  284. 09:30is already done. And the third thing you
  285. 09:32might ask is okay that means this should
  286. 09:33take a long time to do. We actually cap
  287. 09:36production pilots to be between 4 to 8
  288. 09:38weeks long. No more than that. Even the
  289. 09:40most complex of use cases that we do is
  290. 09:43entirely done within 8 weeks. It's very
  291. 09:45important because when a customer hears
  292. 09:48this statement which is like give me a
  293. 09:50nice clean data set and it's nicely
  294. 09:52curated give me a small problem for me
  295. 09:55to show you that it actually works after
  296. 09:57you make the P to get to production it
  297. 09:59ends up taking 6 months or a year and
  298. 10:00most of the projects if you see the
  299. 10:02latest MIT study that said 95% of genai
  300. 10:05projects don't go to production is
  301. 10:07because of that because people do PC and
  302. 10:09it never moves forward but it's a very
  303. 10:11different uh philosophy

  304. Principle 4: Stop Chasing Low Hanging Fruit

  305. 10:16Last count a few months ago, there were
  306. 10:1828,600 startups in the world that
  307. 10:21claimed to be Gen AI startups. I'm sure
  308. 10:24the number is more than doubled since.
  309. 10:26The reason for that is everybody, I
  310. 10:28believe, believes that they're solving
  311. 10:30an important problem. I don't think
  312. 10:32anybody goes out and starts a startup or
  313. 10:34works so hard to solve a problem they
  314. 10:35don't believe is important. However,
  315. 10:37what they're solving for is in different
  316. 10:40domains and what they're solving for is
  317. 10:42in domains that they think is hot. Now,
  318. 10:44there's two ways to solve a problem, two
  319. 10:46ways to tackle a problem. The first way
  320. 10:48is, of course, go after where things are
  321. 10:49hot. That's the red ocean theory. Or go
  322. 10:52after things that are the blue ocean
  323. 10:53theory, where there's nobody there. Now,
  324. 10:55a path less traveled is less traveled
  325. 10:57for a reason, is hard, and it's filled
  326. 11:00with more failures than the path more
  327. 11:02traveled with. But it really depends on
  328. 11:04what you have stomach for and what
  329. 11:05you're actually going and solving and
  330. 11:07what you have resources for that you can
  331. 11:08go do that. I don't think there is any
  332. 11:10one right or wrong answer. It is just
  333. 11:12about what is personal for you. Now do I
  334. 11:14believe that these companies are going
  335. 11:15to make an impact? Absolutely. They
  336. 11:17already are. They've changed the way the
  337. 11:19world in so many fundamental ways that
  338. 11:21it's no longer going to be the same.
  339. 11:23Same thing that OpenAI did. Open AAI was
  340. 11:25a small startup when they started. they
  341. 11:27have fundamentally changed the world and
  342. 11:29I think that's going to continue growing
  343. 11:31that we believe that the lowhanging
  344. 11:33fruit is first and foremost the most
  345. 11:36crowded market out there every company
  346. 11:38is a geni company and every large
  347. 11:40company is trying to do something with
  348. 11:41genai where do they go they usually go
  349. 11:43to marketing they go to finance they go
  350. 11:46to HR partly because it's the lowest
  351. 11:48hanging fruit because you have use cases
  352. 11:50but if you make a mistake not much is
  353. 11:52going to change it's not a a company
  354. 11:54threatening event however if If you look
  355. 11:57at what is the most important thing that
  356. 11:59the company is doing, it's mostly not
  357. 12:01related to any of these things. A
  358. 12:02manufacturing company does
  359. 12:03manufacturing. An aerospace company
  360. 12:05would do design or manufacturing or
  361. 12:07maintenance. We go after use cases in
  362. 12:10those particular segments. And we
  363. 12:11actually go after what makes a big
  364. 12:14difference for the company's topline.
  365. 12:16And the reason for that is if a company
  366. 12:18can actually find new business, find new
  367. 12:21ways to make money out of AI, that is
  368. 12:23something that will be longerlasting.
  369. 12:25Whereas productivity use cases typically
  370. 12:27go after the bottom line which
  371. 12:29everybody's going to do whether you like
  372. 12:32it or not today you're doing it or not
  373. 12:33every company is going to use some
  374. 12:35productivity tool to get their AI use
  375. 12:37case to reduce their burden for what
  376. 12:39they're doing that's not going to give
  377. 12:41differentiation for any company so we
  378. 12:43believe that we work on things that give
  379. 12:45a company differentiation that gives you
  380. 12:47longerlasting value and we also have a
  381. 12:49philosophy that we fundamentally
  382. 12:50grounded on saying look we build our
  383. 12:53business based on the fact that we added
  384. 12:55value to your business. So we can share
  385. 12:57in the value that's how our business
  386. 12:59grows. Our business is not necessarily
  387. 13:01growing by making you more efficient.

  388. No Safety Net in Startups, Just Move

  389. 13:07So the biggest thing that is different
  390. 13:08of course being in a startup is that you
  391. 13:10don't have any safety net like any other
  392. 13:12job I've had there's always somebody who
  393. 13:15I could go to for help but ultimately
  394. 13:17the responsibility lies with me. Meaning
  395. 13:20the buck stops with me which is a huge
  396. 13:23change when it comes to it. That took a
  397. 13:25little bit of time to adjust, but the
  398. 13:26nice thing about being in a startup like
  399. 13:28this is you don't have any time to sit
  400. 13:30and think. You just have to deal with it
  401. 13:32and keep rolling, right? So that's a
  402. 13:34good thing because you don't have the
  403. 13:35time to freak out. Uh I would say the
  404. 13:37biggest thing I'll do differently is to
  405. 13:39make sure that people who come to the
  406. 13:42company understand the fact that there's
  407. 13:44no safety net. And we've had our fair
  408. 13:47share of turnovers. And even people who
  409. 13:48were in the early early teams, we've had
  410. 13:51to let go. Partly because they may be
  411. 13:53perfect candidates, perfect employees in
  412. 13:55a large company, perfect candidates or
  413. 13:57outstanding candidates in a medium
  414. 13:59company, but a terrible misfit in a
  415. 14:01small company. And the reason for that
  416. 14:03is you really need to have a hustle that
  417. 14:06goes way beyond anything you've seen in
  418. 14:08life. Like I'll tell you a general
  419. 14:10conversation that would usually go. We
  420. 14:12ask for say going and meeting with a
  421. 14:14customer or something. we are expecting
  422. 14:15a week goes by you go ask them hey did
  423. 14:18this get done and they'll say no I sent
  424. 14:19an email we are still waiting for a
  425. 14:21response now in a large company that's a
  426. 14:23perfectly fair statement to make no
  427. 14:25problem you'll get a response and
  428. 14:26typically say you're an Amazon or you're
  429. 14:28in Google or Microsoft you send a
  430. 14:29response people respond back the same
  431. 14:31people if you're sending an email from a
  432. 14:34startup they don't respond most of the
  433. 14:35time they don't have time to do that but
  434. 14:37second we don't have the time to wait we
  435. 14:40send an email then call the person then
  436. 14:41text the person and text their wife if
  437. 14:43we have to to make sure we get a
  438. 14:44response. There have been cases where we
  439. 14:47would say, "Look, do you want me to go
  440. 14:48pick up your kids so that I can actually
  441. 14:50get 15 minutes with you?" It's that kind
  442. 14:52of hustle that you rarely almost ever
  443. 14:55need in any other job, but you need it
  444. 14:58here. And for that, you also need
  445. 14:59humility because normally you would say,
  446. 15:01"Well, I am the CEO. I can't be going
  447. 15:05and doing something." It's exactly the
  448. 15:07opposite. In fact, you have to do
  449. 15:09everything. Meaning, I can work 18 hours
  450. 15:11a day. I can work 7 days a week. I would
  451. 15:13still not get to even 30% of my tasks.
  452. 15:16If you're in an early stage startup,
  453. 15:18there is no such thing called work life
  454. 15:19balance. Don't kid yourself. There is
  455. 15:21work, there is life, it's the same
  456. 15:23thing. It's what it is. It's just that
  457. 15:26the people's mindset has to be very
  458. 15:28different. I don't think twice before
  459. 15:30clearing out all the coffee cups in the
  460. 15:32office. If that's what I have to do, if
  461. 15:34I see something and I have the time, I
  462. 15:36no problem doing it. If I have to be the
  463. 15:38one going and talking to the the biggest
  464. 15:40investor out there raising money, I will
  465. 15:42do that too. So there's no such thing
  466. 15:44called a job is beneath you and there's
  467. 15:46also no such thing called you can just
  468. 15:48delegate and walk away. You cannot
  469. 15:49delegate responsibility.
  470. 15:52Some of the people have asked me a
  471. 15:54question like have you always wanted to
  472. 15:55start a startup? Answer is no. I was
  473. 15:57happy being a researcher. I trained to
  474. 15:59be a researcher and I ended up becoming
  475. 16:00one. I was more than happy. The business
  476. 16:02side was entirely accidental. I didn't
  477. 16:05go to business school. Everything I
  478. 16:06learned about business is what I learned
  479. 16:08in the job. For me, it's a kid in a
  480. 16:10candy store moment all over again
  481. 16:11because we're solving problems that very
  482. 16:13few people would solve. Why is
  483. 16:14democratization of technology important
  484. 16:17for me? Partly because it's a universal
  485. 16:20equalizer. And the reason for that is
  486. 16:22think about science and math and even
  487. 16:24say philosophy in general is universal.
  488. 16:27It's deeply liberating. There is no
  489. 16:29inequality respect of where they are.
  490. 16:31And if you actually allow people to use
  491. 16:34that more easily that's the most
  492. 16:36equitable thing you can do back to the
  493. 16:38society and uh that's really what's been
  494. 16:40driving and also if you look at it that
  495. 16:43also comes from how I grew up like I
  496. 16:45grew up in a world where things were not
  497. 16:47easily available like there was no cable
  498. 16:49TV when I was growing up until I was in
  499. 16:51high school there was no internet and as
  500. 16:54crazy as it seems to think about it
  501. 16:55today going from there to sitting in any
  502. 16:59coffee shop accessing anything in the
  503. 17:02world happened in less than a lifetime,
  504. 17:04right? The unlock it gives is massive.
  505. 17:07Think of the child who may not
  506. 17:09necessarily have the resources, may not
  507. 17:11have the backing being able to actually
  508. 17:13improve their life because of this.
  509. 17:15That's really what drives us. And the
  510. 17:16reason why I like the word
  511. 17:17democratization is it's the most
  512. 17:20equitable thing you can do. You are
  513. 17:22enabling people to pick themselves up
  514. 17:24from whatever situation they find
  515. 17:26themselves in. Some of us have been
  516. 17:28super fortunate to to have had education
  517. 17:30to have had to be in places where we
  518. 17:33could actually see these things and
  519. 17:34learn from that. A lot of the people are
  520. 17:36not as fortunate. By the 10 years, we
  521. 17:38would have become the platform of choice
  522. 17:40for domain specific applications
  523. 17:42anywhere in the world in any domain you
  524. 17:44have. And that's really the mission that
  525. 17:45we're after. And to become the platform
  526. 17:47of choice, we also have to be giving
  527. 17:49disproportionate value back to all of
  528. 17:51those outcomes. Right? That's the one
  529. 17:52thing. And the way we would have changed
  530. 17:54the world really is to make sure that we
  531. 17:56enable every single person working in
  532. 17:59every single company to have their own
  533. 18:01digital twins multiple times a day and
  534. 18:04be able to make meaningful outcomes for
  535. 18:06themselves as well as their company.
  536. 18:08Meta announced in their super
  537. 18:09intelligence lab they're doing that's
  538. 18:10what they announced that they're going
  539. 18:12to do for personal life for people
  540. 18:14outside. We're saying we'll do that for
  541. 18:16the enterprise. Nine out of 10 startups
  542. 18:18fail. And even the 1% one startup that
  543. 18:20succeeds doesn't become a Meta or a
  544. 18:23Google. But the pain you endure is so
  545. 18:26massive that you really need to have an
  546. 18:29inner guiding principle to actually get
  547. 18:31you past that. It is, I guess, rooted on
  548. 18:33the belief that we can actually make a
  549. 18:35small difference. Not because we think
  550. 18:38we're smarter than anybody else. It's
  551. 18:39because we have perseverance. So what I
  552. 18:44keep telling my team and I keep telling
  553. 18:45myself as well is we can lose but we
  554. 18:48can't be beat.
  555. 18:51Well, I would say my greatest source of
  556. 18:52strength has been my wife. She's
  557. 18:54somebody who's bought balance in my life
  558. 18:56because I'm a deeply I would say
  559. 18:57imbalanced person. Meaning if I get
  560. 18:59passionate about something, I'll forget
  561. 19:00everything and just dive in. And so
  562. 19:02bringing a semblance of balance is
  563. 19:04something that she taught me. And it's
  564. 19:05also something where I've been enabled
  565. 19:08by a lot of people, but I don't think
  566. 19:10I've been enabled as much by anybody
  567. 19:11else other than her.
What Young Founders Get Wrong About Startups | Articul8, Arun Subramaniyan — Transcriptly