YouTube2Text

The First Big Victim of the AI Bubble — Transcript

by Crayon Capital · 3,018 words · 575 segments · language en · Watch on YouTube

Full transcript

  1. 0:00July 2026, this 24-year-old loses $35
  2. 0:04billion in a single month betting on AI.
  3. 0:07The largest trading loss in history.
  4. 0:10But he didn't lose because his
  5. 0:12prediction was wrong.
  6. 0:13>> [music]
  7. 0:13>> It was coming true.
  8. 0:14So how do you see the future correctly
  9. 0:17and still lose $35 billion?
  10. 0:20To answer that, we need to start with
  11. 0:22the biggest problem in 2024.
  12. 0:27The world has an AI problem. Not the AI
  13. 0:30is coming for your job problem.
  14. 0:32>> [screaming]
  15. 0:32>> AI was getting good, scary good. And
  16. 0:36whoever ended up controlling it would
  17. 0:37hold more power than any company in
  18. 0:39history.
  19. 0:40So the solution seems obvious, right? If
  20. 0:43you can't beat them, join them.
  21. 0:45Invest in the winner. Problem solved.
  22. 0:48Except nobody could tell you who the
  23. 0:49winner was. We're first. We're the best.
  24. 0:53>> [music]
  25. 0:53>> We invented this. Cute. We have your
  26. 0:55entire internet and a nuclear reactor.
  27. 0:59Is everyone using him now?
  28. 1:01Every few months, the leaderboard
  29. 1:03flipped. A company that was dead last in
  30. 1:05March was untouchable by September.
  31. 1:08Then somebody else leapfrogged them
  32. 1:10>> [music]
  33. 1:10>> and the whole thing reset. For years,
  34. 1:12Wall Street tried to analyze it.
  35. 1:14Nothing. No pattern. It's too
  36. 1:16unpredictable.
  37. 1:18Then a 22-year-old claimed he'd figured
  38. 1:20it out. Not just who would win,
  39. 1:22but how the entire decade would unfold.
  40. 1:25Meet the boy wonder, Leopold
  41. 1:27Aschenbrenner. [music]
  42. 1:29Germany, 2001.
  43. 1:32Leopold Aschenbrenner is born to two
  44. 1:34doctors. Normal childhood except for
  45. 1:36[music] one detail. The kid is early to
  46. 1:38everything. He enrolled at Columbia
  47. 1:40University at 15, graduated
  48. 1:42valedictorian with a degree in economics
  49. 1:44and math, and a nearly perfect test
  50. 1:46score he was so proud of, he put it on
  51. 1:48LinkedIn.
  52. 1:49Wow. Okay. And your greatest weakness?
  53. 1:53I finish things too early.
  54. 1:56So in 2022, he did what every brilliant,
  55. 1:59idealistic 21-year-old did that year.
  56. 2:01Went to work for the smartest guy in the
  57. 2:03room.
  58. 2:04That guy was Sam Bankman-Fried.
  59. 2:06Leopold joined the FTX Future Fund, the
  60. 2:09charity arm of the crypto exchange,
  61. 2:11staffed by young people who genuinely
  62. 2:12believed they were about to give away
  63. 2:14billions to save the world.
  64. 2:16For a few months, it felt like the
  65. 2:18center of the universe. Everyone was
  66. 2:20brilliant. Everyone was certain. The
  67. 2:23money seemed infinite. And then,
  68. 2:25November 2022 happened. FTX collapsed,
  69. 2:28the billions vanished, and Leopold got a
  70. 2:30front-row seat to what happens when a
  71. 2:32brilliant person, a compelling story,
  72. 2:34and borrowed money collide.
  73. 2:37After FTX, Leopold moved to the company
  74. 2:40building the future he was most worried
  75. 2:41about, OpenAI.
  76. 2:44He joined its superalignment team,
  77. 2:45>> [music]
  78. 2:45>> whose job was essentially to make sure a
  79. 2:47superhuman AI did not destroy
  80. 2:49civilization.
  81. 2:50Small assignment, but less than a year
  82. 2:52later, OpenAI fired him for allegedly
  83. 2:55leaking internal information.
  84. 2:57Leopold disputed that description.
  85. 2:59He said he had shared an AI safety
  86. 3:00document with outside researchers.
  87. 3:03Either way, he was out.
  88. 3:04So, picture him now.
  89. 3:0622,
  90. 3:07no job, no money,
  91. 3:10fired from the most important company on
  92. 3:11the planet. He had exactly one thing
  93. 3:13left, an opinion.
  94. 3:15In June 2024, Leopold published 165
  95. 3:19pages explaining exactly what he
  96. 3:21believed was about to happen.
  97. 3:23He called it Situational Awareness: The
  98. 3:25Decade Ahead.
  99. 3:27Its argument was simple enough to sound
  100. 3:28insane. AI models were getting smarter
  101. 3:31on a straight, [music] predictable line.
  102. 3:33Draw that line forward, he said, and
  103. 3:35machines hit human-level intelligence
  104. 3:37around 2027. Shortly after that, [music]
  105. 3:40smarter than us. Game over.
  106. 3:42Most people read that and started
  107. 3:44arguing about robots stealing their
  108. 3:45jobs. Leopold saw a shopping list.
  109. 3:48Because smarter AI needs bigger models.
  110. 3:51Bigger models [music] need more chips.
  111. 3:53Those chips need memory.
  112. 3:55All of it needs vast humming data
  113. 3:57[music] centers.
  114. 3:59And those data centers need so much
  115. 4:00electricity that the power companies
  116. 4:02hadn't even started planning for it.
  117. 4:04To build all of this, the world might
  118. 4:06[music] need to spend a trillion
  119. 4:07dollars.
  120. 4:08And whoever supplied that construction
  121. 4:09boom was about to get very, very rich.
  122. 4:13Okay, so you might be wondering why call
  123. 4:15it situational awareness. Well, it is
  124. 4:17exactly because screw-ups like you
  125. 4:19[music] are not aware of this situation.
  126. 4:21Only a few hundred people saw what was
  127. 4:23coming.
  128. 4:24And obviously,
  129. 4:25>> [music]
  130. 4:25>> I was one of them.
  131. 4:26But you should be aware of this quick ad
  132. 4:28break, though.
  133. 4:30How many AI subscriptions are you paying
  134. 4:31for right now?
  135. 4:33Chat GPT for research, Claude for
  136. 4:35coding, something else for images, and
  137. 4:37every one of them wants another $20 a
  138. 4:39month.
  139. 4:40Chat LLM by Abacus AI fixes [music]
  140. 4:42that. For $10 a month, you get GPT-6
  141. 4:45Astra, Claude Fable 5.1, Gemini,
  142. 4:48>> [music]
  143. 4:49>> Grok, Deep Seek, and more in one
  144. 4:51dashboard.
  145. 4:52New models appear the day they [music]
  146. 4:53launch, and if you don't know which one
  147. 4:55to use, Root LLM picks the best model
  148. 4:57for your prompt. Use it to research and
  149. 4:59write documents, create presentations,
  150. 5:01and generate images [music] and videos.
  151. 5:04You can also describe an idea and have
  152. 5:06its AI agent build the whole website or
  153. 5:08app. Front end, back end, hosting,
  154. 5:10[music]
  155. 5:11even payments. You can also create AI
  156. 5:13agents that work like personal
  157. 5:14assistants 24/7. [music]
  158. 5:18Head over to chat.llm.abacus.ai
  159. 5:21or [music] click the link in my
  160. 5:22description to get started. Trust me,
  161. 5:24it's the only AI bet that won't cost you
  162. 5:26billions.
  163. 5:28Okay, back to the video.
  164. 5:30The essay spread through Silicon Valley.
  165. 5:32Investors passed it around. Podcasters
  166. 5:35read [music] it. People started calling
  167. 5:37Leopold the Nostradamus of AI.
  168. 5:40Now, most writers would have taken the
  169. 5:41fame, launched a podcast, and sold a
  170. 5:43course for $499.
  171. 5:46Leopold did something far more
  172. 5:48dangerous.
  173. 5:49He decided to bet real money that his
  174. 5:50prophecy was true.
  175. 5:52With other people's money.
  176. 5:53There was just one problem. In his
  177. 5:55entire life, he had never professionally
  178. 5:58managed a single dollar.
  179. 6:02July 2024.
  180. 6:04Leopold launches a hedge fund. He names
  181. 6:06it what else?
  182. 6:08>> [music]
  183. 6:08>> Situational Awareness.
  184. 6:10By every rule of Wall Street, raising
  185. 6:12money should have been impossible.
  186. 6:14>> [music]
  187. 6:14>> I am 22. I had no track record, and my
  188. 6:17resume included one collapsed crypto
  189. 6:19empire
  190. 6:19>> [music]
  191. 6:19>> and one recent firing.
  192. 6:22But I want to manage your money.
  193. 6:24Right.
  194. 6:25You've read the essay.
  195. 6:26Do you think it's wrong?
  196. 6:28No.
  197. 6:29I think it's right.
  198. 6:31Then trust your own conclusion.
  199. 6:33I'm early at everything. Give me my
  200. 6:35money.
  201. 6:36It worked. His backers reportedly
  202. 6:38included some of the most respected
  203. 6:39names in tech, the founders of Stripe, a
  204. 6:42former CEO of GitHub, big-name investors
  205. 6:44who'd sat through a thousand pitches and
  206. 6:46seen every trick in the book.
  207. 6:48Then Leopold built the trade. And
  208. 6:50honestly,
  209. 6:52it was gorgeous.
  210. 6:54See, all of Wall Street was obsessed
  211. 6:56with the same question.
  212. 6:58Which company wins AI? OpenAI? Google?
  213. 7:01[music] Anthropic? Some startup that was
  214. 7:04currently two guys and a Discord server?
  215. 7:07Leopold thought that was the amateur's
  216. 7:08question.
  217. 7:09During a gold rush, he figured, you
  218. 7:11don't need to guess which prospector
  219. 7:12strikes gold. Half of them go broke.
  220. 7:15Half of them go crazy.
  221. 7:17You don't bet on the prospectors at all.
  222. 7:19You sell every single one of them a
  223. 7:21shovel, and you get paid whether they
  224. 7:22strike gold or die of dysentery.
  225. 7:25So, step one,
  226. 7:26buy the picks and shovels. Memory. Data
  227. 7:29centers. Power.
  228. 7:31Names like SK Hynix, Micron, CoreWeave,
  229. 7:34Nebius, Bloom Energy.
  230. 7:36The unglamorous companies that simply
  231. 7:38had to exist, no matter who won the
  232. 7:40actual race.
  233. 7:42Step two, and this is the clever half.
  234. 7:45Short the victims.
  235. 7:47Bet against the software companies that
  236. 7:49are going to be killed by AI, such as
  237. 7:51Adobe. Why pay a monthly subscription
  238. 7:53for the tool when the robot does the
  239. 7:55same job for free?
  240. 7:57That was the brilliance of the
  241. 7:58portfolio.
  242. 7:59Owning both sides was supposed to
  243. 8:01protect him from a market crash. If the
  244. 8:03whole market fell, he would lose money
  245. 8:05on the stocks he bought, but make money
  246. 8:07on the stocks he shorted. Ideally, one
  247. 8:09would cancel out the other.
  248. 8:11He was simply betting that the AI
  249. 8:13winners would do better than the AI
  250. 8:14victims.
  251. 8:15But it was also the trap.
  252. 8:17Leopold ran a few dozen positions, but
  253. 8:19almost every one of them depended on the
  254. 8:21same sentence being true.
  255. 8:23AI will transform the world, and it will
  256. 8:25happen on schedule.
  257. 8:27For a while, that sentence looked
  258. 8:28correct. First half of 2025, up 47%
  259. 8:32after fees, managing about 1.5 billion
  260. 8:36dollars.
  261. 8:37Solid.
  262. 8:38>> [music]
  263. 8:38>> Then the returns became absurd. By late
  264. 8:40June 2026, the fund was reportedly up
  265. 8:43around 439% for the year, and more than
  266. 8:451,000% since launch.
  267. 8:48Assets passed to 20 billion dollars. At
  268. 8:50its peak, [music] the firm controlled
  269. 8:52roughly 45 billion dollars. In less than
  270. 8:54two years, a man who had never traded
  271. 8:56professionally had built one of the most
  272. 8:58watched funds on Wall Street. Other
  273. 9:00investors [music] started admiring him,
  274. 9:02and
  275. 9:03the filing is out.
  276. 9:05What's he holding?
  277. 9:07Memory? Power? Data centers? Same as
  278. 9:10last quarter, but bigger. Buy all of it.
  279. 9:12You don't want to look at the
  280. 9:13fundamentals first?
  281. 9:15He read the future. I'm reading him.
  282. 9:17It's faster.
  283. 9:18Every new buyer pushed Leopold's
  284. 9:20holdings higher. Those higher prices
  285. 9:22improved his returns. Better returns
  286. 9:24attracted more investors.
  287. 9:26>> [music]
  288. 9:26>> More investor money allowed the fund to
  289. 9:28buy even more. So, success created more
  290. 9:32success. But, underneath it was a the
  291. 9:35engine.
  292. 9:36Leverage.
  293. 9:37Leopold wasn't just investing his
  294. 9:39investors' money, he was borrowing
  295. 9:41against it. Roughly three to four
  296. 9:43dollars of debt for every dollar he
  297. 9:44actually had.
  298. 9:46Here is why that feels wonderful. You
  299. 9:47have $100. The bank lends you $400. Now
  300. 9:51you own $500 of stock.
  301. 9:54If those stocks rise 20%, you make $100
  302. 9:57and double your original money.
  303. 9:59That sounds incredible. It is.
  304. 10:02But please do not ask what happens if
  305. 10:04the stocks fall 20% because leverage
  306. 10:06does not simply make the good days
  307. 10:08better. It puts a clock on the bad days.
  308. 10:11If the portfolio falls, the bank asks
  309. 10:13for more collateral. It's called a
  310. 10:15margin call.
  311. 10:16If you cannot provide cash, you must
  312. 10:18sell stocks.
  313. 10:19Not next year when your prediction comes
  314. 10:21true.
  315. 10:22Today. [music]
  316. 10:24Mr. Leopold, history is littered with
  317. 10:26investors who saw the future correctly,
  318. 10:28but still went broke because their
  319. 10:29timing was catastrophically wrong.
  320. 10:32So what happens if your stocks go down
  321. 10:34before your thesis plays out?
  322. 10:36>> [music]
  323. 10:36>> Yeah, great question. I think the only
  324. 10:38thing going down is my estimate of your
  325. 10:41situational awareness.
  326. 10:43I'm serious. So am I. [music] Obviously,
  327. 10:46not blowing up is task number one and
  328. 10:48two.
  329. 10:49Then on July 10th, 2026, [music]
  330. 10:52the company at the center of Leopold's
  331. 10:53entire thesis rang the opening bell in
  332. 10:56New York. It was supposed to be his
  333. 10:58victory lap.
  334. 10:58>> [music]
  335. 10:59>> Instead, it started the clock.
  336. 11:03The company was SK Hynix, the South
  337. 11:06Korean memory chip giant. If Leopold's
  338. 11:08AI prediction was correct, the world
  339. 11:10would need an extraordinary [music]
  340. 11:12amount of high-performance memory.
  341. 11:14That company made it. And Leopold had
  342. 11:16reportedly put 6.5% of his money there.
  343. 11:19This was the purest version of his
  344. 11:21trade.
  345. 11:22SK Hynix crossed onto the American
  346. 11:24market with enormous [music] hype.
  347. 11:26It rose by 13% on the first day, but it
  348. 11:29didn't last.
  349. 11:30The stock fell.
  350. 11:32At first, that should not have mattered.
  351. 11:34Nothing about the actual future had
  352. 11:36changed.
  353. 11:36>> [music]
  354. 11:37>> No data center got canceled. No AI
  355. 11:39company stopped buying chips. It was
  356. 11:41just the ordinary weather of the stock
  357. 11:42market. On a 10-year chart, pure noise.
  358. 11:45[music]
  359. 11:46But Leopold didn't own SK Hynix with
  360. 11:4810-year money. He owned it with money
  361. 11:50that could demand to be repaid today.
  362. 11:53Leveraged bets on Korean stocks began
  363. 11:54unwinding. Regulators in Seoul rushed to
  364. 11:57make traders [music] put up more cash.
  365. 11:59Anyone who could not provide it was
  366. 12:01forced to sell.
  367. 12:02That selling pushed prices lower, which
  368. 12:04triggered more margin calls,
  369. 12:06>> [music]
  370. 12:06>> and created more sellers.
  371. 12:08SK Hynix stock eventually fell more
  372. 12:10[music] than 30% from their peak.
  373. 12:13Then the fire spread to the rest of
  374. 12:14Leopold's portfolio.
  375. 12:16Micron fell. Coreweave, Nebius.
  376. 12:19Some of the AI infrastructure stocks
  377. 12:21dropped by half or more.
  378. 12:24And then his supposedly protective short
  379. 12:25positions turned against him, too.
  380. 12:28Remember the clever half?
  381. 12:29As terrified investors fled the flashy
  382. 12:31AI names, they ran back into the boring,
  383. 12:34safe software companies.
  384. 12:36The exact companies Leopold had bet
  385. 12:37against. So his longs fell while his
  386. 12:40shorts rose.
  387. 12:41Now the leverage took control.
  388. 12:43Introducing the doom loop.
  389. 12:46Your stocks fall.
  390. 12:48Your cushion of real money shrinks.
  391. 12:50The bank calls, [music] asking for more
  392. 12:52cash to cover the loan. But you don't
  393. 12:55have cash.
  394. 12:56So you [music] sell stocks. And selling
  395. 12:58that much, that fast, pushes the price
  396. 13:00down, which shrinks your cushion again,
  397. 13:02and triggers the next call, [music] and
  398. 13:04the next sale.
  399. 13:06Fall, call, sell. Repeat. Faster every
  400. 13:10turn.
  401. 13:11By Friday, July 24th, the damage was
  402. 13:13severe enough that Leopold [music] sent
  403. 13:15a letter warning investors that the fund
  404. 13:16had suffered heavy losses. And also
  405. 13:19asked them for more money.
  406. 13:20In his mind, nothing has changed about
  407. 13:22the thesis. Not one data center got
  408. 13:25canceled this week.
  409. 13:26Only the price had. Everything he wanted
  410. 13:28to own had just gone on sale. [music]
  411. 13:31So, he called it the best buying
  412. 13:32opportunity since early 2025. But, the
  413. 13:36letter actually exposed how vulnerable
  414. 13:37he was. His rivals already knew what he
  415. 13:39owned.
  416. 13:40Now, they knew he might be forced to
  417. 13:42sell.
  418. 13:43So, they sold first, pushing prices down
  419. 13:45before he could get out.
  420. 13:47Over the next three trading days,
  421. 13:48several of his largest positions dropped
  422. 13:51another 30% or more.
  423. 13:53The prophecy had become a countdown.
  424. 13:55And Leopold [music] had somewhere else
  425. 13:57he was supposed to be that week.
  426. 13:58Guests were flying into Carmel, a
  427. 14:00seaside town on the California coast,
  428. 14:02[music] for his wedding.
  429. 14:03A multi-day celebration, a villa
  430. 14:05ceremony, a forest spa honeymoon
  431. 14:07send-off, and
  432. 14:08>> [music]
  433. 14:09>> uh because this was Leopold, a
  434. 14:10pre-wedding colloquium with panels and
  435. 14:12breakout sessions. So, while his guests
  436. 14:14packed, the groom worked the phones.
  437. 14:17"Hey, Jim Bob. [music] I'm offering 10
  438. 14:19billion in positions at a discount."
  439. 14:21"How big a discount?" Uh "Name it."
  440. 14:24"Nah."
  441. 14:25"I'll wait."
  442. 14:26"Tomorrow, it's cheaper."
  443. 14:28That was the answer everywhere he
  444. 14:30called. Without a deal, he'd have to
  445. 14:32dump billions in stocks onto the market,
  446. 14:34pushing [music] prices down even
  447. 14:35further.
  448. 14:36He needed someone who could take the
  449. 14:37entire portfolio off his hands in one
  450. 14:39transaction.
  451. 14:41Longs, shorts, everything.
  452. 14:43Someone with enough money and the balls
  453. 14:45to buy what the rest of Wall Street
  454. 14:47wouldn't touch.
  455. 14:48And one man was willing to do it.
  456. 14:52Introducing
  457. 14:53Ken Griffin, founder of Citadel.
  458. 14:56Griffin and his lieutenants [music]
  459. 14:57pulled an all-nighter tearing through
  460. 14:59Leopold's book, position by position,
  461. 15:01working out what could actually be sold
  462. 15:04and what couldn't. By morning, [music]
  463. 15:06he had an offer.
  464. 15:07"Okay, my offer is
  465. 15:09I would take the entire [music] $16
  466. 15:11billion position, but I want a 10%
  467. 15:13discount." "10%? The prices are already
  468. 15:16down a lot."
  469. 15:18"Correct. That's what makes them worth
  470. 15:20owning."
  471. 15:21"What if they keep falling?
  472. 15:23Then I hold them.
  473. 15:25That's the difference between us.
  474. 15:26You're selling time.
  475. 15:28I'm buying it.
  476. 15:29And Griffin knew exactly what time was
  477. 15:31worth because 18 years earlier he'd been
  478. 15:34the one who couldn't afford it. Back in
  479. 15:362008, Citadel's biggest funds lost
  480. 15:38[music] 55%.
  481. 15:40Griffin had spread his bets across
  482. 15:42stocks, bonds, credit. And when the
  483. 15:44system broke, everything broke together.
  484. 15:47Then his investors asked for their money
  485. 15:49back. Paying them meant selling into the
  486. 15:51panic, locking in every loss before
  487. 15:53anything could [music] recover.
  488. 15:55So Griffin didn't pay them.
  489. 15:57He locked the doors and refused
  490. 15:58withdrawals for months.
  491. 16:00Investors were furious. Investors were
  492. 16:02stuck.
  493. 16:03And Citadel got the one thing Leopold no
  494. 16:05longer had, the ability to choose when
  495. 16:08to sell.
  496. 16:09Now in 2026, [music]
  497. 16:11he was buying from a man who had lost
  498. 16:13that choice completely.
  499. 16:14The trade cleared overnight.
  500. 16:16And the moment it did, the doom loop had
  501. 16:18nothing left to eat.
  502. 16:20Coreweave, SK Hynix, Bloom Energy.
  503. 16:23The same stocks that had been falling
  504. 16:25for three straight weeks started
  505. 16:26climbing.
  506. 16:27Some of them finished the day up more
  507. 16:29than 25%. [music]
  508. 16:31Citadel had reportedly resold most of
  509. 16:33the portfolio at a substantial profit.
  510. 16:35Its stock picking fund
  511. 16:36>> [music]
  512. 16:36>> finished July up 14.2%
  513. 16:39the best month in the firm's history.
  514. 16:42Meanwhile, for Leopold,
  515. 16:44his fund had fallen from roughly $45
  516. 16:46billion at its peak
  517. 16:47>> [music]
  518. 16:47>> to around $10 billion.
  519. 16:49A private stake in Anthropic remained,
  520. 16:51helping preserve what was left.
  521. 16:53$35 billion [music] gone.
  522. 16:56He wrote to his investors and said,
  523. 16:58"Dear partners,
  524. 17:00we let you down this month.
  525. 17:02We came closer [music] to permanent
  526. 17:03capital impairment than is acceptable to
  527. 17:05us.
  528. 17:06Over the past 2 years, we have delivered
  529. 17:08outstanding [music] results despite
  530. 17:10occasional sharp pullbacks. But our fund
  531. 17:13must always be structured [music] such
  532. 17:15that we can take a loss and fight
  533. 17:17another day.
  534. 17:18I take full responsibility for these
  535. 17:20events.
  536. 17:21We are continuing [music] to operate.
  537. 17:23However, we will manage our public book
  538. 17:25on a fully paid for basis.
  539. 17:27These were very expensive scars,
  540. 17:29>> [music]
  541. 17:29>> but my core promise to you is that we
  542. 17:31will not waste the opportunity to learn
  543. 17:33from these events.
  544. 17:35On the portfolio itself, we are very
  545. 17:37[music] optimistic. The underlying
  546. 17:39fundamentals are accelerating at the
  547. 17:41very same time that prices have declined
  548. 17:43significantly.
  549. 17:45Thank you for your patience and your
  550. 17:46partnership.
  551. 17:47>> [music]
  552. 17:48>> I am fully invested alongside you.
  553. 17:50Virtually all of my capital is in the
  554. 17:52fund. I'm available next week for
  555. 17:54one-to-one [music] calls with each of
  556. 17:55you to discuss all of the above.
  557. 17:59As an interim update, [music] our
  558. 18:00current net month-to-date performance is
  559. 18:02minus 67%
  560. 18:04and of net year-to-date [music]
  561. 18:05performance is plus 80%.
  562. 18:08We will follow up to share further
  563. 18:09details.
  564. 18:11All the best, Leopold.
  565. 18:14That is the story of Leopold
  566. 18:16Aschenbrenner,
  567. 18:17>> [music]
  568. 18:17>> the kid who called the decade,
  569. 18:19outperformed almost everyone on Wall
  570. 18:21Street, and still said $35 billion on
  571. 18:24fire in a single month.
  572. 18:26>> [music]
  573. 18:26>> Early to college, early to the future,
  574. 18:29and almost finished his story
  575. 18:31early.

About this transcript

This page contains the full transcript of The First Big Victim of the AI Bubble by Crayon Capital, generated from the public captions YouTube serves with the video. The transcript has 3,018 words across 575 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.