YouTube2Text

보수적인 하워드 막스가 이번엔 AI에 의외의 낙관을 꺼냈습니다... 모른다면서도 돈을 넣은 단 한 곳은 따로 있었습니다 [구루 11월 - 하워드 막스] — Transcript

by 이효석아카데미 · 2,956 words · 492 segments · language en · Watch on YouTube

Full transcript

  1. 0:00To start with the conclusion, despite
  2. 0:01Howard Marks being a very conservative
  3. 0:03and cautious person, he shared some
  4. 0:05surprisingly positive remarks. The world
  5. 0:07moves along a spectrum from good to bad
  6. 0:09, but the market has a habit of greatly
  7. 0:11exaggerating this. Looking at past
  8. 0:13examples, when people become immune to
  9. 0:15worry, market sentiment becomes
  10. 0:16overheated. When it comes to fear, it’s
  11. 0:18not when the market fluctuates like it
  12. 0:20does these days, but when the market
  13. 0:21sees wild swings and people aren’t
  14. 0:23afraid—isn't that when the real crisis
  15. 0:25starts?
  16. 0:30>> Hello. This is Lee Ho-seok. Today, I'm
  17. 0:32joined by Eric for the first time in a
  18. 0:34while. Hello.
  19. 0:35>> Yes, hello.
  20. 0:36>> It's been a while, so what interesting
  21. 0:37topic should we start with?
  22. 0:39>> Well, isn't it a time of chaos?
  23. 0:42I've gathered insights on what
  24. 0:44perspectives the masters hold during
  25. 0:46these chaotic times.
  26. 0:48>> That's right.
  27. 0:49>> First, our first main guest is a legend
  28. 0:51worth listening to during times like
  29. 0:53these. Well, not just "worth" listening
  30. 0:55to. He’s a legend you absolutely must
  31. 0:57listen to.
  32. 0:57>> Right, right.
  33. 0:58>> It’s Howard Marks. Howard Marks is
  34. 1:00famous for his memos, but he’s also
  35. 1:02known by another nickname—the market’s
  36. 1:03thermometer, right? He has an incredible
  37. 1:05knack for judging whether the market is
  38. 1:08overheated, cooled off, or at a
  39. 1:09reasonable level.
  40. 1:11So, we’re going to find out how Marks
  41. 1:13assesses this current chaotic market
  42. 1:15situation.
  43. 1:17To start with the conclusion, despite
  44. 1:19Howard Marks being a very conservative
  45. 1:21and cautious person, he shared some
  46. 1:23surprisingly positive remarks.
  47. 1:25>> I also read his memo and thought, he's
  48. 1:27usually not the type to say things like
  49. 1:29this, but there were many parts that
  50. 1:31read quite positively regarding AI.
  51. 1:34>> That’s correct. Let's start with the
  52. 1:36conclusion first. He stated that some of
  53. 1:38the overheated sentiment in the current
  54. 1:40market has faded. But, this statement
  55. 1:42could lead to some misunderstandings.
  56. 1:45Is the cooling a result of hitting a
  57. 1:46peak and the bubble bursting on the way
  58. 1:48down, or was it a process heading toward
  59. 1:51a bubble where things just settled a bit
  60. 1:52?
  61. 1:55I've been tracking this, and Marks has
  62. 1:57never called a bubble.
  63. 1:59He's said it's overheated.
  64. 2:01He has said we might be in the early
  65. 2:02stages of a bubble, but since he’s never
  66. 2:04called a bubble, it’s not that we’ve hit
  67. 2:07a peak and are crashing down.
  68. 2:08>> It's not that.
  69. 2:09>> That's right. He said that the
  70. 2:10overheating has faded while we are in a
  71. 2:12phase heading toward a bubble. He also
  72. 2:14mentioned that the market seems unaware
  73. 2:16of what smart investors already know.
  74. 2:19Marks emphasizes that the world moves
  75. 2:21along a spectrum from good to bad, but
  76. 2:23the market has a habit of greatly
  77. 2:25exaggerating this.
  78. 2:27It has a habit of swinging between
  79. 2:29heaven and hell, and that’s exactly the
  80. 2:30atmosphere we’re seeing right now. The
  81. 2:32situation isn't actually that bad, but
  82. 2:34people are exaggerating it into hell,
  83. 2:36and smart investors are saying this is a
  84. 2:38market where you can spot that
  85. 2:40discrepancy.
  86. 2:41>> So, are you a smart investor? How about
  87. 2:44you?
  88. 2:45The key is understanding what smart
  89. 2:47investors know that the general public
  90. 2:49doesn't.
  91. 2:50>> Right, and they spoke quite positively
  92. 2:52about the stock market while saying that
  93. 2:53. Then there's the credit market, which
  94. 2:55is another important sector, isn't it?
  95. 2:57Since everyone has flocked to the credit
  96. 2:59market to invest in AI, and Marks '
  97. 3:01specialty is actually credit.
  98. 3:03To start with the conclusion on credit:
  99. 3:05it’s in good shape.
  100. 3:07He stated there is no concern of a
  101. 3:09credit crunch or the credit market
  102. 3:11collapsing, as some people worry about.
  103. 3:14He pointed out that just like the stock
  104. 3:16market, the credit market also has
  105. 3:18cycles, and he explained this using this
  106. 3:20graph.
  107. 3:23Looking at the current cycle, he
  108. 3:25mentioned we were near a peak until
  109. 3:27recently.
  110. 3:30That’s because there are signals that
  111. 3:31appear at the top.
  112. 3:35Since the market is so good, everyone
  113. 3:38wants to lend money, which attracts more
  114. 3:40people, and that creates a momentum
  115. 3:42effect that generates returns, leading
  116. 3:44to a point where people are practically
  117. 3:46begging to lend money, which lowers
  118. 3:48lending standards—that’s a classic peak
  119. 3:50signal, and that’s the atmosphere we had
  120. 3:53.
  121. 3:55Because of AI, the mood was "AI will
  122. 3:57definitely succeed, just lend the money
  123. 3:59," but haven't there been some negative
  124. 4:01events recently, like the CoreWeave
  125. 4:03situation? As a result, doubts have
  126. 4:05started to emerge in the market.
  127. 4:06>> Hmm.
  128. 4:08>> The tone has shifted from "AI cannot
  129. 4:10fail" to "maybe there are some risks,"
  130. 4:12and he says this cooling off has
  131. 4:14actually moved us down from the peak
  132. 4:16into a much healthier situation.
  133. 4:19>> Doubt is a good sign. I think this is an
  134. 4:21important point you should write down.
  135. 4:23The market is doubtful? Well, then
  136. 4:25that’s a good sign. It repeats this
  137. 4:27cycle of going up and down, and since AI
  138. 4:30has such high volatility, when it drops
  139. 4:32, it feels like the world is ending. But
  140. 4:34that process of doubt—if AI wasn't the
  141. 4:36real deal, you'd have to keep doubting
  142. 4:38it, but if it is, that doubt is actually
  143. 4:40a positive indicator.
  144. 4:41>> He also talked about the AI industry. He
  145. 4:43made some very positive comments
  146. 4:44regarding the industry as well. The AI
  147. 4:46cycle is still in its very early stages
  148. 4:48. They say it’s like being in the third
  149. 4:50inning of a baseball game.
  150. 4:53The reason is that the infrastructure
  151. 4:56built so far—like data centers and
  152. 4:58semiconductor plants—was designed with
  153. 5:00future demand in mind, and that
  154. 5:02anticipated demand is now emerging, with
  155. 5:04strong demand still ahead and many
  156. 5:06large-scale construction projects
  157. 5:08remaining, so they view the AI cycle as
  158. 5:11still being in its early stages.
  159. 5:15>> Yes.
  160. 5:16>> And in terms of corporate adoption, if
  161. 5:18you look at computing demand, it’s
  162. 5:21currently concentrated in a few frontier
  163. 5:23labs, and the cost of AI adoption
  164. 5:25relative to S&P 500 revenue is only at
  165. 5:28about the 0.1%level.
  166. 5:31That’s why we’re just at the beginning
  167. 5:33of the adoption phase, and since Max
  168. 5:35manages an investment firm, he must be
  169. 5:37using AI quite a lot, right?
  170. 5:39He’s found that companies need to
  171. 5:41completely reorganize their work
  172. 5:42processes and organizational structures
  173. 5:44to adopt AI more easily. Also, they need
  174. 5:46to feel the cost-effectiveness for
  175. 5:48full-scale adoption to happen; looking
  176. 5:50at things like Palantir’s earnings
  177. 5:52reports and seminars from domestic tech
  178. 5:53firms, while people used to focus on how
  179. 5:55much AI was used or how many tokens were
  180. 5:57consumed, now they’re asking, "What did
  181. 5:59you actually build?" They’ve started
  182. 6:01checking for cost-effectiveness, asking
  183. 6:03what progress has been made.
  184. 6:04>> That’s right.
  185. 6:05>> Even Amazon used to base performance
  186. 6:07evaluations on token usage, but they’ve
  187. 6:09scrapped that and changed their metrics
  188. 6:11to ask, "So, what did you actually
  189. 6:13achieve?"
  190. 6:15>> I think the point that inference costs
  191. 6:17are at 0.1%is the key.
  192. 6:19I mean, look at it this way. If S&P 500
  193. 6:21companies’ revenue doesn't drop, it will
  194. 6:24continue to grow. And they will spend a
  195. 6:26portion of that revenue on costs. But
  196. 6:29the current spending on AI is only 0.1%
  197. 6:32. How much do you think they could end
  198. 6:33up spending? When you think about it,
  199. 6:35people are worried that a significant
  200. 6:37portion of labor costs, raw material
  201. 6:39costs, and other expenses could be
  202. 6:41replaced by AI. Whenever I talk about
  203. 6:43this, I tell people to pick one side or
  204. 6:45the other. Isn’t it contradictory to
  205. 6:48worry about losing jobs to AI while
  206. 6:49simultaneously thinking this expenditure
  207. 6:52won’t go up? I mean, just pick one.
  208. 6:54Either don’t worry about it, or consider
  209. 6:57that if the current AI expenditure is
  210. 6:590.1%of revenue, hitting 1%would be a 10
  211. 7:02-fold increase. And 2%would be a 20-
  212. 7:04fold increase. So, looking at the speed
  213. 7:06of current progress, you might think, "
  214. 7:09What should we do?" but I feel like you
  215. 7:11should either stop worrying or just
  216. 7:13commit to it; you have to pick one. Yes
  217. 7:15.
  218. 7:15>> He also talks about what comes after
  219. 7:17adoption. As mentioned earlier, having
  220. 7:19used it himself, he speaks in terms of
  221. 7:21the commercialization of AI models.
  222. 7:23Going forward, performance within the
  223. 7:25same job groups will inevitably become
  224. 7:27similar, so the differentiator will be
  225. 7:29how humans use AI to make judgments;
  226. 7:31that gap in quality is where corporate
  227. 7:33competitiveness will be decided. That’s
  228. 7:35why people talk so much about how the
  229. 7:37market has entered the AI adoption phase
  230. 7:39. He emphasized that in this adoption
  231. 7:41phase, companies that utilize AI
  232. 7:43effectively will be the ones that truly
  233. 7:45stand out.
  234. 7:45>> That’s good.
  235. 7:46>> Yes. Regarding the possibility of a
  236. 7:48bubble, while he spoke positively, he
  237. 7:50suggests that when discussing whether a
  238. 7:52bubble is coming or the current status,
  239. 7:54we must break AI down into three
  240. 7:55categories. First, technology, capital
  241. 7:57expenditure, and stocks. As for the
  242. 7:59technology, it is not a bubble. It is an
  243. 8:01undeniably transformative technological
  244. 8:03innovation, and as for capital
  245. 8:04investment, it is unclear if that is a
  246. 8:06bubble. Because although people worry
  247. 8:08about the large amounts being spent,
  248. 8:09>> Hmm.
  249. 8:10>> if the earnings generated are just as
  250. 8:12large, then it isn't a bubble. Therefore
  251. 8:14, the correct answer is that we don't
  252. 8:15know yet, and regarding whether stocks
  253. 8:17are a bubble, as I mentioned before, the
  254. 8:19overheated state seems to have calmed
  255. 8:20down significantly. Meaning it’s not a
  256. 8:22bubble.
  257. 8:22>> Yes.
  258. 8:23>> He diagnosed that the current bubble
  259. 8:25progression is relatively healthy. Yes.
  260. 8:27At the same time, if he talks so
  261. 8:29positively about AI, how is he investing
  262. 8:31? Regarding AI investment, he said that
  263. 8:33while the situation is healthy, he is
  264. 8:35staying out because he doesn't know.
  265. 8:36That’s what he said. Of course, I will
  266. 8:38mention this later. He does say that. He
  267. 8:40claims that AI is the most uncertain
  268. 8:41technology he has seen in his 60-year
  269. 8:43career. It’s difficult to predict how
  270. 8:45far the technology will advance, how the
  271. 8:47industrial landscape will change to
  272. 8:49determine winners and losers, and
  273. 8:50exactly how the business models will
  274. 8:52work to generate money from AI. For
  275. 8:54instance, Anthropic is receiving a lot
  276. 8:56of attention right now, but if you ask
  277. 8:58what new business they will be in five
  278. 9:00years before their IPO, no one can
  279. 9:02answer. Hmm. It’s that transformative of
  280. 9:04a technology, and since Marks is a value
  281. 9:06investor, wouldn’t you agree? He needs
  282. 9:08to be able to forecast what future cash
  283. 9:10flows will look like. You need to build
  284. 9:12scenarios, but AI is difficult to value
  285. 9:15because of the many uncertainties
  286. 9:16mentioned earlier. He mentioned that
  287. 9:19data centers and infrastructure are at
  288. 9:21least a bit better. Does that mean you
  289. 9:22shouldn't invest? That's not it. Because
  290. 9:24this is such a revolutionary technology
  291. 9:26and, as we saw earlier, we are still in
  292. 9:28the early stages of the cycle, so it's
  293. 9:30not about not investing, but the
  294. 9:31important thing is to invest while
  295. 9:33acknowledging there is much you don't
  296. 9:35know. So, don't just think, "Oh, this
  297. 9:36seems like it will do well." Don't go
  298. 9:38all-in with leverage just because you
  299. 9:40think it’s bound to go up; don't do
  300. 9:41things like that.
  301. 9:42>> When we hear the word "uncertain," we
  302. 9:44often see it as a negative. People only
  303. 9:46think in the negative sense, but we also
  304. 9:48have something called "upside risk." So
  305. 9:49, saying "I don't know what will happen
  306. 9:51with AI" can mean AI might not be able
  307. 9:53to do anything, but conversely, it can
  308. 9:55also mean I don't know what AI might be
  309. 9:57capable of. It’s that kind of talk as
  310. 9:58well. Watching the interview, what I
  311. 10:01felt about Howard Marks' remarks is that
  312. 10:03he is truly and thoroughly humble about
  313. 10:05AI technology.
  314. 10:08I don't know how far this will go.
  315. 10:09That’s why he says he won’t even short
  316. 10:11it. Because shorting it is dangerous.
  317. 10:14Who knows what will happen? That was one
  318. 10:16of the most impressive things from his
  319. 10:18interview, I think.
  320. 10:20When talking about Anthropic, he said, "
  321. 10:22I have no idea what Anthropic will be
  322. 10:24selling in three years."
  323. 10:26Because three years ago, their revenue
  324. 10:28was zero.
  325. 10:30But look at where they are now, and I
  326. 10:32can't even imagine what they’ll be
  327. 10:33selling next. So, obviously, you
  328. 10:35shouldn't short it. Yes.
  329. 10:37>> If so, regarding the fields he is
  330. 10:38investing in, as mentioned earlier, he
  331. 10:40is investing in data centers.
  332. 10:42>> I see.
  333. 10:42>> He stated that he is investing in
  334. 10:44related projects. When you look into it
  335. 10:45, he approaches this very conservatively
  336. 10:47and cautiously as well. He doesn't just
  337. 10:49invest in any data center; since there
  338. 10:51are a ton of people saying they'll build
  339. 10:53them, he said he only invested in
  340. 10:55projects where he believed a limited
  341. 10:56number of data centers would actually be
  342. 10:58built, considering geographic traits,
  343. 11:00local systems, and regulations.
  344. 11:02Also, considering the Marks I have known
  345. 11:05so far, he made a very bold statement,
  346. 11:07as he rarely uses the word "must."
  347. 11:10He attached the condition "must" to
  348. 11:13power. Power will "must" be a bottleneck
  349. 11:15. He said that AI infrastructure will
  350. 11:17increase by three to ten times in the
  351. 11:19future, so where will we get three to
  352. 11:21ten times the power? So, another memo
  353. 11:23was uploaded this morning.
  354. 11:26I saw that he even scored the
  355. 11:27predictability of these AI technology
  356. 11:29fields.
  357. 11:32Electricity is the most predictable
  358. 11:34sector.
  359. 11:35Score.
  360. 11:36>> The fact that power will be in short
  361. 11:38supply is highly predictable.
  362. 11:40However, I personally find it difficult
  363. 11:42to use the word "certainly" here.
  364. 11:45The reason is that AI infrastructure
  365. 11:47will likely require ten times more power
  366. 11:50.
  367. 11:52While current technology requires this
  368. 11:54much power, new, emerging technologies
  369. 11:57could lead to memory or other
  370. 11:58innovations that don't need as much.
  371. 12:02>> Power efficiency is key.
  372. 12:04>> Since we don't know how much improvement
  373. 12:06will come from those areas, "certainly"
  374. 12:08is a tough word, but under the
  375. 12:10assumption that there are no such
  376. 12:12technological surprises, then yes, it's
  377. 12:14certain.
  378. 12:15Yes.
  379. 12:16>> As I mentioned earlier, Marks is a
  380. 12:18credit investor, isn't he? He must be
  381. 12:20doing a tremendous amount of deal
  382. 12:22sourcing.
  383. 12:23I suspect he made this judgment after
  384. 12:25seeing things on the ground.
  385. 12:26>> The answer lies in the field.
  386. 12:28>> That’s correct. Regarding the bubble,
  387. 12:30will a bubble come in the future? Didn't
  388. 12:32the Fed say something about that
  389. 12:33recently? They claimed that raising
  390. 12:35interest rates had the effect of cooling
  391. 12:37down market overheating, but Marks
  392. 12:39doesn't agree with that.
  393. 12:40>> Aha.
  394. 12:40>> A bubble cannot be prevented by policy
  395. 12:43alone. Because it is a matter of human
  396. 12:44nature, specifically irrational
  397. 12:46exuberance, it cannot be avoided or
  398. 12:48stopped; haven't we seen recent
  399. 12:49phenomena like sharp drops? Sharp
  400. 12:51fluctuations. There must be these kinds
  401. 12:53of corrections to let the heat out and
  402. 12:55allow things to last longer.
  403. 12:57He concluded the interview by saying
  404. 12:59this.
  405. 13:00If I may add a few of my own thoughts.
  406. 13:02>> It means those who need to be shaken out
  407. 13:04should be shaken out.
  408. 13:05>> We were talking about bubbles, right? I
  409. 13:07personally believe that a bubble is born
  410. 13:09only after experiencing a crisis. I
  411. 13:11actually really like looking at old
  412. 13:13articles. I search for them. The period
  413. 13:16being compared to now is the dot-com
  414. 13:17bubble, isn't it?
  415. 13:19This graph is the NASDAQ chart from 1996
  416. 13:23to 1998. The black box is the period
  417. 13:25when the LTCM crisis occurred. At that
  418. 13:28time, when the LTCM crisis broke out, it
  419. 13:30fell 30%from its peak.
  420. 13:32>> It dropped 30%.
  421. 13:33>> It went from 2,000 to 14,000, so it fell
  422. 13:3630%, but for reference, it's almost
  423. 13:3830,000 now, right? Yes. That is correct
  424. 13:41. Hasn't it fallen about 10%from its
  425. 13:43peak now?
  426. 13:43>> 27,000 or something like that.
  427. 13:45>> If it's 30%, based on a rough
  428. 13:46quantitative standard.
  429. 13:47>> It goes down to 22,000. 22,000, or
  430. 13:4921,000. I think that could actually
  431. 13:51happen. Yes.
  432. 13:52>> So, it fell 30%like that, and if you
  433. 13:53look at the articles from that time,
  434. 13:55what was the mood? The financial market
  435. 13:56is ruined.
  436. 13:57>> Aha.
  437. 13:58>> Oh, what are we going to do? Back then,
  438. 13:59the Fed had to urgently summon
  439. 14:01commercial banks to create a bailout
  440. 14:02package. So the consensus was, "Yeah,
  441. 14:04the internet is great, but the stock
  442. 14:06market is broken." It's over. "I don't
  443. 14:08care about anything else anymore."
  444. 14:10That was the prevailing sentiment. "The
  445. 14:12market is doomed."
  446. 14:13>> Mhm.
  447. 14:14>> But what happened after that? This is
  448. 14:16the NASDAQ index leading up to the 2000s
  449. 14:19, specifically the peak in 2000. In the
  450. 14:21year following the LTCM crisis, the
  451. 14:24NASDAQ 100 rose by 102%. The index
  452. 14:27itself. Looking at things like this, we
  453. 14:29usually discuss market cycles in terms
  454. 14:32of market sentiment.
  455. 14:34The stages of that sentiment reflect Sir
  456. 14:37John Templeton’s famous quote: "Bull
  457. 14:39markets are born on pessimism, grow on
  458. 14:42skepticism, mature on optimism, and die
  459. 14:44on euphoria."
  460. 14:47So, the market cycle is determined by
  461. 14:49how much worry disappears.
  462. 14:53Usually, we tend to think that market
  463. 14:55sentiment overheats in a somewhat linear
  464. 14:57fashion, just getting better and better
  465. 14:59.
  466. 15:00But looking at past cases, market
  467. 15:02sentiment overheats when we become
  468. 15:03immune to worry.
  469. 15:04>> Wow.
  470. 15:05>> Even back then, people thought, "We
  471. 15:06survived the LTCM crisis, so there's no
  472. 15:08need to worry about the market breaking
  473. 15:10." They said, "We’re entering a new
  474. 15:11world," and that's how it ended up like
  475. 15:13this.
  476. 15:13>> Wow.
  477. 15:14>> So, you shouldn't be afraid when the
  478. 15:16market is volatile like it is today; the
  479. 15:19real danger is when the market swings
  480. 15:21wildly and people are not afraid.
  481. 15:23>> Uh-huh.
  482. 15:23>> That's when I start thinking that it
  483. 15:25might be a real crisis.
  484. 15:26>> Hmm.
  485. 15:27>> That’s all for today’s content.
  486. 15:29>> Yes.
  487. 15:31Listening to Howard Marks’ insights
  488. 15:33really calms the mind and helps me
  489. 15:35resolve how I should approach my
  490. 15:36investments.
  491. 15:38So, I really enjoyed hearing from Howard
  492. 15:41Marks.

About this transcript

This page contains the full transcript of 보수적인 하워드 막스가 이번엔 AI에 의외의 낙관을 꺼냈습니다... 모른다면서도 돈을 넣은 단 한 곳은 따로 있었습니다 [구루 11월 - 하워드 막스] by 이효석아카데미, generated from the public captions YouTube serves with the video. The transcript has 2,956 words across 492 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.