1 00:00:00,040 --> 00:00:03,720 You know when you first learn to bowl, they put those inflatable 2 00:00:03,720 --> 00:00:06,200 bumpers in the gutters. You roll the ball, it bounces 3 00:00:06,200 --> 00:00:08,920 left, it bounces right, but it like it always hits the pins. 4 00:00:09,000 --> 00:00:10,520 Yeah, it feels amazing. Right. 5 00:00:10,520 --> 00:00:13,400 You feel totally invincible, but you are actually learning how to 6 00:00:13,400 --> 00:00:14,920 bowl. You're just learning how to 7 00:00:14,920 --> 00:00:17,280 throw a heavy object without any consequences. 8 00:00:18,040 --> 00:00:22,960 And the second somebody deflates those bumpers, panic sets in. 9 00:00:23,480 --> 00:00:26,440 Suddenly, the reality of the lane actually matters. 10 00:00:26,520 --> 00:00:30,000 'Cause you haven't developed the underlying skills to navigate 11 00:00:30,000 --> 00:00:32,920 the actual physical environment. I mean, you've just developed 12 00:00:32,920 --> 00:00:36,240 this reliance on an artificial safety net that guarantees A 13 00:00:36,240 --> 00:00:39,080 somewhat positive outcome, you know, no matter how poorly you 14 00:00:39,120 --> 00:00:40,720 actually perform. Exactly. 15 00:00:41,040 --> 00:00:44,600 And that illusion of safety, that tension between relying on 16 00:00:44,600 --> 00:00:47,480 the bumpers and facing the very real unforgiving reality of the 17 00:00:47,480 --> 00:00:50,040 lane, is exactly what we are unpacking today. 18 00:00:50,440 --> 00:00:53,400 But before we reveal the realities of the actual lane, a 19 00:00:53,400 --> 00:00:55,480 quick note for those looking to do their own lane reading. 20 00:00:56,200 --> 00:00:59,280 To access more advanced researches, follow Guru's latest 21 00:00:59,280 --> 00:01:02,320 movement MAD and leverage our latest Deep Values agents. 22 00:01:02,600 --> 00:01:06,520 Welcome to download Deep Values APP from App Store to track 23 00:01:06,520 --> 00:01:09,280 stock market temperature. Welcome to download iOS app 24 00:01:09,280 --> 00:01:12,640 Buffett Indicator from Apple App Store or Bubble Watch from 25 00:01:12,640 --> 00:01:14,200 Google Play. Links to the description 26 00:01:14,200 --> 00:01:17,960 Disclaimer All analysis is based on publicly available sources 27 00:01:18,280 --> 00:01:20,880 and maybe subject to revisions or different interpretations. 28 00:01:21,360 --> 00:01:23,440 All contents from this channel is for research and 29 00:01:23,440 --> 00:01:26,160 entertainment purposes only and does not constitute any 30 00:01:26,160 --> 00:01:28,120 investment or financial advice or service. 31 00:01:28,480 --> 00:01:30,720 Conduct your own due diligence before making investment 32 00:01:30,720 --> 00:01:34,520 financial decisions. OK, so with that other way, 33 00:01:34,520 --> 00:01:37,560 let's look at today's source. Yeah, it's a really fascinating 34 00:01:37,560 --> 00:01:41,000 transcript from Bloomberg Television, specifically an 35 00:01:41,000 --> 00:01:43,400 interview recorded at the Qatar Economic Forum. 36 00:01:43,600 --> 00:01:45,040 Right. And the person sitting in the 37 00:01:45,040 --> 00:01:48,160 hot seat is Howard Marks. Yeah, and for those who track 38 00:01:48,160 --> 00:01:51,000 the really smart minds in finance, he is the Co chair of 39 00:01:51,000 --> 00:01:54,280 Oak Tree Capital Management. He's just a legendary figure, 40 00:01:54,280 --> 00:02:00,280 mostly known for his incredibly thoughtful, meticulously argued, 41 00:02:00,320 --> 00:02:02,280 and often really contrarian memos. 42 00:02:02,280 --> 00:02:03,480 Yeah. And if you're listening to this, 43 00:02:03,480 --> 00:02:05,800 you're here because you are a learner. 44 00:02:06,120 --> 00:02:09,280 You know, you are constantly trying to find the actual signal 45 00:02:09,280 --> 00:02:10,800 and the absolute noise of the market. 46 00:02:10,800 --> 00:02:12,800 You want to figure out how these brilliant minds operate. 47 00:02:12,800 --> 00:02:17,440 But like, without drowning in impenetrable financial jargon. 48 00:02:17,440 --> 00:02:19,160 Which is exactly what we're going to try to do here. 49 00:02:19,160 --> 00:02:20,840 Right. So our mission for this deep 50 00:02:20,840 --> 00:02:24,600 dive is to unpack Mark's genuinely surprising views on 51 00:02:24,600 --> 00:02:27,520 how central banks should operate, why the market is 52 00:02:27,520 --> 00:02:30,480 currently entirely ruled by optimists, and how a traditional 53 00:02:30,480 --> 00:02:33,640 value investor wrestles with the massive, unquantifiable 54 00:02:33,640 --> 00:02:35,440 uncertainty of this whole AI boom. 55 00:02:35,440 --> 00:02:38,400 It really is a master class in how to process uncertainty in 56 00:02:38,400 --> 00:02:40,240 real time. And you know, the conversation 57 00:02:40,240 --> 00:02:43,080 starts with a topic that usually puts casual observers right to 58 00:02:43,080 --> 00:02:45,520 sleep. But Marx completely flips the 59 00:02:45,520 --> 00:02:47,960 conventional wisdom on its head. And we're talking about the 60 00:02:47,960 --> 00:02:50,280 communications strategy of the Federal Reserve, right? 61 00:02:50,280 --> 00:02:53,080 Exactly. The interviewer asks Marx how 62 00:02:53,080 --> 00:02:55,720 he's processing the recent remarks from Fed Chair Jerome 63 00:02:55,720 --> 00:02:57,600 Powell. Because, you know, there's been 64 00:02:57,600 --> 00:03:00,280 this whole narrative lately about the Fed giving very 65 00:03:00,280 --> 00:03:01,480 limited speeches. Right. 66 00:03:01,480 --> 00:03:04,080 Like they aren't giving the market perfectly clear signals 67 00:03:04,080 --> 00:03:05,400 about what they're going to do next. 68 00:03:05,440 --> 00:03:07,080 Yeah. And you would think a 69 00:03:07,080 --> 00:03:11,520 billionaire investor managing massive portfolios wants all the 70 00:03:11,520 --> 00:03:14,160 forward-looking information possible, like every little 71 00:03:14,160 --> 00:03:15,760 crumb of data. Obviously, yeah. 72 00:03:16,360 --> 00:03:19,440 But Mark says less communication is actually better. 73 00:03:20,160 --> 00:03:23,960 He says he does not want the Fed chair to tell him what to do or 74 00:03:23,960 --> 00:03:25,680 even tell him what the Fed is going to do. 75 00:03:25,720 --> 00:03:28,200 Wait, really? Because it seems like the entire 76 00:03:28,200 --> 00:03:30,920 market hangs on every single word Powell says. 77 00:03:30,960 --> 00:03:32,240 Right. But Marks is pointing out a 78 00:03:32,240 --> 00:03:34,880 fundamental flaw and how the modern financial system has 79 00:03:34,880 --> 00:03:37,160 evolved over the last couple of decades. 80 00:03:37,680 --> 00:03:41,320 The system has grown just deeply, deeply reliant on 81 00:03:41,320 --> 00:03:44,080 central bank signaling. Oh right, like the dot plots. 82 00:03:44,280 --> 00:03:47,600 Yes, the interviewer actually brings up the dot plots, which 83 00:03:47,600 --> 00:03:50,600 are those literal charts the Fed publishes showing where each 84 00:03:50,600 --> 00:03:52,920 official thinks interest rates are heading over the next few 85 00:03:52,920 --> 00:03:56,840 years, and marks kind of jokes that he's had to learn way too 86 00:03:56,840 --> 00:04:00,320 much about dot plots lately. His core argument, though? 87 00:04:00,680 --> 00:04:02,760 It's about the systemic health of society. 88 00:04:03,480 --> 00:04:06,600 If the central banker stops handholding the market and tells 89 00:04:06,600 --> 00:04:10,160 people less about what's coming, businesses and investors are 90 00:04:10,160 --> 00:04:12,480 forced to make their portfolios anti fragile. 91 00:04:12,480 --> 00:04:15,640 Anti fragile, meaning you build a system that can withstand 92 00:04:15,640 --> 00:04:19,680 unforeseen shocks rather than a system perfectly optimized for 93 00:04:19,680 --> 00:04:22,120 like 1 specific guaranteed output. 94 00:04:22,600 --> 00:04:26,160 Kind of like driving with AGPS that dictates every single micro 95 00:04:26,160 --> 00:04:29,360 turn, every single lane change like a mile in advance. 96 00:04:29,880 --> 00:04:33,280 If you drive like that for years, you completely lose your 97 00:04:33,280 --> 00:04:36,280 innate sense of direction. 100% you stop looking at the road 98 00:04:36,280 --> 00:04:39,160 signs entirely right? And if that GPS signal suddenly 99 00:04:39,160 --> 00:04:42,760 cuts out, or you know the map is outdated, you have absolutely no 100 00:04:42,760 --> 00:04:44,360 idea how to navigate, you just crash. 101 00:04:44,600 --> 00:04:46,960 And that's exactly what Marx is warning about. 102 00:04:47,240 --> 00:04:50,240 If every investor and business owner thinks they know exactly 103 00:04:50,240 --> 00:04:53,480 what the Fed is going to do, they all calibrate their entire 104 00:04:53,480 --> 00:04:57,200 financial structure for that exact telegraphed outcome. 105 00:04:57,200 --> 00:05:01,040 They build around the bumpers. Yes, the entire society becomes 106 00:05:01,040 --> 00:05:03,320 completely dependent on that single path. 107 00:05:03,720 --> 00:05:07,000 So if an unexpected macroeconomic shock hits. 108 00:05:07,040 --> 00:05:09,400 Like a pandemic. Or geopolitical conflict. 109 00:05:09,400 --> 00:05:12,720 Or supply chain collapse. Yeah, the system just shatters 110 00:05:12,720 --> 00:05:14,800 because no one built in a margin of safety. 111 00:05:14,800 --> 00:05:15,920 They didn't think they needed one. 112 00:05:15,920 --> 00:05:18,600 So Marx wants the Fed to just step back. 113 00:05:18,680 --> 00:05:22,320 Yeah, he advocates for a decidedly less activist central 114 00:05:22,320 --> 00:05:24,000 bank. He wants the economy to just 115 00:05:24,000 --> 00:05:27,720 function on its own unless it's in danger of going completely 116 00:05:27,720 --> 00:05:30,120 off the rails. You know, like hyperinflation or 117 00:05:30,400 --> 00:05:32,600 severe depression. But if the economy is just 118 00:05:32,600 --> 00:05:34,120 humming along, somewhere in the middle. 119 00:05:34,160 --> 00:05:36,960 His advice is to leave it alone. He points directly to this 120 00:05:36,960 --> 00:05:38,840 concept of natural interest rates. 121 00:05:39,120 --> 00:05:40,720 Right. And he defines natural interest 122 00:05:40,720 --> 00:05:43,680 rates as the rates that result purely from negotiations between 123 00:05:43,680 --> 00:05:46,600 borrowers and lenders. Yeah, like actual market 124 00:05:46,600 --> 00:05:48,640 dynamics. We're talking about the free 125 00:05:48,640 --> 00:05:53,000 market pricing the cost of money based on actual risk and reward. 126 00:05:53,920 --> 00:05:56,720 Think about it, if you are starting a highly speculative 127 00:05:56,720 --> 00:06:00,160 tech company, the free market should naturally demand a higher 128 00:06:00,160 --> 00:06:03,520 interest rate from you than from someone taking out, say, a 30 129 00:06:03,520 --> 00:06:05,680 year mortgage on a stable piece of real estate. 130 00:06:05,840 --> 00:06:08,440 Because the risk dictates the cost of the capital. 131 00:06:08,480 --> 00:06:12,120 Exactly when a central bank artificially suppresses interest 132 00:06:12,120 --> 00:06:15,360 rates and broadcasts exactly how long they're going to stay 133 00:06:15,360 --> 00:06:18,480 suppressed, they distort that natural negotiation. 134 00:06:19,080 --> 00:06:21,920 Capital flows into places it probably shouldn't simply 135 00:06:21,920 --> 00:06:24,160 because the money is artificially cheap. 136 00:06:24,320 --> 00:06:26,080 But wait, let me push back on this a little bit. 137 00:06:26,560 --> 00:06:30,080 Isn't a lack of communication from the Fed exactly what causes 138 00:06:30,080 --> 00:06:33,280 massive market panics? I mean, investors hate 139 00:06:33,280 --> 00:06:35,080 uncertainty. Oh, they despise it. 140 00:06:35,240 --> 00:06:36,840 Right. So if the Fed goes completely 141 00:06:36,840 --> 00:06:39,880 silent and stops providing forward guidance, doesn't 142 00:06:39,880 --> 00:06:42,280 everyone just panic and sell off because they don't know what the 143 00:06:42,280 --> 00:06:43,880 cost of capital will be next month? 144 00:06:43,920 --> 00:06:47,080 Investors absolutely crave certainty in the short term, but 145 00:06:47,080 --> 00:06:50,200 Marx is saying that constantly feeding that craving creates a 146 00:06:50,200 --> 00:06:54,400 long term systemic danger. Providing artificial certainty 147 00:06:54,400 --> 00:06:57,600 breeds this very specific and honestly sometimes dangerous 148 00:06:57,600 --> 00:07:01,480 type of market psychology where risk assessment is essentially 149 00:07:01,480 --> 00:07:03,400 just outsourced to the government. 150 00:07:04,040 --> 00:07:07,280 So in central bank stops hand holding the market, investors 151 00:07:07,280 --> 00:07:09,720 are suddenly forced to confront their own psychology. 152 00:07:09,760 --> 00:07:13,240 Yes, and right now, according to Marx, that psychology is 153 00:07:13,240 --> 00:07:16,640 overwhelmingly 1 sided. Yeah, I mean, the interviewer 154 00:07:16,640 --> 00:07:20,120 asks him what this current environment actually feels like 155 00:07:20,120 --> 00:07:23,320 for a value investor. And Marx makes a very distinct 156 00:07:23,320 --> 00:07:25,480 categorization of himself to answer this. 157 00:07:25,480 --> 00:07:29,560 He says he identifies as a value investor but with a small V, not 158 00:07:29,560 --> 00:07:31,880 a capital VA. Small V value investor. 159 00:07:31,880 --> 00:07:34,800 OK, So what does that actually look like in practice is when 160 00:07:34,800 --> 00:07:38,600 someone is analyzing a stock, because usually value investing 161 00:07:38,600 --> 00:07:40,720 sounds pretty rigid. Well, a Capital V's value 162 00:07:40,720 --> 00:07:42,560 investor tends to be heavily dogmatic. 163 00:07:42,600 --> 00:07:45,480 You know, they might have these rigid, unbreakable rules about 164 00:07:45,480 --> 00:07:48,440 price to earnings ratios, or they only buy companies trading 165 00:07:48,440 --> 00:07:51,040 below their book value. Right, old school Buffett stuff. 166 00:07:51,200 --> 00:07:53,320 Yeah. But those strict historical 167 00:07:53,320 --> 00:07:56,760 metrics often lock them completely out of modern asset 168 00:07:56,760 --> 00:08:00,280 like growth companies. A small value investor strips 169 00:08:00,280 --> 00:08:03,680 the discipline down to its core philosophy into what you figure 170 00:08:03,680 --> 00:08:07,120 out the intrusic value of an asset, and you try to buy it at 171 00:08:07,120 --> 00:08:09,160 a fair price relative to that value. 172 00:08:09,160 --> 00:08:13,520 Intrinsic value being the actual measurable worth of the business 173 00:08:13,520 --> 00:08:15,440 based on the future cash it's going to generate. 174 00:08:15,440 --> 00:08:18,520 Right, Basically, yeah. You calculate how much cash a 175 00:08:18,520 --> 00:08:21,920 business will produce over its lifetime, discount that back to 176 00:08:21,920 --> 00:08:26,240 today's dollars and that's your intrinsic value. 177 00:08:26,480 --> 00:08:29,400 If the stock market is offering you the company for a less than 178 00:08:29,400 --> 00:08:32,480 that number, you buy it. But doing that right now is 179 00:08:32,480 --> 00:08:35,440 incredibly difficult, isn't it? I mean, Marks points out that 180 00:08:35,440 --> 00:08:38,919 since roughly October of 2022, the markets have been entirely 181 00:08:38,919 --> 00:08:41,960 ruled by optimists. He knows the S&P 500 has more 182 00:08:41,960 --> 00:08:43,640 than doubled in that specific period. 183 00:08:43,679 --> 00:08:46,800 Yeah, and when optimists dictate the flow of capital, prices 184 00:08:46,800 --> 00:08:49,920 naturally detach from that grounded intrinsic value 185 00:08:49,920 --> 00:08:52,640 population Marx is looking for. People start paying for best 186 00:08:52,640 --> 00:08:55,000 case scenarios rather than current realities. 187 00:08:55,080 --> 00:08:57,840 Which is wild. The interviewer asks a really 188 00:08:57,840 --> 00:09:00,320 crucial question about the source of this optimism. 189 00:09:00,880 --> 00:09:04,440 Because even when diplomatic or economic deals in Washington, 190 00:09:04,440 --> 00:09:07,800 Europe or the Middle East seem completely unlikely to 191 00:09:07,800 --> 00:09:11,520 materialize, the market just responds positively to just the 192 00:09:11,520 --> 00:09:14,560 trace of a rumor. It's just wild optimism, and 193 00:09:14,560 --> 00:09:17,680 Marx lists a few core reasons driving this behavior. 194 00:09:18,000 --> 00:09:21,400 First, to be fair, the underlying truths are actually 195 00:09:21,400 --> 00:09:25,040 pretty solid. the US economy is doing well in the aggregate. 196 00:09:25,240 --> 00:09:27,320 He actually calls it the best performing developed world 197 00:09:27,320 --> 00:09:28,480 economy. OK, that's fair. 198 00:09:28,600 --> 00:09:31,000 What's the second reason? Second, there's the Tina trade. 199 00:09:31,200 --> 00:09:33,280 Tina, there is no alternative. Right. 200 00:09:33,320 --> 00:09:36,040 For years, bond yields were incredibly low. 201 00:09:36,240 --> 00:09:38,720 If you kept your money in a savings account or a safe 202 00:09:38,720 --> 00:09:42,240 government bond, you were actually losing purchasing power 203 00:09:42,240 --> 00:09:44,280 to inflation. So you had to buy stocks? 204 00:09:44,440 --> 00:09:47,160 Investors were essentially forced to put their capital into 205 00:09:47,160 --> 00:09:50,080 the stock market to get any sort of real return, and that in 206 00:09:50,080 --> 00:09:53,240 green habit is hard to break even as interest rates have 207 00:09:53,240 --> 00:09:55,280 shifted recently. OK, so that's two. 208 00:09:55,640 --> 00:10:00,200 Third is obviously the massive excitement over AI, which we 209 00:10:00,200 --> 00:10:03,320 will dig into shortly. But the fourth reason is the 210 00:10:03,320 --> 00:10:05,960 most fascinating to me from a psychological standpoint. 211 00:10:06,040 --> 00:10:08,880 Yeah, the time frame. Marks points out that we haven't 212 00:10:08,880 --> 00:10:12,880 had a truly devastating, prolonged bad time since March 213 00:10:12,880 --> 00:10:16,560 of 2009, when the stock market bottomed out during the global 214 00:10:16,560 --> 00:10:18,960 financial crisis. It's incredible to think about 215 00:10:19,240 --> 00:10:21,720 an entire generation of investors and financial 216 00:10:21,720 --> 00:10:24,680 professionals has has essentially never managed money 217 00:10:24,680 --> 00:10:27,600 through a catastrophic multi year bear market. 218 00:10:27,800 --> 00:10:29,760 They don't know what it feels like when the bumpers are gone. 219 00:10:29,760 --> 00:10:32,040 Exactly. People get used to good times, 220 00:10:32,120 --> 00:10:35,080 and human nature naturally extrapolates those good times 221 00:10:35,080 --> 00:10:38,040 into the future. Optimists, as Marx observes, 222 00:10:38,080 --> 00:10:41,600 they just Slough off bad news. They get thrilled by the rumor 223 00:10:41,600 --> 00:10:44,960 of a settlement, but they do not adjust their valuations downward 224 00:10:44,960 --> 00:10:46,880 when that deal inevitably falls through. 225 00:10:47,080 --> 00:10:50,960 But wait, if optimists are just blindly ignoring bad news, 226 00:10:51,120 --> 00:10:54,720 shrugging off failed deals, and extrapolating good times 227 00:10:54,720 --> 00:10:57,800 indefinitely, aren't they just building a massive House of 228 00:10:57,800 --> 00:10:59,800 Cards? Like buying stocks at 229 00:10:59,800 --> 00:11:03,040 historically high prices relative to their intrinsic 230 00:11:03,040 --> 00:11:06,320 value? Sounds like the exact scenario 231 00:11:06,320 --> 00:11:09,120 where a value investor should just liquidate everything and 232 00:11:09,120 --> 00:11:11,920 hold cash. You'd think so, but Mark's take 233 00:11:11,920 --> 00:11:15,440 away is remarkably pragmatic. He really avoids extreme 234 00:11:15,440 --> 00:11:17,680 reactions. He doesn't say sell everything. 235 00:11:17,680 --> 00:11:20,080 No, he doesn't advocate panic selling at all. 236 00:11:20,280 --> 00:11:23,160 He says you have to clearly acknowledge the positive bias in 237 00:11:23,160 --> 00:11:24,960 the room. You have to recognize who was 238 00:11:24,960 --> 00:11:27,680 setting the prices. When optimists are in control, 239 00:11:27,680 --> 00:11:29,920 prices are high. That doesn't mean a crash is 240 00:11:29,920 --> 00:11:32,840 imminent tomorrow. Markets can remain elevated for 241 00:11:32,840 --> 00:11:34,840 years. It's simply means your margin 242 00:11:34,840 --> 00:11:37,040 for error is smaller, so it's time for caution. 243 00:11:37,040 --> 00:11:39,480 You build in a little defense. Exactly. 244 00:11:39,480 --> 00:11:42,320 You don't leave the game entirely, but you start favoring 245 00:11:42,320 --> 00:11:44,280 companies with stronger balance sheets. 246 00:11:44,520 --> 00:11:47,960 You take a lot more care in how you allocate new capital because 247 00:11:47,960 --> 00:11:51,920 the forces driving this current wave of optimism carry risks 248 00:11:51,920 --> 00:11:55,280 that are genuinely unprecedented in modern financial history. 249 00:11:55,400 --> 00:11:59,040 And that need for caution is exactly why Marx's view on AI is 250 00:11:59,040 --> 00:12:01,400 so surprising to me. I mean, you think the ultimate 251 00:12:01,400 --> 00:12:05,160 market optimist engine, the promise of artificial general 252 00:12:05,160 --> 00:12:08,080 intelligence reshaping everything, would absolutely 253 00:12:08,080 --> 00:12:10,200 terrify A disciplined value investor? 254 00:12:10,400 --> 00:12:13,080 It does, but his concerns are focused completely outside of 255 00:12:13,080 --> 00:12:15,360 the stock market. In previous memos, Marx has 256 00:12:15,360 --> 00:12:18,680 expressed acute concern about AI wiping out jobs, particularly in 257 00:12:18,680 --> 00:12:20,840 middle America. Right, so the interviewer asked 258 00:12:20,840 --> 00:12:22,560 him. Given all the hype and all the 259 00:12:22,560 --> 00:12:25,920 recent existential dread we hear about, AI have his fears 260 00:12:25,920 --> 00:12:27,960 compounded. And this is where Mark shares a 261 00:12:27,960 --> 00:12:30,920 quote from his wife that frames his entire approach to this 262 00:12:30,920 --> 00:12:35,440 macro level anxiety. She defines worrying as bleeding 263 00:12:35,440 --> 00:12:37,960 before you're shot. Bleeding before you're shot. 264 00:12:38,200 --> 00:12:41,880 Wow, yeah. A sentient human being observing 265 00:12:41,880 --> 00:12:44,640 the landscape obviously has to think about these massive 266 00:12:44,640 --> 00:12:47,360 technological shifts, but you also have to acknowledge when 267 00:12:47,360 --> 00:12:50,200 there is absolutely nothing you can personally do to alter the 268 00:12:50,200 --> 00:12:52,120 trajectory. Right, He can't put the AI 269 00:12:52,120 --> 00:12:54,760 technology back in the box. He can't stop companies from 270 00:12:54,760 --> 00:12:56,280 adopting it, so he rolls with it. 271 00:12:56,520 --> 00:12:58,840 Yeah, he actively prevents it from becoming a daily 272 00:12:58,840 --> 00:13:01,000 preoccupation that clouds his judgment. 273 00:13:01,000 --> 00:13:04,160 But that personal Stoic acceptance doesn't mean he 274 00:13:04,160 --> 00:13:07,280 thinks society as a whole should just ignore the implications, 275 00:13:07,280 --> 00:13:08,440 right? Not at all. 276 00:13:08,760 --> 00:13:11,880 He points his finger directly at the lack of structural 277 00:13:11,880 --> 00:13:15,160 governmental action. People are endlessly debating 278 00:13:15,160 --> 00:13:19,160 philosophical existential threats like killer robots, but 279 00:13:19,160 --> 00:13:21,560 he doesn't see any concrete policy coming out of the 280 00:13:21,560 --> 00:13:23,840 government to manage the economic transition. 281 00:13:23,920 --> 00:13:26,640 Right, He argues there needs to be a dedicated government task 282 00:13:26,640 --> 00:13:29,840 force to address the coming societal effects of AI. 283 00:13:30,120 --> 00:13:33,560 Yes, and he poses A purely practical balance sheet question 284 00:13:33,560 --> 00:13:36,400 that completely bypasses the usual sci-fi debates. 285 00:13:36,760 --> 00:13:41,320 He asks if AI puts millions of people out of work, who pays the 286 00:13:41,320 --> 00:13:42,600 taxes? Oh wow. 287 00:13:42,880 --> 00:13:46,480 If a software program takes the place of a human worker, how 288 00:13:46,480 --> 00:13:48,160 does it take their place as a taxpayer? 289 00:13:48,560 --> 00:13:51,800 What happens to municipal, state and federal government revenues 290 00:13:51,800 --> 00:13:55,600 when payroll taxes just vanish? It is a devastatingly simple 291 00:13:55,600 --> 00:14:00,200 question with massive macroeconomic implications, like 292 00:14:00,200 --> 00:14:02,600 we hear constantly about universal basic income as a 293 00:14:02,600 --> 00:14:05,960 solution for displaced workers. But Marx is looking at the other 294 00:14:05,960 --> 00:14:09,040 side of the Ledger. Exactly how do you fund any 295 00:14:09,040 --> 00:14:12,920 societal safety net if the primary revenue engine human 296 00:14:12,920 --> 00:14:16,360 income tax just evaporates? It is like a corporate board 297 00:14:16,360 --> 00:14:19,440 deciding to RIP out the heavy gas engine of a delivery truck 298 00:14:19,680 --> 00:14:22,600 to save on weight and improve efficiency, completely 299 00:14:22,600 --> 00:14:24,880 forgetting that the engines alternator is what powers the 300 00:14:24,880 --> 00:14:27,720 headlights, the dashboard and the refrigeration unit in the 301 00:14:27,720 --> 00:14:29,080 back. That is a perfect. 302 00:14:29,080 --> 00:14:32,960 Analogy If citizens lose jobs on mass, the government's revenue 303 00:14:32,960 --> 00:14:35,400 engine collapses. You can optimize corporate 304 00:14:35,400 --> 00:14:38,640 efficiency all you want with AI, but if you break the underlying 305 00:14:38,640 --> 00:14:41,400 funding mechanism of the society those corporations operate 306 00:14:41,400 --> 00:14:44,680 within, the efficiency gains literally don't matter. 307 00:14:44,960 --> 00:14:47,440 Right. It's a macro level societal risk 308 00:14:47,440 --> 00:14:49,680 that is entirely unquantifiable right now. 309 00:14:49,680 --> 00:14:52,440 Nobody knows how fast job displacement will actually 310 00:14:52,440 --> 00:14:55,200 happen, or if new industries will replace the old ones 311 00:14:55,400 --> 00:14:57,320 quickly enough to maintain that tax base. 312 00:14:57,560 --> 00:15:01,360 And this sober kind of pragmatic view of the macro risks 313 00:15:01,760 --> 00:15:06,160 translates directly into how Marx is utilizing AI inside his 314 00:15:06,160 --> 00:15:08,960 own billion dollar firm right now, because the interviewer 315 00:15:08,960 --> 00:15:12,960 asks if he's out there actively using ChatGPT and if Oak Tree 316 00:15:12,960 --> 00:15:15,280 has integrated AI tools into their workflows. 317 00:15:15,480 --> 00:15:17,360 And they have. Oak Tree has absolutely 318 00:15:17,360 --> 00:15:20,240 integrated AI to Marshall massive amounts of data and 319 00:15:20,240 --> 00:15:23,840 perform early analysis. The technology is phenomenal at 320 00:15:23,840 --> 00:15:27,240 taking tedious manual tasks that used to consume days or weeks 321 00:15:27,240 --> 00:15:28,760 and compressing them into minutes. 322 00:15:28,920 --> 00:15:31,320 It is an undeniable massive time saver. 323 00:15:31,320 --> 00:15:33,960 Like, instead of a junior analyst spending three grueling 324 00:15:33,960 --> 00:15:37,000 weeks reading 400 pages of obscure corporate debt covenants 325 00:15:37,080 --> 00:15:40,600 just to find a specific legal loophole, the AI model flags 326 00:15:40,600 --> 00:15:42,840 that exact loophole in three seconds. 3 seconds. 327 00:15:43,000 --> 00:15:45,560 Yeah, that is the prep work, Yeah, But Marx is adamant that 328 00:15:45,560 --> 00:15:48,240 they absolutely do not turn actual business or investment 329 00:15:48,240 --> 00:15:50,080 decisions over to the AI. Right. 330 00:15:50,120 --> 00:15:52,600 And the reasoning gets to the absolute core of what Marx 331 00:15:52,600 --> 00:15:56,080 believes makes a great investor. He argues that exceptional 332 00:15:56,080 --> 00:15:59,120 investing doesn't come from a computer or an algorithm, simply 333 00:15:59,120 --> 00:16:01,080 having access to more information. 334 00:16:01,120 --> 00:16:04,400 OK, then what does it come from? It comes from a human being 335 00:16:04,400 --> 00:16:07,640 doing a superior job of assessing the import of that 336 00:16:07,640 --> 00:16:10,240 information. The import, like the weight, the 337 00:16:10,240 --> 00:16:13,320 subtle meaning, the broader context of the data. 338 00:16:13,400 --> 00:16:15,880 Exactly. The AI can gather the data 339 00:16:15,880 --> 00:16:17,680 perfectly. It can summarize the debt 340 00:16:17,680 --> 00:16:21,200 covenants, Florida State. But judging what that loophole 341 00:16:21,200 --> 00:16:24,680 actually means in the messy, irrational real world is a 342 00:16:24,680 --> 00:16:28,800 uniquely human skill, right? A human analyst has to decide if 343 00:16:28,800 --> 00:16:31,480 exploiting that loophole will ruin a vital long term 344 00:16:31,480 --> 00:16:33,320 relationship with a corporate partner. 345 00:16:33,320 --> 00:16:36,480 The AI doesn't understand reputation or human psychology. 346 00:16:36,480 --> 00:16:39,800 Marx explicitly states that this level of nuanced superiority is 347 00:16:39,800 --> 00:16:41,480 just not available through AI. And. 348 00:16:41,680 --> 00:16:44,480 The interviewer asks a really insightful follow up to this. 349 00:16:45,000 --> 00:16:47,760 When Oak Tree hires younger people, young analysts who grew 350 00:16:47,760 --> 00:16:50,640 up digital natives, are they seeing them outsource their 351 00:16:50,640 --> 00:16:52,520 critical thinking to these AI tools? 352 00:16:52,720 --> 00:16:55,680 Which is a huge concern, Mark says. 353 00:16:55,680 --> 00:16:58,120 Oak Tree actively discourages that reliance. 354 00:16:58,520 --> 00:17:01,480 He warns that if you stop doing the hard, slow work of critical 355 00:17:01,480 --> 00:17:04,359 thinking, you will inevitably lose the muscle. 356 00:17:04,640 --> 00:17:06,920 The analogy of a kitchen works perfectly here, I think. 357 00:17:07,319 --> 00:17:09,200 AI is like the ultimate sous chef. 358 00:17:09,440 --> 00:17:12,720 It can dice 10,000 onions perfectly and instantly, measure 359 00:17:12,720 --> 00:17:16,119 every complex spice blend to the microgram, and prep all the 360 00:17:16,119 --> 00:17:19,119 ingredients faster than any human brigade ever could. 361 00:17:19,119 --> 00:17:22,359 But wait, if the AI sous chef is that flawless? 362 00:17:22,520 --> 00:17:26,200 If it processes a million times more data and makes 0 363 00:17:26,200 --> 00:17:29,200 mathematical errors, why wouldn't we just trust it to 364 00:17:29,200 --> 00:17:31,920 cook the final meal? Because the sous chef doesn't 365 00:17:31,920 --> 00:17:34,880 know who's sitting in the dining room, The sous chef doesn't know 366 00:17:34,880 --> 00:17:37,840 that it's an unusually hot day outside, so the soup needs to be 367 00:17:37,840 --> 00:17:40,480 balanced with a little more acidity to feel refreshing to 368 00:17:40,480 --> 00:17:42,480 the guests. Context. 369 00:17:42,480 --> 00:17:45,000 Right context. You still require the human head 370 00:17:45,000 --> 00:17:48,720 chef to taste the broth, assess the broader atmosphere, context, 371 00:17:48,960 --> 00:17:52,000 and know what the dish actually needs to succeed in reality. 372 00:17:52,120 --> 00:17:55,680 And Marx believes AI will do to active investing exactly what 373 00:17:55,680 --> 00:17:59,160 index funds did decades ago. It will put average, 374 00:17:59,200 --> 00:18:01,560 unexceptional active investors out of business. 375 00:18:02,240 --> 00:18:04,480 But it won't replace the exceptional human ones. 376 00:18:04,680 --> 00:18:08,800 The true alpha, the actual edge, remains human judgement. 377 00:18:08,840 --> 00:18:11,800 But applying that human judgement to the current AI 378 00:18:11,800 --> 00:18:14,440 landscape is where things get really complicated. 379 00:18:14,880 --> 00:18:18,880 How does a human mind evaluate the pure chaos and intense hype 380 00:18:19,280 --> 00:18:22,040 of AI companies trading on the stock market today? 381 00:18:22,080 --> 00:18:23,400 Oh. This prompts the interviewer to 382 00:18:23,400 --> 00:18:25,560 ask the bubble question. The bubble question. 383 00:18:25,560 --> 00:18:28,480 Yeah, the interviewer references Alan Greenspan's incredibly 384 00:18:28,480 --> 00:18:32,000 famous phrase from 1997, Irrational exuberance. 385 00:18:32,360 --> 00:18:35,280 They ask Marx how publicious this current AI driven market 386 00:18:35,280 --> 00:18:38,120 environment actually looks to him through his small value 387 00:18:38,120 --> 00:18:39,760 lens. And Marx agrees that there is 388 00:18:39,760 --> 00:18:42,280 definitely exuberance, right? The optimism is visible in the 389 00:18:42,280 --> 00:18:45,800 evaluations everywhere, but he hesitates very specifically to 390 00:18:45,800 --> 00:18:48,960 call it irrational. Which is a striking, deeply 391 00:18:48,960 --> 00:18:51,480 analytical distinction for a traditional value investor to 392 00:18:51,480 --> 00:18:54,960 make, because usually high prices driven by overwhelming 393 00:18:54,960 --> 00:18:58,720 optimism are immediately labeled irrational by value purists. 394 00:18:58,800 --> 00:19:00,440 So why doesn't he call it irrational? 395 00:19:00,800 --> 00:19:04,600 Marx explains his logic. To definitively call a market 396 00:19:04,600 --> 00:19:08,360 behavior irrational, you have to know with certainty what the 397 00:19:08,360 --> 00:19:11,720 rational baseline is. And right now, absolutely nobody 398 00:19:11,720 --> 00:19:14,280 can spec out exactly what artificial intelligence is going 399 00:19:14,280 --> 00:19:17,440 to be doing in five years. No one knows what specific 400 00:19:17,440 --> 00:19:20,680 transformative effect it will have on global productivity, or 401 00:19:20,680 --> 00:19:24,640 how profitable these underlying models will actually be, because 402 00:19:24,640 --> 00:19:27,600 that future is completely unwritten and technologically 403 00:19:27,600 --> 00:19:30,680 unprecedented. You cannot definitively state 404 00:19:30,680 --> 00:19:32,640 that the current excitement is irrational. 405 00:19:32,720 --> 00:19:36,560 But he draws a really sharp practical contrast between 406 00:19:36,560 --> 00:19:39,120 different types of technology companies currently involved in 407 00:19:39,120 --> 00:19:41,280 this boom. He points to the hyperscalers. 408 00:19:41,320 --> 00:19:44,560 Right, massive companies like Microsoft, Amazon and Google, 409 00:19:44,680 --> 00:19:47,640 Yes, they are investing 10s of billions of dollars into AI 410 00:19:47,640 --> 00:19:51,280 infrastructure, but they also have massive, existing, highly 411 00:19:51,280 --> 00:19:54,920 robust legacy businesses. Like global retail logistics or 412 00:19:54,920 --> 00:19:57,880 enterprise cloud hosting. Things that actively support 413 00:19:57,880 --> 00:19:59,440 their current earnings estimates. 414 00:19:59,600 --> 00:20:03,080 Exactly, you can actually build a financial model around their 415 00:20:03,080 --> 00:20:06,680 existing cash flow. But then he points to pure AI 416 00:20:06,680 --> 00:20:09,440 companies. Companies whose entire existence 417 00:20:09,440 --> 00:20:13,480 and valuation are dependent solely on AI technology, many of 418 00:20:13,480 --> 00:20:16,840 which have never produced a single dollar of actual 419 00:20:16,920 --> 00:20:19,000 operating profit. So for anyone trying to 420 00:20:19,000 --> 00:20:22,960 calculate intrinsic value, a pure AI company presents an 421 00:20:22,960 --> 00:20:26,120 almost insurmountable degree of fundamental uncertainty. 422 00:20:26,720 --> 00:20:30,200 You are trying to discount future cash flows, but you can't 423 00:20:30,200 --> 00:20:32,880 even quantify what the basic revenue model looks like. 424 00:20:33,280 --> 00:20:36,080 Yeah, Marks brings up Warren Buffett's famous strategy for 425 00:20:36,080 --> 00:20:38,680 dealing with exactly this kind of scenario, putting it in the 426 00:20:38,680 --> 00:20:41,520 too hard pile. It is practically impossible for 427 00:20:41,520 --> 00:20:44,360 a small value investor to confidently slap a fair 428 00:20:44,360 --> 00:20:46,640 valuation on a pure AI company today. 429 00:20:46,680 --> 00:20:49,280 But hold on, if we have absolutely no idea what a pure 430 00:20:49,280 --> 00:20:51,840 AI company will generate in revenue in five years if we 431 00:20:51,840 --> 00:20:54,800 can't quantify the future earnings at all, isn't throwing 432 00:20:54,800 --> 00:20:58,000 billions of investment dollars out of the literal definition of 433 00:20:58,000 --> 00:21:00,720 irrational behavior? It sounds exactly like venture 434 00:21:00,720 --> 00:21:03,680 capital gambling, not discipline public market investing. 435 00:21:03,840 --> 00:21:07,120 It does sound like that, but Marx addresses this exact 436 00:21:07,120 --> 00:21:09,840 tension with a piece of counterintuitive logic that 437 00:21:09,840 --> 00:21:12,200 forms the bedrock of his risk philosophy. 438 00:21:12,960 --> 00:21:15,440 He essentially argues that just because you do not know the 439 00:21:15,440 --> 00:21:18,560 exact outcome, that does not mean the underlying premise is 440 00:21:18,560 --> 00:21:21,640 untrue or silly. OK, the fact that there is an 441 00:21:21,640 --> 00:21:24,920 unusual, historically high amount of uncertainty present 442 00:21:24,920 --> 00:21:27,920 doesn't mean you just freeze up, sell everything and do nothing. 443 00:21:27,920 --> 00:21:31,240 You don't just ignore a massive technological shift simply 444 00:21:31,240 --> 00:21:34,400 because the math is hard. The presence of immense, 445 00:21:34,440 --> 00:21:37,440 unquantifiable uncertainty. You shouldn't paralyze you. 446 00:21:37,760 --> 00:21:40,200 It should simply inform how you structure your investments. 447 00:21:40,200 --> 00:21:43,560 It informs which specific form of AI exposure you choose to 448 00:21:43,560 --> 00:21:45,280 take. Like leaning heavily toward the 449 00:21:45,280 --> 00:21:48,320 established hyperscalers that offer a downside floor instead 450 00:21:48,320 --> 00:21:51,440 of, you know, rolling the dice on the pureplay startups. 451 00:21:51,680 --> 00:21:54,600 Exactly. And more importantly, it informs 452 00:21:54,600 --> 00:21:57,960 exactly how much of your total portfolio you allocate to that 453 00:21:57,960 --> 00:21:59,720 theme. You acknowledge the heavy 454 00:21:59,720 --> 00:22:03,000 uncertainty, you size your financial bets accordingly, and 455 00:22:03,080 --> 00:22:05,640 when you actively build in that defensive posture we talked 456 00:22:05,640 --> 00:22:08,680 about earlier, you don't walk away from the table out of fear, 457 00:22:09,080 --> 00:22:12,120 but you carefully manage your exposure to a hand you cannot 458 00:22:12,120 --> 00:22:14,720 fully see. Man, pulling all of these 459 00:22:14,720 --> 00:22:17,320 threads together to view the entire landscape we've just 460 00:22:17,320 --> 00:22:19,600 covered. What does this all mean for you, 461 00:22:20,120 --> 00:22:22,120 the learner? Howard marks his entire 462 00:22:22,120 --> 00:22:24,440 worldview, whether he's analyzing the silent strategy of 463 00:22:24,440 --> 00:22:27,000 the Federal Reserve, the relentless optimist driving 464 00:22:27,000 --> 00:22:31,240 market valuations, or the unquantifiable societal threat 465 00:22:31,240 --> 00:22:34,840 of artificial intelligence, is fundamentally about embracing 466 00:22:34,840 --> 00:22:37,600 uncertainty without losing your analytical discipline. 467 00:22:38,120 --> 00:22:40,840 The goal of investing, and perhaps navigating the modern 468 00:22:40,840 --> 00:22:43,520 economy isn't to perfectly predict a future that is 469 00:22:43,520 --> 00:22:46,760 inherently unpredictable. The goal is to build a robust 470 00:22:46,760 --> 00:22:50,040 anti fragile system for yourself that can survive multiple 471 00:22:50,040 --> 00:22:53,240 different futures. And you can actively apply that 472 00:22:53,240 --> 00:22:57,200 small V value mindset to so much more than just a financial 473 00:22:57,200 --> 00:22:59,520 portfolio. You can apply it to how you 474 00:22:59,520 --> 00:23:02,840 consume Daily News or how you make long term career decisions 475 00:23:02,840 --> 00:23:04,920 in an increasingly AI heavy world. 476 00:23:05,520 --> 00:23:08,160 You calmly acknowledge the positive bias in the room. 477 00:23:08,440 --> 00:23:11,240 You accept that you cannot control macro forces like 478 00:23:11,240 --> 00:23:14,280 central bank policy or technological evolution. 479 00:23:14,280 --> 00:23:16,320 You don't waste energy bleeding before you're shot. 480 00:23:16,520 --> 00:23:19,040 Exactly. But you always build in your own 481 00:23:19,040 --> 00:23:22,400 margin of safety. And crucially, as Marx warns, 482 00:23:22,880 --> 00:23:26,000 you never let the algorithms do your critical thinking for you, 483 00:23:26,400 --> 00:23:29,440 because the moment you stop actively assessing context, you 484 00:23:29,440 --> 00:23:31,400 lose the only real advantage you have. 485 00:23:31,440 --> 00:23:34,840 If the AI models can eventually Marshall all the world's data 486 00:23:34,840 --> 00:23:37,440 and perform the basic fundamental analysis instantly, 487 00:23:37,440 --> 00:23:40,800 and Florida State, the only genuine edge left in the market 488 00:23:40,840 --> 00:23:43,720 and in the workplace is the human capacity to assess subtle 489 00:23:43,720 --> 00:23:46,080 context. To weigh the nuanced meaning of 490 00:23:46,080 --> 00:23:48,720 that data against the messy reality of human behavior. 491 00:23:48,840 --> 00:23:51,720 Which leaves us with something really fascinating to chew on as 492 00:23:51,720 --> 00:23:54,320 we wrap up. If the ultimate investing edge, 493 00:23:54,360 --> 00:23:56,960 and quite possibly the ultimate professional edge of the next 494 00:23:56,960 --> 00:24:01,520 decade is exceptional human critical thinking and artificial 495 00:24:01,520 --> 00:24:04,240 intelligence is rapidly automating all the basic 496 00:24:04,240 --> 00:24:07,680 analysis and data gathering. Is the most valuable asset of 497 00:24:07,680 --> 00:24:10,200 the future simply the human ability to ask the right 498 00:24:10,240 --> 00:24:13,560 unprompted questions? And if so, how are you actively 499 00:24:13,560 --> 00:24:16,440 training that muscle today? Are you practicing your swing on 500 00:24:16,440 --> 00:24:18,880 the real lane, or are you still relying on the bumpers? 501 00:24:19,080 --> 00:24:21,040 Thanks for joining us on this deep dive into the mind of 502 00:24:21,040 --> 00:24:22,960 Howard Marks. Keep exploring, keep 503 00:24:22,960 --> 00:24:25,360 questioning, and we'll catch you on the next deep dive.