May 4, 2026 · 1h 43m · wtf
Howard Marks: AI, Debt vs Equity & The Next 40 Years Of Investing | Nikhil Kamath | People by WTF · Nikhil Kamath
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In an in-depth conversation with host Nikhil Kamath, veteran investor Howard Marks explores the mechanics of market cycles, credit investing, and risk management while examining the cognitive frameworks needed to achieve superior returns alongside emerging artificial intelligence.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Nikhil holds 12.9% of the talking time here. How this is scored →
Nikhil as informed peer 4.6 Guest teaching 4.3 Guest disagreement 1.7 Nikhil pushing back 2.4
speaking balance: gold is Nikhil, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 54:05 Marks firmly rejects Nikhil's default break-even calculation
Marks swiftly cuts off Nikhil's assumption that a 10% yield allows a 10% default rate, explaining that a 10% principal loss with 10% interest results in zero return rather than profitability.
Hardest push from Nikhil ▶ 12:25 Nikhil challenges Marks on the coin toss vs preparation contradictionNikhil directly confronts Marks's coin flip analogy, pressing him on how one can prepare for the future using historical patterns if preceding events are truly independent.
Biggest teaching moment ▶ 24:19 Marks corrects Nikhil's fundamental definition of a recessionMarks corrects Nikhil's premise about recessions being excess production relative to organic consumption, clarifying that economic contractions are fundamentally driven by demand reduction.
Nikhil holds their own ▶ 30:12 Nikhil identifies AI compute and energy infrastructure as modern inventory build-upWhen Marks claims there are no classic inventory or construction excesses visible in the economy, Nikhil pushes back by identifying AI compute hardware, data center construction, and power infrastructure as the current cycle's boom area.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Nikhil as informed peer | Guest teaching | Guest disagreement | Nikhil pushing back | Why |
|---|---|---|---|---|---|---|
| Episode Preview and Highlights | 2 | 1 | 1 | 0 | Nikhil sets the stage for the episode, outlining the target audience of young entrepreneurs before inviting Marks to share his origin story from Queens to Wharton and Citibank. | |
| Japanese Philosophy, Mujo, and Independent Probabilities | 4 | 4 | 2 | 2 | Nikhil introduces Japanese philosophy and cultural attitudes toward mean reversion. Marks clarifies the concept of mujo and emphasizes the mathematical independence of events using a coin toss analogy. | |
| Preparing for the Unpredictable vs. Making Forecasts | 3 | 5 | 1 | 1 | Nikhil inquires about game theory, prompting Marks to elaborate on why forecasts are unreliable and how investors should instead prepare by suboptimizing across a range of plausible middle outcomes. | |
| Human Psychology as the Engine of Market Patterns | 6 | 5 | 2 | 5 | Nikhil sharply challenges Marks, pointing out a tension between the independent coin toss analogy and using past history to prepare. Marks clarifies that while physical events are independent, human psychological reactions create semi-predictable cycles. | |
| The Mechanics of Market Cycles: Excesses and Corrections | 5 | 4 | 2 | 3 | Marks outlines how cycles stem from psychological excesses and corrections around trendlines. Nikhil pushes on the S&P 500 benchmark by highlighting that index constituents continuously change over time. | |
| Impact of Artificial Intelligence on Market Volatility | 5 | 6 | 3 | 4 | Nikhil explores AI-driven markets and questions whether GDP is an appropriate metric for recessions when consumption drops artificially. Marks directly corrects Nikhil's definition of recession, explaining that drop in demand drives economic contractions. | |
| Current Economic Health and AI Infrastructure Booms | 6 | 3 | 1 | 4 | When Marks notes a lack of classic cyclical excesses like building cranes or inventory surges, Nikhil counters by identifying AI data centers, energy, and compute as the contemporary equivalent of an inventory boom. | |
| Staying Mentally Sharp and Engaging with AI | 4 | 3 | 1 | 2 | Nikhil asks Marks how he stays sharp at 80, leading to a discussion on genetics, learning from Claude, and whether baseline ambition is fundamentally driven by financial reward or staying relevant. | |
| Drifting Down the River vs. Intentional Decision-Making | 3 | 4 | 1 | 1 | Nikhil probes whether major life transitions stem from conscious reflection or luck. Marks reflects candidly on spending 35 years drifting down the river before co-founding Oaktree with proactive intention. | |
| Overcoming Inertia and the Spark of Entrepreneurship | 4 | 3 | 1 | 2 | Nikhil asks how one transitions from drifting to taking active control. Marks explains that co-founding Oaktree was driven by external encouragement from his wife and partner rather than an innate entrepreneurial impulse. | |
| Why Debt Over Equity: The Rise of High-Yield Bonds | 5 | 6 | 3 | 4 | Nikhil questions why an investor would prefer debt over exponential equity upside. Marks recounts the Nifty 50 collapse and corrects Nikhil's calculation regarding break-even default rates in high-yield portfolios. | |
| Oaktree's Credit Analysis Framework and Analytical Edge | 5 | 5 | 1 | 3 | Nikhil explores Oaktree's edge in calculating credit spreads and asks about re-branding distressed debt as opportunistic. Marks details Oaktree's bottom-up, disciplined eight-point investigative process. | |
| The Next 40 Years: AI, Human Edge, and Indexation | 5 | 5 | 2 | 3 | Nikhil asks where the front of the line is for the next 40 years, prompting Marks to argue that understanding AI is critical. Marks also demonstrates that indexation succeeded primarily because active managers performed poorly. | |
| Second-Level Thinking, Market Psychology, and Charlie Munger's Wisdom | 5 | 5 | 2 | 3 | Nikhil asks if indexing succeeded because of a continuous bull market. Marks explains that market panics theoretically open opportunities for active investors, but widespread psychological paralysis prevents them from capitalizing. | |
| Nikhil's Investment Thesis on India and Marks's Perspective | 6 | 4 | 2 | 1 | Nikhil presents his long-term investment thesis on India's growth. Marks declines to speculate on Indian sectors and shares an insight from his son Andrew that readily available quantitative data offers zero analytical edge. | |
| Humility, Adaptability, and the Essence of Second-Level Thinking | 4 | 5 | 1 | 1 | Nikhil compliments Marks on his intellectual flexibility and asks how investors can cultivate adaptability. Marks asserts that second-level thinking is essential but cannot be taught, likening it to coaching height in basketball. | |
| Can AI Execute Contrarian Thinking and Generate Alpha? | 6 | 5 | 2 | 3 | Nikhil suggests programming AI models to be contrarian or contradictory to consensus. Marks points out that simply deviating from consensus without superior insight guarantees being wrong rather than successful. | |
| History Rhyming, Human Nature, and Market Crashes | 5 | 4 | 2 | 2 | Nikhil quotes Napoleon on history being written by winners, while Marks argues financial figures provide objective anchors. Marks concludes by framing investing as a probabilistic puzzle where one must accept being wrong occasionally. |
Statements from this episode (32)
Marks: Prepare portfolios for multiple middle-probability outcomes rather than maximizing single scenarios
“In investing, you can put together a portfolio rather than one that will produce maximum success. If a certain future unfolds, you can put together a portfolio Which will do fine if several, any of several futures unfold, and not too terribly if a bunch of oth…”
Howard Marks: Market history and human psychology have semi-predictable future implications
“History, the things that have happened so far and people's reaction to history have Semi-predictable implications for the future.”
Howard Marks: Ability to infer market behavior rises after extreme events
“The things that have happened, for example, in the economy and the markets are likely to induce certain behavior in the future, and you can know a little bit about that. And, but when the history and the past behavior has been extreme, the ability to infer fro…”
Marks: Market cycles are behavioral excesses and corrections, not just ups and downs
“Rather than thinking of cycles as ups and downs, which I think most people do, think of them as excesses and corrections.”
Marks: S&P 500 returned 20% annually in the 1990s and zero in the 2000s
“In the nineties, it returned 20% a year, and in the aughts, it returned zero per year.”
Marks: Annual S&P 500 returns are almost never between 8% and 12%
“The return on the S&P, which averages 10, is almost never between eight and 12.”
Marks: AI-driven markets without emotional bias should produce steadier returns
“It stands to reason that it would. If America, if we programmed AI to find the elements and expect the patterns that produced success in the past and left it alone. And since we're told AI does not have greed or fear or optimism or pessimism, you would think t…”
Marks: Real-World Cycles Occur Predictably Despite Variable Timing and Amplitude
“Patterns in the real world, the non, non-scientific, non-mechanical world Certainly aren't regular, but they do occur ups and downs. They do it predictably, but not necessarily predictable in time or predictable in amplitude, but they certainly occur predictab…”
Howard Marks: US Economic Recovery Is in Healthy Middle Age
“Well, I think that the recovery is, ah, I think it's, ah, it's gone on a good while, and, ah, it's, ah, it's not nascent, it's not adolescent, ah, it's, but it's also not, ah, ah, geriatric. I think it's in middle age, ah, and I think the US economy, if that's…”
Marks: AI Infrastructure Is Booming but Bust Risk Remains Unknowable
“That, that, well, that is the area that's booming in our economy. And I need people who know more about AI than me to tell me whether it's going to bust. You know, the question is, is all of this, or is some of this construction unwarranted? I don't think anyb…”
Marks: Ambition Starts with Money but Transitions Away in Enlightened Cases
“I think that most ambition starts with the dollar sign. In enlightened cases, it transitions to non-dollar.”
Marks: Founding Oaktree at 49 Was His First Intentional Career Decision
“When I think about my early decades, not years, decades, and I think about you know, what I did from, let's say, I would say from the beginning of, let's say high school until starting Oak Tree in, in 95. So that 35 year period, which took me up to age let's s…”
Marks: Buying Nifty 50 stocks in 1969 led to 95% losses
“And if you bought those stocks, the day I got to work in September of 69, if you held them tenaciously for five years, you lost about 95% of your money. 95%. Because for many of them, something did go wrong, and for all of them, the price was too high. The P.E…”
Marks: 99% of Oaktree's high-yield bonds paid full principal and interest
“I've been in high yield bonds for 48 years. And in our experience, 99% of the bonds have paid interest in principle as promised.”
Marks: Bruce Karsh managed $70B in distressed debt with 90%+ profits
“He's managed about 70 odd billion dollars since 1988 in that field, by far the biggest. And of his total profits and losses, well over 90% are profits, less than 10% are losses.”
Marks: Exceptional, low-risk returns come from unloved assets, not popular ones
“If you go into a field that everybody likes and they like it a lot and have bid up the price, then your returns are not likely to be high. If they, if you go into a field where everybody else has been blind to the merits and says, I wouldn't touch that with a …”
Marks: Superior Credit Investing Requires Predicting Default Probabilities Better Than Competitors
“The base, the most important fundamental skill set is predicting the probability of the fault better than other people. That's where if we don't have that, that's the necessary condition for superior performance.”
Marks: Oaktree's 40-Year Default Rate Is Roughly One-Third Market Average
“Over the last 40 years, on average, something like 3.6 or 3.7% of all high yield bonds have gone into default every year. And our default rate has been Roughly a third.”
Marks: 80% Stock Drops Usually Stem From Incomplete Initial Analysis
“When you look at a portfolio and you see a stock that's down 80%, It's usually because something wasn't covered. And because the analysis was not disciplined enough to touch all the bases.”
Marks: The investor who best understands AI will dominate the next decade
“Who is the person in the investment business today who will find the most success in the next 10 years? The answer is, in my opinion, the person who best understands AI and its capabilities and implications.”
Marks: Exceptional human investors will retain an edge over AI
“Neither can most people. So, so what that says is there is a role for exceptional people, even in an AI world.”
Marks: Indexation won because active management was bad, not because indexing is great
“My answer is that indexation has taken over as it has, not because it's so good, but because active management was so bad.”
Marks: Psychological barriers prevent most investors from buying during market panics
“The bad times create an opening for active management. But it's still hard, and not many people can do it well. Why do bad times create an opening? Because panic drives down, let's say, stock prices. But the same panic causes most people to panic, which means …”
Marks: Average insight into AI will not provide an investment edge
“So if your excitement level, and by that I mean your insight level, is average, you don't have an edge. Your, and you get involved in AI, your performance will be average. You will go along with the tide if it works. You'll be schmeiced if it doesn't work.”
Kamath predicts Indian 6-7% GDP growth will drive consumer and energy outperformance
“India's growing at, say, seven percent or six percent, it will continue to do so for a period of time. GDP per capita will go up, so consumer will do well energy will do well, energy consumption goes up as GDP per capita goes up, maybe above 5000 dollars.”
Howard Marks: Readily Available Present Data Cannot Create Above-Average Returns
“Readily available, quantitative information about the present cannot hold the key to success. Because everybody has it. Success in investing is doing better than others. Investing is a funny field because it's really easy to be average, and it's really hard to…”
Howard Marks: Investing success depends on outperforming others, not absolute correctness
“This is a competitive game... You don't have to, it's not a matter of being right or wrong. It's a matter of being more right than the other person or less wrong.”
Howard Marks: Second-level thinking cannot be taught, like height in basketball
“Asking me how to become a second level thinker is like in basketball, we say you can't coach height. All the coaching in the world will not make your team taller. So can you, I can tell you that you must become a second level thinker to be successful as an inv…”
Howard Marks: Contrarian thinking must be different and better than consensus
“Second level thinking or contrarian thinking is not just different from the herd. That's not enough. It has to be different from the herd and better.”
Marks: Forcing AI to generate non-consensus ideas usually leads to errors
“If you believe that consensus is as close to right as most people can get, then if you tell Claude, I only want non-consensus thinking, then it has a high probability of being wrong.”
Marks: History rhymes because human nature changes very slowly
“There are certain themes that re that rhyme from instance to instance in history. And They mostly have to do with human nature. And I think human nature doesn't change incredibly slowly. I mean, when we talk about fight or flight, people will start talking abo…”