In a recent online interview, Howard Marks, co-founder of Oaktree Capital, shared his latest insights on investment topics including the AI bubble, gold, Bitcoin, and bonds.
Marks compared the current phase to the internet bubble period of 1998-2000, suggesting both revolve around revolutionary technologies that ignite boundless market imagination.
However, he pointed out that "changing the world" and "investors making money" are two entirely different matters.
As Warren Buffett once remarked, "There's no question the internet will dramatically improve productivity, but whether it will have a positive impact on any particular company's profitability remains unclear."
Marks believes this logic applies equally to AI today.
He contends that, compared to the internet bubble era, people currently lack a clear understanding of AI's specific commercialization pathways and profit models.
Regarding AI investment, Marks cautioned retail investors against falling into a "lottery mentality" and "binary betting" mindset.
He distinguished between two types of investment choices: one involves betting on pure AI concept companies, which carry high risk but potentially high returns; the other entails investing in large technology companies with established stable businesses, where AI merely serves as a growth catalyst.
Marks emphasized that investors must have a sober understanding of their own risk tolerance and investment plans.
On gold and Bitcoin, he reiterated his value investment philosophy, noting that such assets generate no cash flow and therefore cannot be evaluated based on intrinsic value, with their prices depending entirely on market supply-demand dynamics and sentiment.
People who bought gold a year ago have made substantial profits, while those who purchased gold at the end of 2010 have achieved approximately 7.7% annualized returns to date.
Had they invested in the S&P 500 index during the same period, they would have obtained 12.7% annualized returns.
Using historical data, Marks demonstrated that gold's long-term returns trail those of the stock market—while certainly not worthless, investors shouldn't be misled by short-term price surges.
Investment Report (liulishidian) has meticulously translated and distilled Howard Marks' key insights:
The current situation can be compared to the internet bubble period.
Host: I'm wondering, if we compare the current period to 1973-1974, 1999-2000, or 2007-2008, which era do you think our present situation most closely resembles?
Howard: I believe, if forced to choose a reference point, the most appropriate analogy—though not a perfect match—is the 1998-2000 internet bubble.
I'm not suggesting they're identical in scale, but rather most similar in nature.
The "Nifty Fifty" period differed because it wasn't driven by technological imagination, but primarily focused on established large companies.
Back then people said the internet would change the world—what was the result?
Can you imagine today's world without the internet?
It has thoroughly transformed the world in every aspect, including the fact we're communicating through it right now.
So this is comparable—AI represents another technological innovation that I believe will change the world.
But I recall we had clearer vision back then about how the internet would transform the world, and many ideas have since materialized.
There was tremendous excitement surrounding e-commerce, which has indeed become an economic pillar today.
In my view, our trajectory predictions back then largely came true.
Currently, I believe we lack that same clarity.
I'm no expert in this field—far from it—but I've yet to hear anyone explain precisely how AI will change the world.
We know it's a powerful force capable of thinking and processing data.
It can access all previously compiled information.
But what exactly it will do, how it will become a business, how people will profit from it, and how it will impact daily life—these aspects remain unclear.
Still, I do see comparability between the two eras.
Both involve new developments that capture people's imagination, and most bubbles form around such novel concepts.
In 1969, it was growth stock investing; in 2006, subprime mortgages; in 1999, the internet; in 1720, the South Sea Company; in 1620, Dutch tulips.
I've always believed bubbles form around new things because imagination becomes unconstrained, free to soar.
But you'll never see bubbles in paper or lumber stocks because these are too easily understood.
It's always the unknown new developments that make trees grow to the sky.
"Changing the world" doesn't equal investors making money.
Host: You recently had a fascinating conversation with Edward Chancellor, author of "Devil Take the Hindmost."
During this discussion, you made two bold assertions: first, that AI will change the world.
Second, that most companies people invest in today—or expect to profit from the AI wave—will ultimately become worthless.
Howard: Indeed, "changing the world" and "investors making money" are separate matters.
In fact, Warren Buffett once pointed out—I believe he said this about the internet during the 2000 annual shareholders meeting—"There's no question the internet will dramatically improve productivity, but whether it will have a positive impact on any particular company's profitability remains unclear."
I think the same applies to AI.
What worries me is that during my travels, I saw a CNN program where the host told a guest, "You believe AI has the capacity to eliminate half of entry-level positions..."
The crucial point is, this might be true.
Obviously, if US GDP remains constant while eliminating half of entry-level jobs, productivity could indeed increase.
But the question remains: will it actually increase profits?
Who will benefit from these cost savings?
When different companies compete to provide AI services, they might engage in price wars that ultimately leave no profitability.
Alternatively, when employers using AI compete for market share, all cost savings might be passed to consumers through lower prices.
Therefore, how this labor-saving tool translates into profits is something I believe nobody can clearly explain.
AI investment should avoid lottery mentality and binary betting.
Host: Although none of us can predict the final outcome, when you ask yourself, what potential mistakes seem most worthy of vigilance right now?
What specific errors should we try to avoid?
Howard: What I've observed repeatedly during market manias is: First, you might assume today's leaders will remain leaders tomorrow.
They might maintain their lead, but you shouldn't bet your entire fortune on it.
Second, don't assume investing in laggards is wise simply because leading stocks have become too expensive—just because they're cheaper.
Some say, "Although the probability of success is low, the potential returns could be enormous, so I should buy"—this is precisely what I call the "lottery mentality."
You must understand—if the success probability is low, you should accept this reality.
It means the likelihood of failure is greater.
Regarding AI, what I understand is that you can invest in pure AI concept companies that have nothing but the AI narrative.
This represents typical "binary betting"—either soar to heaven or crash to zero.
Alternatively, you can choose to invest in established large technology companies.
If AI succeeds, they'll achieve moderate gains; if AI proves less successful, they still have businesses and remain profitable.
Now we return to our initial discussion: Do you want to invest in startups with no revenue, no profits, relying solely on AI storytelling, but with potential for breakthrough success if they succeed?
Or do you prefer investing in large, steadily profitable technology companies where AI might bring incremental benefits but not revolutionary change?
This represents a choice.
It depends on your style and plan.
If you intend to make binary bets on innovative companies, you must understand the risks involved.
Gold and Bitcoin cannot be valued based on intrinsic worth.
Host: Gold prices recently broke through $4,000 per ounce, while Bitcoin similarly triggers massive speculative enthusiasm.
You once wrote in your memo, "In my opinion, for assets that generate no cash flow, there's no way to conduct valuation through analysis." I've never been able to convince myself to buy Bitcoin either.
How do you view Bitcoin and gold now? Should they be included in investment portfolios?
Howard: Both Oaktree Capital and I consider ourselves value investors.
What value investors do is examine an asset—whether stocks, companies, bonds, real estate, or anything else—attempting to determine its intrinsic value.
Then we compare today's market price against that intrinsic value.
Intrinsic value always stems from an asset's profitability—its ability to generate profits and cash flow.
For example, if I own a building generating $1 million annual net income, and I want to sell while you want to buy, we can negotiate pricing based on your expected return rate.
I might ask for $12 million, giving you an 8% return.
You might counter with $8 million, demanding 12% return due to risk factors.
I might settle for $11 million, providing 9% return with future growth potential.
So we can negotiate, but our bargaining revolves around the asset's value.
However, if you buy Bitcoin, gold, diamonds, or artwork, there's no intrinsic value to discuss.
I often use oil as an example.
I recall oil prices reached $147 per barrel in June-July 2007, then dropped to $35 per barrel six months later.
The oil remained the same oil.
Oil itself has no intrinsic value—all these assets' values and prices depend entirely on what the market is willing to pay.
If I ask a gold enthusiast or Bitcoin believer, "What price do you expect in one year?" how do they make analytical investments?
Without cash flow as foundation, what basis do they use for judgment?
But as I mentioned in my memo, you can buy gold because you believe in it, think it will rise, or invest because it's historically been a store of value.
However, you cannot analyze whether it's undervalued based on so-called "intrinsic value."
I still maintain this perspective.
People who bought gold a year ago have made substantial profits.
Right before our call, I checked the numbers: if you bought gold at the end of 2010, your annualized return to date is approximately 7.7%.
If you bought the S&P 500 index during the same period, you'd have achieved 12.7% returns.
So gold certainly isn't worthless, but you shouldn't be misled by its short-term price surges.
It has consistently been a mediocre investment.
High-yield bonds remain relatively attractive assets.
Host: In December 2022, you wrote a crucial memo titled "Sea Change," discussing the end of an era—the declining interest rate environment since 1980 has disappeared.
If we want equity-like returns in this rather concerning environment, what should we do?
Howard: High-yield bonds belong to what I call "credit assets."
The market calls them debt, fixed income, bonds, notes, or loans—all fall into the same category.
These are financial instruments where you lend money to others who pay you rent, promising semi-annual interest payments with principal repayment at maturity.
You can calculate based on these facts, determining what return rate you'll achieve by lending money at current rates until maturity.
It's called fixed income because the outcome is predetermined.
This represents a contractual relationship promising fixed returns.
Once you purchase, only one variable remains in the equation: the likelihood that the counterparty will honor interest payments and principal repayment.
Assessing this probability is credit analysts' work—a relatively specialized field.
For most people, I recommend that amateurs in most investment areas choose funds, ETFs, or similar products managed by professional teams.
Remember what Charlie Munger said: "Investing isn't easy—anyone who thinks it's easy is a fool."
This applies not only to what I do, but equally to stocks, mutual funds, ETFs, index funds, and various other areas.
Investing represents an interesting business where achieving average returns is easy, but obtaining excess returns proves difficult.
However, if you're satisfied with average returns, you can purchase products at relatively low costs with high probability of achieving average performance.
Taking high-yield bonds as an example, they currently offer approximately 7% yields, with other credit products potentially slightly higher.
If you can accept this return range, numerous products already exist in the market to help you achieve it.
Comments