Success accumulates capability, but it can also harden a once-effective method into habit and identity. Google’s hesitation before the ChatGPT moment, Munger’s description of Buffett’s “brain block,” and the later performance gaps of Steve Eisman and John Paulson all point to the same caution: spotting one historic mispricing is not the same as possessing an advantage that adapts across market regimes. The essay distinguishes stable principles from revisable conclusions, and a circle of competence from path dependence, arguing that companies and investors should remeasure their boundaries when facts change so that moats and past victories create room to adapt instead of becoming walls.
Recently, I came across a recollection from someone involved in Google’s early work on large language models. Before ChatGPT appeared, DeepMind had already come remarkably close to what would later be called the “GPT moment,” but never truly brought it to market. Thibault Sottiaux later recalled that when ChatGPT arrived, he realised it was something DeepMind had been “sitting on” for almost two years. (WIRED)
What makes the story most revealing is not whether Google built some particular product before OpenAI. It is whether an organisation that is already successful enough has the courage to release something that might redefine it.
On paper, innovation is always welcome. As long as it improves efficiency, increases revenue, and leaves the existing order untouched, almost no one objects. Truly difficult innovation is rarely about adding one more feature to the old system. It asks a company to admit that the part of its business that is most profitable, most familiar, and its greatest source of pride today may matter less in the future.
At that point, advantage itself becomes resistance.
A company that has yet to succeed has little to lose, so it can bet on the future more easily. For a company that has already built a vast business, organisation, and web of interests, every step forward feels like an erosion of its own past. It may not be blind to change. The old world is simply still working well, which means there is always a reason to wait a little longer.
The Kodak story is retold so often not because Kodak failed to see digital photography at all, but because seeing the future and being willing to let it replace the present have always been two different things.
Many companies do not die from ignorance. They are held back for too long by judgements that were once correct.
Success gradually builds an entire system of explanation. The product remains profitable. The distribution channels remain strong. Customer habits do not seem to have changed. The new technology is still immature, and the new business model has yet to prove itself. Each reason may be valid on its own. Together, they are enough to keep an organisation rationally and respectably hesitant at the very moment it most needs to act.
Investors, too, find this hesitation hard to escape.
There is a passage in Poor Charlie’s Almanack that I particularly like. After rejecting the idea that he had been Buffett’s decisive teacher, Munger did not portray Buffett as someone naturally immune to the grip of old experience. He acknowledged that after Buffett worked under Ben Graham early in his career and made a great deal of money using that approach, Buffett developed a touch of what he called “brain block.”
Munger did not attribute this entirely to a flaw of character, nor did he use it to dismiss Graham’s method. What he identified was closer to a universal and largely hidden fact: once a method has made you rich, leaving it is far harder than learning a new one.
People often say that failure leaves scars, but speak less often about the switching costs created by success. Failure at least forces us to question an old answer. A sufficiently brilliant victory can turn a method from a tool into a habit, and from a habit into an identity.
Some of the celebrated short sellers of the financial crisis offer an interesting footnote.
Steve Eisman was one of the real-life inspirations for Mark Baum in The Big Short. He and his team returned 66.2% in 2007 by shorting subprime mortgages, making him one of the financial crisis’s best-known investors. Years later, the firm he founded, Emrys Partners, did not repeat anything like that result: the fund returned about 3.6% in 2012 and 10.8% in 2013, before closing in 2014. (Reuters)
John Paulson’s reversal was more dramatic still. His bet against the US subprime market in 2007 became one of the most famous trades in financial history. Yet in 2011, his Advantage Plus fund lost about 52% and the Advantage Fund about 36%. Performance remained uneven after that and assets under management kept shrinking. In 2020, Paulson & Co. returned its outside investors’ capital and converted into a family office managing Paulson’s own wealth. (Reuters on the 2011 performance; Reuters on the 2020 conversion)
Of course, those numbers alone do not prove that Eisman or Paulson underperformed because of path dependence. Market conditions and strategies differed; changes in fund size and risk appetite also affect returns. Reducing an investor’s later failures to “being trapped by the past” is the same kind of intellectual laziness as turning him into a legend after a single success.
But these cases do remind us of at least one thing:
Spotting one historic mispricing and possessing an edge that continually adapts to different market environments are not the same thing.
The danger created by an enormous victory may be subtler than simply repeating one kind of trade.
What someone truly begins to depend on may no longer be the original trade, or even a concrete method such as “shorting real estate.” It may be the weight that victory has taught him to assign to his own judgement.
If someone saw the problem when everyone else believed house prices could not fall, and made unimaginable wealth by rejecting the consensus, then the next time he confronts a consensus he will naturally trust his doubts more than most people do.
That confidence has value. Without it, he might never have held on to the original trade.
The problem is that the market does not raise the odds that someone will be right next time merely because he was right before.
Especially in short selling.
Someone who once made a fortune when a bubble burst can easily become dependent on the feeling of “seeing through the boom.” He grows more sensitive to the cracks in valuation, narrative, and sentiment, but also more likely to overlook another possibility: some absurd-looking prices are not simply waiting to collapse. They may reflect a change that the old framework has not yet learned to contain.
The most seductive thing about short selling is the sense of clarity it can confer on someone standing against the majority. Its greatest danger lies in exactly the same place.
In short selling, being right about direction is nowhere near enough. Timing, position size, liquidity, and market sentiment can each determine the outcome before the logic itself does. A short squeeze, a rally detached from fundamentals, or simply a market that takes longer than expected to recover its reason can erase accumulated gains in a remarkably short time.
Markets are under no obligation to vindicate an investor on a timetable that investor can survive.
More troublingly, when someone has once been richly rewarded for rejecting the consensus, “being unlike the majority” can gradually acquire a meaning of its own.
At first, the thought is:
I have found a problem the market has overlooked.
Later, it can slowly become:
If most of the market believes it, there must be a problem.
The two sentences look alike, but their logic runs in opposite directions.
The first begins with facts and ends at a conclusion that differs from the market’s. The second begins with “I must differ from the market,” then goes looking for facts that can prove the market wrong.
When an investor starts searching for the next bubble that can prove he is still more clear-sighted than the market, his dependence is no longer merely on a trading method. It is on the version of himself that was once so richly rewarded for being different.
This may be the most dangerous form of path dependence created by success: a past victory does not merely tell you that a method works. It quietly tells you what kind of person you are.
And people often find it harder to give up an identity than a strategy.
The problem, of course, is not short selling itself.
Nor is the problem a person’s long-term commitment to value investing, trend following, or any other method they genuinely understand.
A circle of competence is not narrow-mindedness, much less path dependence. An excellent value investor can remain a value investor for life, with no need to study every speculative technique merely to prove an open mind; nor does a trend follower need to force themselves to forecast a company’s cash flows ten years from now.
What truly needs updating may not be the method, but the understanding of the world beneath it.
A value investor can remain a value investor, but their understanding of business models, moats, industry structure, and capital efficiency cannot remain permanently fixed in the past. A trend follower can continue to respect price, but cannot assume that market participants, sources of liquidity, and trading mechanisms have never changed.
Principles can remain stable. Conclusions must not be allowed to harden simply because of that stability.
Learning does not mean chasing every new fashion, much less endlessly overturning one’s entire system. More often, it simply means admitting that an answer which once worked does not automatically retain the right to explain the future.
I increasingly think that very few things are worth preserving for the long term.
An awareness of risk. Honesty in the face of facts. A willingness to admit what we do not know. The patience to wait when no real opportunity exists.
These are closer to principles.
How a particular industry should be valued, what multiple a given asset should always command, what kind of company truly possesses a moat, or even which trading method has an advantage in the present market—these are closer to conclusions.
The purpose of principles was never to preserve our conclusions forever. It is to help us reach new conclusions when the environment changes.
The same is true of a circle of competence.
A circle of competence may have boundaries, but those boundaries should not be permanently drawn by a past success. They can expand slowly through learning, and they may need to be measured anew when the world changes.
Staying within a circle of competence means knowing what you do not understand. Path dependence means gradually losing the ability to judge anew whether you understand it at all.
Companies are no different.
A truly good company may not always see the future first, but it should at least allow something new to challenge it. It need not repudiate every old business the moment a new technology appears, but it must distinguish between the core capabilities worth preserving and the generous returns that an old era happens to have left behind.
Moats matter, of course.
But a moat should give a company more room to make mistakes and transform itself, not make it more comfortable remaining in the past.
The same applies to individuals.
The best use of money earned, experience accumulated, and systems developed in the past is to give a person more freedom in the face of the unknown. Because he already possesses something, he can admit more calmly that he may be wrong. He can learn again. When necessary, he can let go of a judgement that no longer fits.
If success instead leaves someone increasingly compelled to defend the image of his formerly correct self, then wealth, experience, and reputation eventually become switching costs of their own.
A truly mature investor does not need to know how to do everything.
What he must preserve is the ability, once the facts have changed, not to rush to defend the person he was when he was right.
Past success is certainly worth cherishing. But its best use is to give us more room to face the unknown, not to write the future’s answers in advance.
Once a moat loses its movement, it can eventually become a wall.
And the hardest path dependence to escape is often not the wound left by failure, but the road that once carried us forward far too smoothly.
THE CONVERSATION
Leave a thought. 00
Different perspectives are welcome. Be kind. Comments in both languages meet here.
No published comments yet. Your thought could be the beginning of a conversation.