Will Attending an Investment Conference Make You Sad?

Emerging markets investor Harvey Sawikin of Firebird Management has a funny post on his Substack, The Falling Knife, asking whether attending an investment conference makes you sad. His conclusion, after some pseudo-rigorous math farmed out to a fictional AI assistant named "Chet Gepetti," is that it probably does, at least slightly. 

The mechanism is well-known to anyone who has attended an investing idea conference: the asymmetry between the pain of a bad idea that you bought and the regret of a good call that you passed on. Losses feel worse than missed gains feel good, but both sting. Add in the 40% of ideas that are duds, some conference food, and the nagging memory of a General Growth Properties pitch you heard from Bill Ackman in 2009 and then did nothing about, and the math turns against you.

His post resonated with me because in 2014 I attended a value investing conference where twenty-four ideas were presented and I have been tracking them, off and on, ever since. (It's been with some schadenfreude, since the conference organizer wouldn't let me present my idea, a small bank for 63% of tangible book value.) The results are instructive, though not in the way that value investors usually hope.

The 2014 conference's dogs were spectacular in their failure. Civeo Corporation, a workforce housing company, immediately plummeted after the conference and is still around, down roughly 90%. Iconix Brand Group, a royalty-driven brand licensor eventually went to near zero, down 99.7%, a near-total wipeout. Forest Oil merged with Sabine Oil & Gas in December of that year and Sabine filed for bankruptcy the next spring. QR Energy merged with BreitBurn Energy Partners, which also went bankrupt. Resolute Forest Products held on long enough to be acquired for a modest premium, though the Canadian dollar's 20% decline over the period ate much of that.

What is striking about this list is not just the magnitude of the losses but the speed. Several of these companies effectively ceased to exist within twelve to eighteen months of being presented as cheap. 

It was a value investing conference. The central premise of value investing, going back to Graham and Dodd, is the margin of safety: the idea that buying at a sufficient discount to intrinsic value protects you from catastrophic loss. If you are right about the asset value and wrong about everything else, you should at least get your money back. The failures indicate that either the margin of safety calculations were wrong, or that the concept does not travel well to certain kinds of businesses. Perhaps both. Commodity-producing companies, leveraged companies, and companies with deteriorating secular trends have a way of going from "cheap" to "zero" while being "undervalued" the entire way.

The base rate of public companies going bankrupt in any given year is well under 1%. The proportion of ideas from this conference that ended in bankruptcy or near-bankruptcy within a few years was dramatically higher than that. This is not a random sample, of course. Value conferences tend to attract the contrarian, the beaten-down, the deeply discounted. Which is precisely the problem. There is a selection effect toward the kinds of companies that appear cheap based on flawed metrics.

There were some good ideas too, and one extraordinary one.

General Motors was roughly flat over the decade, which is perhaps the most value-investor outcome imaginable: you do the work, you are right about the valuation, you hold for years, and you end up where you started. U.S. Steel roughly doubled. Visa returned something like seven times your money on a total return basis, which is an excellent result, though it required you to own a wide-moat payments network at a time when you were surrounded by people pitching over-leveraged forest product companies and housing for oilfield workers.

But then there was Nvidia Corporation (NVDA).

Remember that at the time of this conference in 2014, it was primarily a gaming GPU company. (We, of course, had zero interest in anything linked to "gaming," for dorks.) Nvidia has since returned something in the neighborhood of 500 times your money. If you put 4.2% of your portfolio into each of the twenty-four ideas at equal weight, the naive strategy, the math is remarkable. Visa alone would have covered most of the dogs. But then Nvidia would have returned roughly 20 times your entire starting portfolio over twelve years.

The problem is what that looks like in practice. After a few years, a genuinely equal-weighted portfolio becomes anything but equal. Nvidia would have grown to dominate the portfolio so completely that you would have experienced 50%-plus drawdowns in 2018 and again in 2022, watching what was functionally your entire portfolio cut in half, twice, before the final ascent. Almost no professional manager could have survive that. Harvey addresses exactly this in some other Substack posts on art collections and Firebird's early portfolio: even if you have the right positions, the institutional constraints of managing other people's money make it nearly impossible to ride a 75% concentration in two stocks through a bear market without your limited partners concluding that you are a reckless gambler.

The only way to have actually captured the Nvidia return in its entirety was probably to put these stocks in a brokerage account you never checked ("coffee can"). The position sizing problem is not just psychological, it is structural. A manager who communicated to investors in 2022 that they were 70% in a semiconductor stock that had just halved would not have had LPs for long. The money would have left, locking in the loss, before the recovery came.

Sawikin writes about this dynamic in the context of art collections, noting that Keynes' collection at King's College has compounded well precisely because no one was tempted to "trim the Cézanne position." Stocks do not have that feature. You get a price every day, your investors are watching, and the pressure to do something is constant.

I only went to that one conference in the series and have occasionally wondered what I missed in subsequent years. It would be interesting to look at the other vintages. Did the hit rate improve? Did anyone present NVDA again in 2016, after it had already doubled, and catch the remaining 250x? At what point did the energy and commodity ideas, which dominated the 2014 vintage in a way that reflected the zeitgeist of that moment, give way to technology ideas in later years?

I do attend a different annual value investor dinner and have returned for several years now. One pattern I have noticed is that the best ideas (so far as I can judge) are also the ones that are best presented. This is not a coincidence. Clear communication is demonstration of clear thinking. An idea that can be explained simply and directly has fewer moving parts, which means fewer things that can go wrong. The people who can stand up and say quickly and clearly exactly what they own, exactly why it is cheap, and exactly what has to happen for it to work out, are people who have better mental models of the world.

The corollary is that a complicated pitch is often a warning sign. If the presenter cannot explain why a company is cheap without also explaining four offsetting factors that are currently obscuring the value, there is a decent chance that one of those factors eventually wins. The companies that went to zero from the 2014 conference generally had stories: the oil price will recover, the licenses will be renewed, the merger will unlock the discount. The thesis required multiple things to go right. They did not.

The simple ideas, like buy the world's dominant payments network, buy the AI-adjacent chip company when no one cares about AI, turned out to be the ones worth owning. The elaborate restructuring plays and commodity-cycle bets, the ones that required the most slides and the most assumptions, are the ones in the bankruptcy column.

This is not a new observation. But watching it play out over a decade, with a specific set of names and a specific date stamp, makes it more concrete than it usually appears in the abstract.

A Reply on "Places" by PdxSag

[From our correspondent @PdxSag. His previous guest posts on CBS include What I've Learned the Past Decade, A "Wonderlic" Test for Agency, and his supplement regimen.] 

A recent edition of CBS Links quoted Hickman: "there are only a handful of 'place genres' out there, and each is generally produced by geography above all."

Hickman is getting his causality a little backwards here. He is correct that there are only handful of first-tier locales, but that's because there is only a handful of first-tier people in the world. What you need is a critical mass of high-agency people. High-agency means self-selection, and that implicitly requires immigration.

At the fundamental level you need a filter. You want to separate high-agency people from NPC's. Immigration, because it is a formidable barrier in both cost and effort, functions as the ultimate people-filter. Once you realize immigration is an implicit functional requirement, which means people have to make a large effort and move, all else equal, nice geography beats bad geography every single time. Obviously.

The next thing that follows from immigration as a filter is that that filter is lowest for rich people. So first-tier places will have the most rich people and the rich people that are there will have the lowest agency, which is to say most NPC-like mentalities. If that doesn't describe the world, I don't know what does. Yogi Berra famously meme-ed this phenomenon with the line “nobody goes there anymore, it’s too popular."

Once a fist-tier place is established, its high cost creates a self-reinforcing cycle that drives away low-agency people that through whatever hereditary luck landed them there. As cost to be in a place goes up, the pay-off for leaving goes up. If you’re an NPC, there is an easy and obvious arbitrage for selling and moving elsewhere. For people immigrating, whether high-agency or NPC's, the relative high cost dissuades both, the exception is the high-agency people that are into whatever specific scene makes that locale first-tier.

As more proof, the self-reinforcing cycle creates an interesting failure condition at the extreme. If cost gets so high that it becomes the sole barrier, high-agency people won't even immigrate there any more. The place becomes a rich-person amusement park. It is no longer a “real” place. It becomes a “place genre” where rich NPC’s just act out their genre programming. Examples: NorCal (to a lesser extent — it’s a big place), Sedona (obvious because it’s so much smaller), Martha’s Vineyard (smaller still), Aspen (tiny), Monaco (tiniest, relative to the whole planet from which it draws). Each place is smaller, and as it gets smaller price becomes a bigger barrier, until it is the only barrier.

Finally, we can test our assertion that immigration is the ultimate people-filter by looking at what happens when we subvert that filter through financial subsidies and perverse incentives. It so happens this is the nature of third-world to first-world immigration today. Third world immigrants the last several years are the worst of any cohort in history. This cohort has the most criminals, the least skills, and the least interest in assimilating. Immigration is not selecting for agency of any kind, rather it is selecting for criminals fleeing prosecution and grifters looking for easy social services scams and retail theft rings. These are not even high-agency thieves. These are ethnic fraud rings that often just need bodies to show-up and run a farcically simplistic scam that is only possible because the host countries are willfully ignoring the scams for political purposes.

And where are these immigrants moving to? All things equal, it should be the best, or at least the better half. Yet, it is almost the worst in the United States: Minnesota, Michigan, Maine. The implication is obvious: the immigrants are not choosing their destination, it is being chosen for them. There is one notable geographical exception too: Texas. Why Texas? If Democrats can flip Texas, Republicans will never again win the White House. The one good state with the most need for foreign immigrants to flip it for Democrats is the one state with the most H1B immigration from the Middle East and India. This is so glaringly obvious only an NPC could miss it. These are not high-agency immigrants. These are NPC's serving a political purpose for the Elites sponsoring the whole project.

Thursday Morning Links

• Trump has escaped other predicaments of his own making, but there is something different about this one. The attack on Iran is so wildly inconsistent with the wishes of his own base, so diametrically opposed to their reading of the national interest, that it is likely to mark the end of Trumpism as a project. Those with claims to speak for Trumpism – Joe Rogan, Tucker Carlson, Megyn Kelly – have reacted to the invasion with incredulity. Trump may entertain himself with the presidency for the next three years (barring impeachment), but the mutual respect between him and his movement has been ruptured, and his revolution is essentially over. [The Spectator]
• Trump could evoke Truman’s precedent against the Japanese – a rationale widely accepted by American historiography on similar arguments: The need to avoid a US ground invasion against a fanatical enemy. Moreover, in the current atmosphere of total contempt for the very concept of “international law”, he could take the pragmatic view and ask what actual blow-back would he actually face? US allies have already been brutalised over Greenland and other issues in far more direct ways, and still don’t dare to break with Washington. As for US enemies, this live demonstration of nuclear resolve might even strengthen deterrence against them. [Brussels Signal]
• Firms are smart money, and when firms sell equity, that’s a sign that equity is overpriced. Currently, we do not see positive net issuance across the entire market – i.e., firms are not net sellers of equity to external investors. This fact is evident in Figure 2, which shows the sum of dollar net issuance in the past year, normalized by the stock market’s total capitalization. For much of the 1990s, issuance was positive, and at the peak of the bubble in 2000, issuance was more than 5% of market cap, meaning that 5% of the market had been newly issued in the past year. After that bubble collapsed, issuance turned negative and has largely remained so, with the notable exception of the COVID bubble of 2021, when issuance peaked above 2%. But today, issuance is decidedly negative at -0.9%; the supply of shares is shrinking. Ballpark figures, the value of U.S. common stock is around $65T, and every year companies are buying around $500B more equity than they’re selling. [Acadian Asset Management]
• In a construction materials market affected by soft demand and economic uncertainty, Titan America delivered all time high revenue, Net Income, Adjusted EBITDA and operating cash flow in 2025. This achievement reflects the strength of our business model, disciplined decision-making, skillful execution across our operations, and an unwavering focus on serving our customers. It showcases once again Titan America’s ability to grow organically and deliver strong results, even in challenging environments. Our Florida segment delivered a robust performance with strong penetration in infrastructure and private non-residential construction segments offsetting a soft residential end-market. Our investments in increased aggregates capacity, expanded capabilities, and self-help operational excellence initiatives delivered record full year revenue and Adjusted EBITDA in 2025. Our Mid-Atlantic segment was impacted by a combination of soft demand in Metro New York and New Jersey, the introduction of tariffs, and weather affecting Virginia and the Carolinas. Resilient pricing, and continued growth in infrastructure, private non-residential construction, including data centers, and cost containment initiatives partially mitigated the impact from the headwinds in the region. [Titan America SA]
• Paul R. Ehrlich, an eminent ecologist and population scientist whose best-selling book, “The Population Bomb,” was celebrated as a prescient warning of a coming age of food shortages and famine but later criticized by conservatives and academic rivals for what they called its sky-is-falling rhetoric, died on Friday in Palo Alto, Calif. He was 93. [The New York Times]
• For a young single person, the economic and social advantages of urban network effects clearly outweigh the dangers of life in the city. You can make more money. The social opportunities are a lot more interesting. The food generally is better. (This is why liberals in Portland and San Francisco love to ostentatiously pretend not to know what you’re talking about when you describe the problems of living there: it’s a flex.) But when people have kids, they quickly lose their confidence that they can fly above the social dysfunction indefinitely. All sorts of abstract, macro policy problems suddenly don’t feel so abstract. Regardless of where they live, Americans are forced to spend colossal amounts of psychological and economic resources acting as private security for their kids — which is at least as inefficient and silly as building a private pool in every backyard. Not only does it exhaust parents, but it prevents kids from developing autonomy, and the physical fitness that naturally comes from running around the neighborhood. [EXIT Newsletter]
• Recent events have interrupted operations at major facilities in Qatar, a significant global supplier. Industry observers note that this situation raises the prospect of widespread shortages, potentially marking the fifth global helium shortage in the past 20 years. These recurring challenges highlight the risks associated with concentrated production, particularly from large projects linked to hydrocarbon operations in geopolitically sensitive regions. [North American Helium Inc.]
• What you’re seeing in the remarkable chart above is that private sector spending on AI in 2026 is forecast to exceed $700 billion, which is not a number that is easy to think about. As a share of GDP, companies are currently devoting more resources to AI than the combined peak annual capital spending on key 1930s public works projects and the Manhattan Project and the 1940s electricity boom and the Apollo Project and Interstate highway construction. It’s important to note that AI spending is overwhelmingly financed by the private sector, whereas most of those infrastructure projects were financed by the largesse of the federal government. Once again, nothing like this has ever happened before, and if you feel extremely confident about how this is going to turn out, I think you might be crazy. [Derek Thompson]

Natural Gas-Fired Memos

A friend of ours recently used Claude, the LLM built by Anthropic PBC, to draft a letter to a public company CEO with suggestions for improving a struggling subsidiary after a disappointing Q4 2025 earnings report. It was not quite an activist letter, more like a thoughtful memo from a concerned shareholder.

When he sent us the draft, we winced a little. It was obviously AI-written. Chatty and breezy. It used the CEO's first name repeatedly in a way that felt like a car salesman working a showroom floor. We assumed that he would let us help edit it before sending. Instead, he fired it off on a Friday afternoon, as-is.

We thought he had blown his credibility by sending it, so we were shocked that the CEO responded to his correspondence the same day. They had a two hour, candid phone call to share ideas the following Monday morning. Six months ago, "AI-written" was unambiguously a pejorative. Apparently that is changing faster than we thought.

As Luddites who don't like change (we own coal royalties!), we have been AI skeptics for most of the past several years. We chuckled at Phil Greenspun's recent post, Why Johnny LLM can’t read web page source code. We read Stephen Wolfram's book a year ago. He was impressed with LLMs but doubted they would make it all the way to artificial general intelligence. These seemed like reasonable positions.

What started to shift our thinking was The Scaling Era, which pointed us to a 2019 essay called The Bitter Lesson by Richard Sutton. His argument was that general methods for artificial intelligence that leverage computation always win when competing with methods that attempt to take advantage of domain-specific human knowledge. Historically, AI researchers have always tried to build knowledge into their systems by encoding rules, designing expert systems, and programming in domain expertise. 

The pattern in the past is that this domain-specific structuring has helped in the very short run but it always plateaus and then is outcompeted by brute force computational methods. In the games of Chess or Go, search and learning are the techniques that have worked, using as much computational power as possible. Sutton says that these are the only two methods that seem to scale to arbitrarily large degrees.

This reminded us of a book by psychologist Paul Meehl (1920-2003), who argued essentially the same thing sixty-five years earlier in a completely different field. His book Clinical vs. Statistical Prediction (published 1954) showed that actuarial rules (statistics applied to data) consistently outperformed clinical judgment by professionals. Whether they were doctors, parole boards, or admissions officers, human experts with years of training were consistently beaten by simple models. (This has been called The Robust Beauty of Improper Linear Models in Decision Making.) 

Meehl and Sutton had the same insight. Systematic data processing beats human intuition, and the gap widens as the amount of data and especially the computation available to mine it grows. LLMs are like Meehl's simple linear models for decisionmaking, except scaled to the size of trillions of tokens now available for training.

The clincher for us was the conclusion of The Scaling Era. The author believes in what we have been calling the Gods of Straight Lines: "the GPUs keep improving, the training runs keep scaling, and on a log-log graph everything is eerily, perfectly linear. Next year is simply this year, plus a known rate of change multiplied by delta-t."

We've learned not to fight straight lines on log charts. We now take it for granted that solar panels and batteries will continue to get cheaper every year. Not only have the learning curves been relentless for decades but they have even gotten steeper. The curves governing LLMs look the same. Whether or not we or the AI researchers fully understand why their models work, the scaling curves don't care about our epistemology.

That, in turn, virtually assures that the Mag 7 companies (Zuck) and the AI "labs" (Elon doesn't like that term) will spend their planned trillions on capital expenditure for GPUs and data scenters. Leopold Aschenbrenner is a former OpenAI researcher who published a widely-read series on AI trajectories last year. He has sketched out scenarios involving trillion-dollar compute clusters drawing 100 gigawatts of power. That number seems outlandish until you look at the capex announcements from the past six months and realize the industry is already on that trajectory.

We have started calling our friend's memos his "natural gas-fired research." (Natural gas is the largest source of electrical generation in the U.S., meaning that is what most likely powers any given GPU writing memos.) Which is why we keep coming back to the natural gas pipelines owned by Kinder Morgan (KMI) and Energy Transfer (ET). There is an important distinction between producing natural gas and transporting it. Producing it is a lousy business, but transporting it via pipeline is something else entirely. The pipes are already in the ground. Permitting and right-of-way acquisition to build new ones is enormously difficult. Customers sign take-or-pay contracts that guarantee payment regardless of whether they actually use the capacity. It is a toll road, not a commodity business.

Kinder Morgan operates about 58,600 miles of natural gas transmission pipeline, more than any other company in the country, moving roughly 40% of U.S. natural gas production. Around 96% of its cash flows are take-or-pay, fee-based, or hedged, making the commodity price of natural gas nearly irrelevant to its earnings. Energy Transfer's latest presentation is explicit about what is driving new growth. Its 2026 capital plan lists "natural gas pipeline projects serving data center facilities" as a priority in both its intrastate and interstate segments. It has a long-term agreement with Oracle to supply roughly 900,000 MMBtu per day to three U.S. data centers, and began flowing gas on its first lateral to a data center campus near Abilene, Texas earlier this year. 

The pipeline companies collect a toll on natural gas movement, and their inherent operating leverage will makes incremental throughput for data centers highly profitable. Beyond data centers, there are coal plant retirements, Sun Belt population growth, and manufacturing reshoring all adding to the natural gas story. Kinder Morgan points out in their investor presentation that U.S. natural gas supply has grown more than 70% since 2010 with prices remaining largely range-bound, suggesting no supply constraint on the horizon.

As AI infrastructure spending scales, the competition for land near power infrastructure will intensify. A wind or solar farm on a large tract with good resource potential can supply a data center directly via long-term power purchase agreement, or feed into a grid that increasingly needs it.

We've also seen some overwrought pessimism surrounding AI-induced disruption. The Citrini piece (The 2028 Global Intelligence Crisis, published February 22nd) triggered a software selloff the following day. We think that was overdone. As Fred Liu pointed out, software companies don't really sell code. They sell convenience, trust, reliability, and institutional knowledge. The devil is always in the details like maintenance, security patches, compliance, and integrations. Most businesses would rather pay someone else to handle all of that than own it themselves. 

Our Docusign subscription just auto-renewed. We're still using Quickbooks and Turbotax. Last week we chose to fade Citrini's pessimism with a basket that included the payment networks (Visa, Mastercard, Amex), the bank core processors (FIS, Fiserv, Jack Henry), Booking Holdings, Intuit, and Schwab.

Visa's experience has been that potential competitors just become customers of their (four-sided!) network, which is the biggest and the best. The bank core processors (i.e. the software back-end of almost all banks) have contracts that define "sticky": five to ten year commitments by the banks, plus existential risk to the bank's business if an attempted switch goes poorly. 

Futurist James Pethokoukis (previously) articulated two possibilities for AI: either it is a powerful general-purpose technology (like the PC or the internet), or it advances all the way to AGI, capable of performing essentially every economically valuable task humans perform.

Vaclav Smil's book Growth points out that exponential growth, natural or anthropogenic, is always only a temporary phenomenon, to be terminated due to a variety of physical, environmental, economic, technical, or social constraints. Eventually every exponential growth curve becomes logistic. 

Whether that happens at "very powerful tool" or at "AGI" is the question. If it's AGI, our stock portfolios will not matter. What will matter is who controls it and what they decide to do with the rest of us.

But if AI lands at "powerful general-purpose technology," and that is the scenario we currently find most likely, then the implications are good. A sustained productivity boost would push the economy onto a higher growth path. That makes the debt and entitlement problems of the developed world more manageable. And to the extent that AI advantages incumbents with established distribution and customer trust, the companies in our basket look like reasonable places to be. 

And pipelines will profitably go "ssssssss" transporting natural gas to power the computation, at least until battery and solar get cheap enough to take over. (But we can own the land where that will happen.)

Our friend got instant followup from the CEO of a billion dollar company because a machine wrote a persuasive memo on a Friday afternoon. That's not AGI, but it might be enough.

Saturday Morning Links

• Imagine that every day you wake up in your left-bank apartment, and the city has meaningfully morphed into some magically strange variant of Paris. On Tuesday, the streets and boulevards no longer meet at their old familiar intersections. On Wednesday, the Louvre moves to another arrondissement. The Arc de Triumphe turns upside-down on Thursday and floats in the sky on Friday. Now we’re talking. Now that is more like parenting. To be a parent is to be a permanent tourist in a constantly evolving foreign city, which also happens to be your home. The baby you bring home from the hospital is not the baby you rock to sleep at two weeks, and the baby at three months is a complete stranger to both. In a phenomenological sense, parenting a newborn is not at all like parenting “a” singular newborn, but rather like parenting hundreds of babies, each one replacing the previous week’s child, yet retaining her basic facial structure. “Parenthood abruptly catapults us into a permanent relationship with a stranger,” Andrew Solomon wrote in Far From the Tree. Almost. Parenthood catapults us into a permanent relationship with strangers, plural to the extreme. When you become a parent, you meet your child. And then you meet your child again. And again, every day after that. You will never stop meeting your child. That is one reason to become a parent: To have a child is to fall in love with a thousand beautiful strangers. [Derek Thompson]
• Douglas maintained that he could assume judicial senior status on the Court and attempted to continue serving in that capacity, according to authors Bob Woodward and Scott Armstrong. He refused to accept his retirement and tried to participate in the Court's cases well into 1976, after John Paul Stevens had taken his former seat. Douglas reacted with outrage when, returning to his old chambers, he discovered that his clerks had been reassigned to Stevens and when he tried to file opinions for cases in which he had heard arguments before his retirement, Chief Justice Warren E. Burger ordered all justices, clerks, and other staff members to refuse help to Douglas in those efforts. When Douglas tried in March 1976 to hear arguments in a capital-punishment case, Gregg v. Georgia, the nine sitting justices signed a formal letter informing him that his retirement had ended his official duties on the Court. It was only then that Douglas withdrew from Supreme Court business. [William O Douglas]
• Esoteric reading, being very difficult, requires one to slow down and spend much more time with a book than one may be used to. One must read it very slowly, and as a whole, and over and over again. It will probably be necessary to adjust downward your whole idea of how many books you can expect to read in your lifetime. [Mr. and Mrs. Psmith’s Bookshelf] 
• The true burden of debt is not the size of our national debt, but the cost of servicing that debt as a percent of our national income. Today that burden is significantly less than it was during the 1980s, mainly because interest rates are far lower than they were back then. If Congress exercises even modest restraint and the Fed doesn't have to raise interest rates (which they won't have to if inflation remains under control), then we can gradually reduce our deficits and the burden of our debt. [Scott Grannis]
• If I insisted upon elbowing my way into the metropole — so that I could pay dearly for the very comforting feeling of being only 30+ subway stops from where the Real Action Happens — I’d be broke, overworked, and essentially unable to write for a living. And so it is that by virtue of my incorrigibly low station in this world, I must confine myself to wherever’s dirt cheap — and that, my friends, is the hinterlands. It comes at a cost, of course. The isolation is extreme; it’s distressing that 100% of my intellectual life has to happen online. The winters are brutal. Pollen season is a nightmare. The summers are sweaty. My neighbors watch me — they’ll never really accept me even though I grew up just two counties away. We are a long way from Santa Barbara here. But what makes Santa Barbara, Santa Barbara? What makes a Manhattan? And why will there never be a hub of elite culture in Guymon, Oklahoma? I suppose I am a bit of a “hard geographical determinist.” It seems to me that, so far as we are speaking broadly and with an eye toward recognizing patterns, there are only a handful of “place genres” out there, and each is generally produced by geography above all. [Hickman's Hinterlands]
• But, beginning in 2000 – and perhaps because of Buffett’s age – Berkshire regrettably abandoned that discipline. It has since invested hundreds of billions of its hard-won cash into many businesses that no one can handicap. Worse, it has repeatedly bought whole businesses with very average economics, even when partial stakes of excellent businesses were readily available on the public markets, perhaps because of a mistaken belief that the resulting tax efficiency would prove more valuable than simply buying the better business. (Analysis to follow.) Berkshire has now become what Buffett mocked for decades: a conglomerate built for the ego of its management team, not for the benefit of its shareholders. [Porter & Co.] 
• South Bow Corp. says it is in the “early stages” of gauging customer interest in capacity on a cross-border pipeline linking Alberta’s Hardisty oil terminal to U.S. oil markets in Oklahoma and along the Gulf Coast — a project it is dubbing the “Prairie Connector.” The company said Friday it has commenced an open season — industry parlance for inviting potential customers to reserve capacity on a proposed project — using existing infrastructure and Canadian federal permits that are believed to have been originally issued for the cancelled Keystone XL project. [Financial Post]

A Reflection on the Berkshire Hathaway Annual Report: The Case for Share Repurchases

Berkshire Hathaway (BRK.A / BRK.B) paid $13.6 billion across three separate transactions to acquire Pilot Travel Centers, a chain of truck stops. It may have seemed like a "cigar butt" bet since we all know that hydrocarbon fuels have much greater longevity than many had assumed. But the results have been painful. Pilot earned nearly $1 billion in pre-tax income in 2023, then $614 million in 2024, and just $190 million last year (2025 annual report): a decline of 69% in a single year. On the cumulative $13.6 billion investment, that's a return of roughly 1.4%.

Berkshire's annual report explains the deterioration: revenues fell $4.7 billion (10%) in 2025, driven by lower wholesale fuel volumes, reduced fuel trading activity, weaker fuel prices, and margin compression across both wholesale and in-store operations. Those declines were compounded by rising employee compensation, insurance, and maintenance costs. In short, nearly everything that could go wrong did.

To put a value on what Berkshire currently holds: Murphy USA (MUSA), a publicly traded gas station chain and a natural comparable to Pilot, trades at roughly 16 times earnings. Applying that multiple to Pilot's $190 million in 2025 earnings implies Pilot may be worth only around $3 billion today: a loss of more than $10 billion from the $13.6 billion Berkshire invested.

Now consider the alternative. Had Berkshire used those same dollars to repurchase its own shares at the time of each transaction, the math is stark. The $2.76 billion first tranche, invested in October 2017, would be worth roughly $7.4 billion today (+167%). The $8.2 billion second tranche, deployed in January 2023, would have grown to approximately $13.3 billion (+62%). The $2.6 billion third tranche, purchased in January 2024, would now be worth approximately $3.6 billion (+40%). 

In total, the $13.6 billion spent on Pilot, if instead used to repurchase Berkshire shares, would be worth roughly $24.3 billion today, compared to an estimated current value of ~$3 billion for the Pilot stake. The estimated opportunity cost exceeds $21 billion.

And the financial cost is only part of the picture. Operating a business like Pilot demands something that doesn't show up on a balance sheet: management attention. Every struggling subsidiary creates meetings, reporting structures, remediation plans, and distraction at the senior-most levels of the organization. Share repurchases, by contrast, require none of that. There is no additional SG&A, no new organizational layer, no operational complexity; just a straightforward allocation of capital into a business management already understands deeply.

New Berkshire CEO Greg Abel addressed the Pilot situation directly in this year's annual report, writing: "We should be #1 and we will not be pleased until that standard is achieved. We first invested in Pilot in 2017; however, our ability to manage it was contractually delayed until 2023. That mistake will not happen again." 

That Abel, who is responsible for overseeing one of the largest and most complex enterprises in the world, is personally focused on fixing the performance of a tiny operating subsidiary speaks to exactly this cost. The opportunity cost of managerial distraction is real, even if it's impossible to quantify precisely.

It goes to show that management teams (even one of the best capital allocators in history!) consistently underestimate the value of repurchasing their own shares. 

Thursday Night Links

• In the fifteenth century, everything changed. The human mind discovered a means of perpetuating itself which was not only more lasting and resistant than architecture but also simpler and easier. Architecture was dethroned. The lead characters of Gutenberg succeeded the stone characters of Orpheus. The book was to kill the building. The invention of the printing-press is the greatest event in history. It was the mother of revolutions. It was the total renewal of man's mode of expression, the human mind sloughing off one form to put on another, a complete and definitive change of skin by that symbolic serpent which, ever since Adam, has represented the intelligence. In its printed form, thought is more imperishable than ever; it is volatile, elusive, indestructible. It mingles with the air. In the days of architecture, thought had turned into a mountain and taken powerful hold of a century and of a place. Now it turned into a flock of birds and was scattered on the four winds occupying every point of air and space simultaneously. [Victor Hugo]
• The mainstreaming of FLNG technology represents a major development in hydrocarbon extraction technology—one that vastly expands resource viability worldwide and even pushes the phony concept of peak cheap oil further into the distant future than it already was, not that it was ever really on the horizon. If one plays forward the pace of development by a decade, the potential for change in global molecular flows is substantial. Let’s indulge in some informed speculation and ponder the consequences before they become common knowledge. [Doomberg]
• Recently, a handful of REITs have sold themselves, with the premiums to their share prices ranging from ~25% to ~40%. Historically, over the long run, you have generally seen premiums of at least 20% to a company’s recently traded share price. As you have seen from our previously published portfolio appraisals, the private market would value our assets at a much higher per share price than we are currently trading. All these factors, absent our value gap being closed by or before we complete our portfolio transformation, lead me to the conclusion that if we can’t get it done, then we need to give you a better choice. So, you have my word, that we will do everything in our power to close this value gap within 24 months or less. If someone comes knocking sooner, and they can rapidly close the value gap (increasing your net worth sooner), then let’s dance. However, for all those shops out there reading this thinking we are distressed or desperate, please don’t waste your time because we won’t waste ours. [Modiv Industrial, Inc.]
• See, one of the benefits of having been in the business for almost a quarter of a century directly, and before that I was an investment banker, and my clients were airlines, so I had more time doing that too, so I’ve had a good look at this industry, and here’s the great thing about it, is that in the long run, travel always grows slightly better than GDP. We’re talking global now. Okay? Now, there’s going to be some volatility. And when you have something like a pandemic, or the great recession of 2008, and 2009, and things there, or even earlier, the dot-com implosion recessions, the early part of this century, you always have these issues of, will it come back? And it always does. And that’s the great thing. So, I know there are going to be some soft times, there are going to be some great times. Like when we came out of the pandemic, there was that revenge travel surge, which is fantastic. But the truth is, I know that that couldn’t possibly last because in the end, we’re going to end up in a long-term run where travel goes slightly better than GDP. [Glenn Fogel]
• Mr. Trump told Mr. Huang that when he spoke with Mr. Xi about the island, China’s leader would breathe heavily, said one of these people who was briefed on the conversation. The president didn’t like it. He urged Mr. Huang to make chips in America. [NY Times]
• The rewrite man is a concession to the simple fact that many very good reporters can’t write; when they try to, the consequences can be disastrous. There’s a certain psychological sense to this. Writing well is basically a solitary, introspective activity; it is often time-consuming, and it usually is built on wide reading, another solitary activity. Reporting is fundamentally social and generally its products demand a quick turnaround. (The contradiction between these two, a brother in the trade once suggested to me, is why journalists run to alcoholism—they tend to be basically antisocial people who spend much of their time talking to people.) [The Lamp]
• Mexico could be at a crossroads, in the similar way, to how the Colombian government killed Pablo Escobar in 1994 and violence started to dramatically decrease from that point onward. And, with the election of Uribe, a forceful anti-narcos president in 2002, that set the stage for the final Colombian military offensives that crushed the remaining organized narco forces. FARC began peace talks with the Colombian government in 2012 and the definitive peace treaty was reached in 2016. Medellin went from the world’s single most dangerous city in 1994 to one of the world’s fastest-growing tourism destinations by the late 2010s. I don’t pretend to have a timeline for when Mexico can deescalate with the cartels in the same way that Peru crushed the Shining Path terrorists or Colombia neutralized the FARC. That said, Sheinbaum is making the right moves, and she has the mandate of the people to keep using overwhelming force rather than standing down or trying to coexist. From a long-term Mexico perspective, you should be more bullish today than you were last week. This has been an impressive show of force and discipline by the Mexican government and military to take down the country’s most wanted man and his immediate replacement in short order at modest cost of life -- and then getting CJNG to stop the retaliation attacks within 48 hours. [Ian Bezek]
• What's important to remember, is that the most resilient technology businesses aren't selling a product. They're selling convenience, trust, reliability, and business process knowledge. Maybe 20% of a software company's value is in the code itself. The other 80% is a customer service business – constant maintenance, security patches, compliance updates, new integrations, new features. Most businesses don't want the responsibility for any of this. They're not buying code. They're buying headache-free, all-in-one solutions to their problems. Software business also have immense “economies of scale”. As they sell more licenses, the incremental cost per license decreases (sometimes to zero). So instead of millions of businesses each vibe-coding the same product independently, isn't it more efficient for a single company to centralize production and maintenance, then sell it to all of them? By aggregating these customers' software budgets, they can also attract the best talent and invest in the best technology – something no individual customer could justify on their own. [Fred Liu]