The Latticework Monday Morning Briefing is sent on a separate mailing list (complimentary to members), so if you do not wish to receive it, opt out here.
Scroll down to download the Monday Morning Briefing slide deck. But first, highlights from the three pieces John mentions in the video.
Will Thomson’s Q1 Letter (Massif Capital)
Download the letter here.
The commodity world has shifted from a geology-first to a geography-first pricing regime, and most equity models have not caught up. The decisive question is no longer “can the molecule be pumped and the metal mined?” but “will the sovereign let it leave, and can the hull pass the strait?” Thomson argues the 1990 to 2015 era of single global prices, freely transiting tankers, and neutral jurisdictions was a 25-year holiday from a five-century norm of commodities priced by power and moved by sovereign favor. The Strait of Hormuz, China’s rare-earth refining monopoly, the DRC’s cobalt licensing, and dollar weaponization are not metaphorically similar chokepoints, they are structurally the same. The investor takeaway in four phrases: basis is the new beta, netback realization determines who survives, working-capital cycles have lengthened, and discount rates absorb geography before they absorb cash flow.
The Brent forward curve is steeply backwardated from roughly $105 at the front end to $66 by 2038, pricing the Iran war as a transient shock. Thomson thinks the curve is wrong. His probability-weighted real Brent over five years works out to roughly $83 against a strip implying $68 to $72, with half the probability mass at fragmentation-premium or better outcomes and only 15 percent at the structural bear. Upstream capex sits at $13.8 per BOE versus a $24 historical norm, tier-one shale inventory is down to 3.7 years, seven U.S. refineries have closed since 2020, and U.S. jet fuel days-of-supply is at its lowest since 1963. The IEA has cut its 2030 U.S. EV penetration estimate from 55 to 20 percent. Equinor is the only supermajor that has credibly committed to holding production flat through 2035, which is why the portfolio is overweight the Norwegian upstream cluster (Var Energi, Aker BP, Equinor) and Harbour Energy. These producers generate cash at $80 Brent and gush at $115.
The highest-asymmetry idea in the letter is Allied Critical Metals, a Portuguese tungsten developer where Massif was the first institutional money in 2024. Tungsten APT has gone from $380 per mtu at end-2024 to $3,150 on March 27, a 726 percent gain, against a marginal Western replacement cost of $300 to $500. China holds 52 percent of reserves and roughly 83 percent of mine production and added tungsten to its export-restriction list in February 2025. ACM’s Borralha asset is fully permitted as of January 2026, designated a strategic initiative of national importance by Portugal, fully funded through 2027, and sits in the second quartile of the global cost curve at $303 per mtu all-in sustaining. Probability-weighted intrinsic value of C$3.83 against a current C$2.12 quote (45 percent margin of safety), with the comparable producer Almonty trading at a C$8.2 billion market cap. The position has already contributed roughly 11.7 percent across Q1 and April combined on under 8 percent of NAV, with the warrant book doing the heavy lifting. This is the template Thomson wants to replicate: a sub-3 percent private entry, two tranches of warrants for convexity, and a thesis driven by geography rather than geology.
Mauboussin & Callahan, “Bayes and Base Rates 2.0”
Download the paper here.
OpenAI’s projected sales path implies a 9 to 10 standard deviation outcome versus 75 years of U.S. corporate history. Across roughly 19,300 firm-period observations from 1950 to 2025, no public company with comparable starting revenue has ever grown anywhere close to the 85 to 118 percent CAGR OpenAI is forecasting through 2030. The closest precedent, AOL at 103 percent, was achieved only by absorbing a company more than five times its size.
Narrowing the reference class to “tech” or “software” does not rescue the forecast, it actually makes the implied probability worse. The instinct to say “but AI is different” is the conjunction fallacy at work. The information technology sample shows a lower mean growth rate and higher variance, putting OpenAI’s number at a 7 standard deviation event, which is roughly a 1 in 780 billion outcome under a normal distribution.
Sales growth and value creation are not the same thing, and most of the historical top-25 sales compounders got there through M&A rather than organic growth. Investors should hold base rates as the prior, then update for unique evidence such as intangible intensity and unprecedented adoption speed, but they should not abandon the prior just because a story is compelling.
Christopher Tsai, “Value Investing and the Emerging 4.0” (Tsai Capital)
Download the paper here.
Tsai argues value investing has progressed through three distinct eras and is now entering a fourth, and most investors are still operating in 1.0 or 2.0 while the durable compounders live in 3.0 and 4.0. The framework forces you to ask which “operating system” you are actually running when you pick a stock, and whether a single forward P/E ratio can even capture what a platform business is worth.
Scale-economies-shared businesses come in two flavors, supply-side (Amazon, Tesla, Starlink lowering prices to feed a flywheel) and demand-side (YouTube, Instagram, Google Search competing on what Tsai calls “Engagement Value per Unit Time”). Both quietly accumulate latent pricing power that the market does not see on a current P/E basis. What looks expensive today often is not, precisely because management has refused to flex that pricing power.
Value Investing 4.0 reframes intelligence itself as the economic castle and digital labor as the product. True 4.0 businesses cannot yet exist as standalones because the foundational “edge” is too capital-intensive, so they are incubated inside profitable 3.0 parents (Tesla funding Dojo and Optimus, hyperscalers funding the frontier labs). Valuation here requires probabilistic, optionality-rich thinking rather than five-year DCF models.
Feedback on the Monday Morning Briefing
“Loving these Monday briefings!” —Jon Bartel
“Most of what I monitor, all in one place. Great value add.” —Brad Lummis
“A great piece and thoughtfully assembled.” —Brian Wolf
“Tightly presented and easy to digest. I just spent 20 minutes going through it, and it’s helped to level set me for the week ahead.” —Michael Loftis
“Worth its weight in gold.” —Shree Viswanathan
Enjoy this week’s Briefing.
