AI, Labor, Capital — Follow the Incentives
Looking at artificial intelligence through the lens of labor and capital incentives — and why understanding human motivation may be more useful than forecasting model capability.
A complete archive, listed from most recent to oldest.
Looking at artificial intelligence through the lens of labor and capital incentives — and why understanding human motivation may be more useful than forecasting model capability.
Why reducing an extended position can improve probability-weighted returns — and how thoughtful calibration, tax structure, and valuation discipline compound over time.
How AI-generated misinformation increases the value of trust infrastructure — and why businesses like Moody’s and S&P Global may be more essential, not less.
Why AI increases operating leverage and durability for protected royalty businesses like FICO, Moody’s, and S&P Global.
Why most long-term mistakes come not from what decisions cost, but from what they quietly crowd out across investing, work, and life.
What actually changes for investors in a U.S.–China-dominated world, what doesn’t, and why business quality matters more than macro narratives.
A first-principles breakdown of what actually moves Canadian housing prices over time.
Why operational excellence alone isn’t enough, and how valuation, scale, and capital allocation shape outcomes.
What mainstream financial advice is designed to do, and why it often produces average outcomes by construction.
A calm long-term case: ecosystem lock-in, capital-light economics, and disciplined capital allocation.
Closed-loop economics, premium positioning, and durable advantages through cycles.
An update on the COKE thesis: what worked, what changed, and what I learned.
Dividends are a side effect, not a goal. Compounding depends on reinvestment and capital allocation.
A practical framework for long-term compounding with disciplined psychology and concentrated bets.
Why concentration can outperform if you can handle volatility and stay rational.
Distribution density and embedded workflows as a durable compounding engine.
Capital-light economics and why the best compounders often require less capital than you’d expect.
AI’s next bottleneck is power, and MPS is built to solve it.
Culture and data create durability, and underwriting discipline compounds.
Standards plus switching costs plus disciplined execution as a steady compounding machine.
FICO is still a great business, but at a demanding multiple I’m holding, not buying.
Selling too early kills compounding; real strength is staying detached from price anchors.
Valuation, improving business quality, and local moats.
Power-management silicon and the design-win flywheel into the AI era.
Decisioning as a standard, pricing power, and platform transition.