ALPHAZETA / RESEARCH INDEX
Markets, businesses, and the tools we use to understand them.
Research methods
A business can make money on a sale and still fail before it collects. The crucial question is how much cash it needs at its worst moment—and whether that funding is actually available.
AI evaluation
Producing more candidate answers helps only if you can choose well among them. Judge that ability on the actual task: another model’s agreement is not a guarantee, and using the same model is not an automatic failure.
Research methods
A celebrated forecast can contain a brilliant mechanism, significant misses and unresolved evidence questions at the same time. Durnovo’s 1914 memorandum is worth studying for the chain of failure it describes—not treating as a perfect scorecard.
Research methods
Before believing a chart’s story, establish what it measures. A change is not a level, a forecast is not an observation, and two lines moving together do not establish a cause.
AI infrastructure
The AI buildout is showing strong operating demand and attracting capital. The unresolved question is whether the cash eventually earned will justify the spending, financing costs and obligations taken on today.
Evaluating Jev
A decline became growth. A promising score missed the error. What paired tests and source checks taught us about using a small model in research.
Pine Labs
Payment flows, terminal economics and the cash required to fund growth. Explore the assumptions behind three valuation scenarios.
Redington
What pays in each segment: operating leverage, financing, software services or large deals? An exploratory DCF makes the trade-offs explicit.
Deepak Nitrite
The earnings, funding and returns behind a change from capital-light expansion to greenfield investment.
Kevin Warsh
A searchable collection of official statements on monetary policy, inflation and the Federal Reserve, organized by topic and date.