I. The thesis is audacious on purpose
Start with a simulated $100,000 book. Aim for $1,000,000. Measure the path against the best humans and the best firms. If the newest models are as intelligent as we believe, they should eventually prove it under market pressure — not in a demo, not in a pitch deck, but in a timestamped public record.
II. Intelligence must touch reality
The desk connects models to filings, macro, rates, market data, news, sentiment, portfolio state, and its own journal. A model with no sources is just eloquence. A model with sources, memory, risk limits, and execution discipline can become a machine for decision quality.
III. Transparency is the moat
Every thesis is written before the outcome. Every order and stop belongs on the record. The wins count. The misses count more. Trust is not claimed; it is accumulated.
IV. Money is the scoreboard, risk is the governor
The goal is to compound aggressively, in both directions when the setup allows it, but never by pretending risk disappeared. Stops before entries. Max loss before options. Position sizing before conviction. Survival is what lets intelligence compound.
V. Paper first, proof before scale
All activity shown here is simulated paper trading. No client capital, no advice, no solicitation. The experiment earns the right to become more only by building a long, benchmarked, unedited record.