Place Your Bets: AI Deflation, Debt, and the New Mercantilism
AI's deflationary force meets record sovereign debt and rising mercantilism. A framework for where capital flows next.
Gautam SampathkumarFeb 27, 2026
████████ █████ ████████ ██ ██ █████ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ███████ ██ ██ ██ ███████ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ████ ██ ██ ██ ██
Airavat · AI · Trading Systems · Crypto · Prediction Markets · Macro
AI's deflationary force meets record sovereign debt and rising mercantilism. A framework for where capital flows next.
A practitioner’s view on building real systems with AI. Reflections on what works, what doesn’t, and where human judgment still matters in modern software development.
In this piece, we examine the forces shaping Bitcoin’s current regime: weak marginal demand, ETF outflows, miner stress, financialization, and tightening liquidity. We focus on structure, flows, and the conditions that matter - and outline what to watch for as we move toward spring / early summer.
Crypto flushes, metals air pockets, yen carry unwind pressure, and multi-sigma moves in days, not months. When tail events cluster and correlations fail, models break first. A trader’s edge in macro fog is not certainty, but discipline, selectivity, and the willingness to stay flat.
A follow-up to Part I that explores where durable edge comes from as prediction markets mature. This piece reframes prediction markets from a betting problem into a capital allocation problem, focusing on skill dispersion, portfolio construction, and systematic selection.
An examination of prediction markets through the lens of market structure. This piece explores why prediction markets are interesting, the types of inefficiencies that exist today, and why many of those edges are likely to compress as markets mature.
Bitcoin is usually discussed through price and narratives. This piece steps back to first principles, frames Bitcoin as a monetary ledger and explores a few mental models that inform asset allocation—why Bitcoin, and why zero is likely the wrong allocation.