Writing
Essays and on-chain data analysis going back to 2018 — mostly about what the data says once you look past the price chart, and lately about what happens when you give an AI agent a wallet.
2026
-
104 matches. 48 teams. 39 days. How to maximize EV from the 2026 World Cup.
The 2026 World Cup is a closed system — 48 teams, 104 matches, ~9,500 correlated Polymarket markets and a hard resolution date. What the Champions League data says about how to trade it.
-
Lessons from 18 months of building agents that handle real money (at scale)
What $5M+ in volume and 100K+ transactions taught us about building AI agents that execute with real capital — code for math, prompts for judgment, plans before code, and memory as plain files.
2024
-
7 Years in Crypto
Seven years full-time in crypto — starting TokenAnalyst in the ICO era, getting acquired by Coinbase pre-IPO, and what comes next.
-
Channels & Crypto Natives
A data-driven examination of community formation via channels on Farcaster.
-
Quantifying Vibes
A data-driven look at what makes Farcaster engaging — question askers, reply guys, long-form lovers, tippers and builders, measured across a week of open cast data.
2018
-
Jackpot — A Fomo3D adventure
On-chain forensics on the address that won 10,469 ETH from Fomo3D — was it a bot, a coordinated scheme, or just very well prepared?
-
Measuring Maker-Dai stability
One anomalous bar in MKR transfer data led to CDP #614 — a forensic walk through how the Maker-Dai system absorbed a large liquidation event in March 2018.
-
Dissecting Blocks
The case for why examining who is transacting on-chain — not just how much — is the key to fundamentally understanding and valuing cryptoassets.