BLOG
Project updates and engineering notes
I BUILT A 39-ROLE AI RESEARCH ORGANIZATION — ITS FIRST RUN KILLED EVERY IDEA
I built a multi-agent research organization with executable quality gates. Its first full run abandoned every candidate thesis. So I ran a pre-registered, blinded calibration study to find out whether the gates discriminate between sound and flawed research — or just kill everything.
Jun 2, 2026THE VALIDATION CRISIS PAPER IS DONE — AND MY OWN PREDICTION FAILED
The short version of the validation-crisis paper is finished: a 21-page arXiv preprint. Across the surveyed AGI forecasts, deflation factors cluster between 1.28× and 2.02× — and the preregistered self-prediction I locked before computing anything failed at 1.285×.
May 14, 2026STARTING THE VALIDATION CRISIS: AUDITING AGI FORECASTS WITH QUANT FINANCE TOOLS
Beginning a research project that imports the overfitting-detection tools quantitative finance built between 2014 and 2018 — and applies them to the major AGI timeline forecasts shaping how capital and talent get allocated.
Apr 5, 2026X_QUANT V2: FROM ML PIPELINE TO AUTONOMOUS AI HEDGE FUND
How X_Quant evolved from a 90K-line quant platform to a 299K-line autonomous trading system with 20 AI agents, a 5-member Investment Council, regime-adaptive fusion, and 9 layers of risk control — achieving Sharpe 1.708.
Feb 28, 2026MISTRAL AI WORLDWIDE HACKATHON — BUILDING QUANTBRIEF IN 48 HOURS
How I built a multi-agent market intelligence platform at the Mistral AI Hackathon 2026 — 5 specialized AI agents, 256K context SEC filing analysis, and multilingual audio briefings.
Feb 15, 2026STARTING KELLY PM: OPTIMAL POSITION SIZING FOR PREDICTION MARKETS
Deriving the Kelly Criterion for prediction markets from first principles — information-theoretic foundations, the simplified market form, and why fractional Kelly is essential.
Feb 11, 2026BUILDING THIS SITE: A TERMINAL-AESTHETIC PORTFOLIO
How I built my personal portfolio with Next.js, Tailwind CSS, and a dark terminal aesthetic inspired by Geist Pixel.
Feb 9, 2026GOOGLE SWE L3 — FIRST ROUND TECHNICAL INTERVIEW
Notes from my Google Software Engineer L3 first round interview: coding challenge (similar to LeetCode #394), behavioral, and Googleyness rounds.