A few months ago, I published an article titled “Teaching My Own J.A.R.V.I.S. to Read the Markets: A Personal Experiment in Augmented Decision Making.” At the time, the idea was simple but ambitious: could I build an AI assistant that helps me make better investment decisions without replacing my own judgment? Today, I’m excited to share the next milestone in that journey. The Problem: Too Much Data, Too Little Time Every trading day, thousands of data points compete for our attention. Top gainers, unusual volume, foreign inflows, momentum shifts, sector rotations, market sentiment, technical indicators, and countless chart patterns.…
One of the most fascinating moments in building technology is when something transitions from idea to interaction. Recently, I reached one of those moments. After a series of experiments and iterations, I successfully integrated my agent system — J.A.R.V.I.S (Just A Rather Very Intelligent System) — with a macOS desktop screen feed. For the first time, the system could see what I see. Not logs.Not structured API data.But the actual visual workspace. And naturally, the very first thought that came to mind was: “What should I teach it to observe?” As a retail investor without a deep financial background,…
Every builder eventually reaches a point where tools are no longer just tools. They become collaborators. For years, my daily workflow — like many engineers — has been a combination of terminals, dashboards, scripts, alerts, logs, and documentation. Powerful, yet fragmented. Every task required context switching, manual translations of intent into commands, and constant cognitive overhead. Efficient, but never truly seamless. This reflection didn’t appear overnight. It actually connects back to something I wrote previously — an article capturing a simple but deeply relatable moment of curiosity and experimentation: “Wake Up, Daddy’s Home.” That piece wasn’t really about technology…