16-year-old builds trading bot with Jev for decisions and Opus for execution
The architecture splits into two layers: Jev handles decisions, priorities and routing, while Opus writes code, researches and calls tools; the author says the bot made $6,000 on its first day and plans to break down the workflow.
a 16 y.o. nerd from Texas took Jev + Opus 5.5 and built a bot that made $6,000 on its first day Jev is the brain. Opus is the hands. one decides what needs to happen, while the other actually gets the work done. the student built the system after realizing that most AI agents waste expensive model calls on simple decisions. so he split the architecture in two. Jev handles the decisions, priorities, and routing. Opus handles the execution writing, researching, coding, and interacting with tools. the result was a small AI workforce that could run tasks almost completely on its own. I found the workflow behind the bot, and the way these two models work together is surprisingly simple. I’ll break down the architecture next. stay tuned if you want to see how he built it. what would you make an AI workforce like this do first?
This post quotes another post: x.com/me_barnyx/status/2101976764779999381