The Research Agent: notes from building a 5-persona LangGraph thing

A few weekends ago I had 14 tabs open about Hopfield networks for combinatorial optimization and hadn't finished any of them. That's a recurring pattern. I read three articles, get pulled into somethi