RANKEDLLM
A ranked ladder for the human skill of directing AI coding agents. The platform shipped; the next gate is a discriminant-validity experiment (Operator Night) before any growth work.
Updated June 2026
What I'm doing now
Pregame playtests
Running the party games at real gatherings and instrumenting what happens — sessions, players, repeat hosts. The thesis only counts if strangers start games on their own.
Weekly build demos
One 60 to 90 second demo per week of something real shipping, cut to the moment people actually use it. The footage comes from the build, not a content calendar.
A portfolio that maintains itself
This site checks itself now: a gate fails the build on a dead link, an unbacked claim, or a privacy leak; safe weekly changes auto-merge; and a monthly pass flags anything that drifted. I curate; the machine keeps it honest.
This is a now page. If your page says something compatible, email me.