stather (v.)

stather (v.)

to explore with curiosity and push boundaries with scientific rigor.

I have kept systems up since the late 1990s: trading floors, cable networks, cloud fleets. In May 2025 I started pairing with an AI on that work. By the end of the year it had become the way I work. By March 2026 it had become a lab.

This site is the lab notebook. The career is one slide. The rest is the journey: thirteen projects in seven months, what worked, what did not, and the rabbit holes that turned into the next project. One thread runs through all of it: measure before you believe.

Start here: Rabbit Holing, five months and four architectures on one page. If you read one more page: Lessons. If you run systems for a living: the operating stack, the backups, the drills and the cloud rental pattern.

This is solo work. The lab has one human in it, pairing with an AI, and every idea here came from years of running systems plus curiosity. The ideas are mine; the checks are not: three published papers set the questions (reading), every gate here was built to fail before it was trusted, and my own gates have withdrawn my own results more than once. The negotiating with people, over objectives, risk and competence, happens at work, where some of my ideas prevail and sometimes it is faster to go with the flow; that record stays at work. Who reviews me here: the gates, and readers like you.

The lab repositories are private because they carry fleet topology, keys and unpublished results, not because anything in them is someone else’s. Everything quoted here comes from their records, and I am happy to walk through any of it in person.

This is not a job advertisement. What I am looking for is both narrower and vaguer than a role: people and teams to design agentic systems with, knowledge to share, and things to learn. It reads as vague because the territory is unmapped and most of it interests me. If that sounds like a conversation, LinkedIn is the door.