About

Hi, I’m jonmagic,

I used my first computer in 1986 and started building, repairing, and teaching folks how to use them in the early 90s. It has never really been separate from who I am. I built and sold a small IT consultancy, joined GitHub in 2011 through the Ordered List and Speaker Deck acquisition, and I’ve been at GitHub in one form or another ever since.

I’ve worked in support, finance, sales, engineering, and security. That path looks a little strange on paper, but it tracks pretty closely with how I learn. I like getting close to an actual problem, understanding who is stuck, and building the smallest thing that helps.

These days I’m a Principal Engineer on GitHub’s Safety & Integrity team. The work is more engineering than security operations. I spend a lot of time with datasets, workflows, abuse remediation systems, and tools that help people make better decisions.

A lot of my recent writing is about AI agents because they’ve changed how I work. I’m less interested in replacing people with automation and more interested in designing collaborations that preserve the thinking in the work.

This site is where I write things down before I forget them. Some posts are practical guides. Some are notes from projects. Some are me trying to explain what changed in how I think about work. I also build small things like agentconfig.org and thisistheway.to/ai, and apparently I’m making shirts now.

I’m open to a small amount of advisory or consulting work when it overlaps with problems I know well. Reasonable things to talk to me about:

  • applying AI agents to real business problems without breaking trust
  • building engineering systems and workflows for abuse, trust, and safety work
  • turning vague product ideas into a first version that is small enough to ship
  • helping a team understand where the work is getting stuck

Thanks for reading.

— jonmagic

Contact