Also known as rcyw. Curious about AI agents, the spaces they work in, and the systems that keep them running.
Switch on a lamp, look around, and follow a thread:
- AI / Agent — What makes an agent useful beyond a good demo?
- Agent Sandbox — How can agents explore within boundaries we can trust?
- AI Infra — What makes all of this reliable enough to build on?
If you're exploring similar questions, I'd be glad to compare notes, trade a useful link, or hear a different approach.




