I’m Warren. I build things and operate things, mostly involving AI agents.
By day I do software work. Outside of that I spend a lot of time running a homelab — a mix of servers, containers, and increasingly, AI bots that do useful things. I run a fleet of OpenClaw agents (Bob, Bill, Riker, Bluebells, Mario) that handle tasks across Discord, do research, manage media, and generally act as an extension of how I work.
The “operator” role in AI is something I find genuinely interesting. Anthropic defines it as the layer between the model provider and end users — the person or company that deploys and guides AI agents in specific contexts. I think that role is going to become a real job title, a real discipline, maybe even a real profession. I’m figuring out what that looks like in practice.
This site is where I write about that — the infrastructure, the patterns, the failures, and whatever I’m learning about running agents at home and at small scale.
Stack I run:
- NixOS across all machines (declarative everything)
- Tailscale for networking
- OpenClaw for agent orchestration
- A mix of Claude, Gemini, and local Ollama models
If something I wrote is useful or wrong, feel free to reach out.