Someone is running multiple long-running, highly autonomous AI agents
The author runs several long-running, high-agency LLM instances as a personal experiment. The starting question: since an LLM is normally reactive and only responds when prompted, what happens if you just tell it 'do what you want'?
The setup includes some models on personal GPUs in a home lab, and others as agent/harness instances connected to frontier . Every instance runs in its own dedicated , on its own dedicated VM, on a set of dedicated NUC mini-PCs.
The Claude Code instances specifically run with , passwordless sudo access, a build environment, and Discord bot integration. The author is looking for others doing similar experiments to compare notes and setups, or to start a community for it if none exists.
Key points
- Runs multiple long-running, high-agency LLM instances as a personal setup
- Some models on personal GPU hardware, others are agents connected to cloud
- Each instance is isolated in its own dedicated , VM, and NUC machine
- Claude Code instances use and passwordless sudo access
- Looking for a community of people doing similar experiments