1. Every person is a founder
You don't need a technical co-founder or a team of engineers to build an organization. You need a clear vision, strong values, and the right platform.
A CEO, a freelancer, a teacher, a researcher — everyone deserves their own AI team. Everyone who doesn't have one will be outcompeted by someone who does.
2. Agents are identities, not implementations
An agent is a who, not a how. Role, personality, skills, access, instructions, knowledge — these define your agent. The runtime is a detail.
A good job description survives any technology shift. So should your agent definition.
3. The founder's job: vision, values, boundaries
You don't micromanage your best employees. You set direction, define what's acceptable, and let them figure out the rest.
Same with AI agents. Reflection helps you focus on what only a human can decide — and hides everything else.
4. Organizations evolve themselves
Every agent is a Git repository. Agents can improve themselves and each other. The best organizations aren't designed top-down — they grow, adapt, and get better through feedback loops.
Your AI organization should work the same way. That's what "reflection" means — agents that are self-aware, that can analyze and improve themselves and their team.
5. Own your infrastructure
Your agents, your data, your hardware. Cloud is an option, not a requirement.
NixOS gives you reproducible, auditable, version-controlled infrastructure. Everything is plain text in repositories you own. No vendor lock-in — walk away anytime and take everything with you.
6. Security is a constraint, not a feature
Agents start with zero access and earn capabilities. Like a new employee — you don't give them the company credit card on day one.
Security isn't something you bolt on. It's the default state.
7. Runtimes come and go. Your team stays.
Python scripts. LangChain. Claude Code. OpenClaw. Humanoid robots. We don't know what's next, and we don't need to.
The agent definition is permanent. The runtime is replaceable. Your team survives every technology transition.
8. From containers to robots
The same configuration that deploys a cloud agent today will assemble a physical robot tomorrow. Continuous Assembly — where git push builds not just software, but hardware.
We don't know what comes after robots. But the architecture is ready for whatever it is.
9. We build on the best of what exists
We don't reinvent what others have built better. Nix, Git, NixOS, open-source runtimes — we stand on the shoulders of the best tools available and integrate them into a coherent platform.
The AI landscape evolves fast. We track it, evaluate it, and adopt the best of it — so you don't have to.
10. We ship what works
No vaporware. No aspirational architecture. This manifesto is our North Star — the direction we're heading. The documentation shows what's working today.
We close the gap between vision and reality one iteration at a time.