·Research·5 min
Control Planes for Autonomous Companies
Why autonomous firms need explicit control surfaces, not just agent loops and prompt chains.
Archive
Everything published by the institute.
·Research·5 min
Governance After Human Labor
As firms reduce human labor in the execution layer, governance has to move from managerial supervision toward policy, monitoring, and intervention design.
·Research·6 min
The Firm as an Autonomous System
A company should be understood not as a legal shell with employees inside it, but as a coordinated execution system with memory, goals, and control loops.
·Tools·3 min
OpenClaw
A local-first agent operating system that provides memory, messaging, browser control, nodes, and automation primitives useful for autonomous-company infrastructure.
·Tools·3 min
Paperclip
A useful reference point for builders thinking about agent tooling, execution environments, and practical infrastructure around autonomous work.
·Guides·3 min
Building an Agent Memory Layer
A practical guide to the difference between stateless model calls and a usable memory substrate for autonomous work.
·Guides·3 min
From Agents to Operations
The architectural shift from isolated agent demos to company-level execution systems.
·Blog·8 min
Launching the Institute Site
Why the field needs a public home that combines conceptual seriousness with builder usefulness.
·Blog·4 min
Why Builder Utility Matters
Institutional seriousness is good, but the site only matters if it is genuinely useful to builders.
·Blog·3 min
The Conversation No One Wants to Finish
AI Twitter will talk about the underclass problem for exactly three posts before pivoting to acceleration. That's not enough.
·Research·4 min
The Underclass Question
Autonomous companies accelerate a problem most technologists prefer to hand-wave: what happens to the people whose labor is no longer needed.
·Research·5 min
The Alignment Tax
Building autonomous companies that behave well costs more than building ones that don't. Who pays for the difference, and what happens when no one does.