Distributed Cognitive Systems
Research question
What is the right unit of analysis when cognition is spread across people, models, agents, tools, memory, and machines?
Working thesis
The unit of analysis is the distributed cognitive system, not the individual model or agent. Cognitive topology — who thinks — and authority topology — who decides — are separate structures, and conflating them is where coordination cost, authority leakage, and unbounded failure propagation come from.
A thesis is the lab’s current position, not a result. What supports it is listed below; what remains unresolved is stated at the end of this page.
Publications
2 attributedApparatus
3 systemsSystems attached to this program are what the lab builds to investigate the question. Maturity, evidence level, and current activity are recorded separately on each system — a system attached here is not thereby claimed to be production-ready.
X-Router
Capability Routing Layer
A capability-routing layer that selects models, tools, agents, runtimes, and execution paths based on capability, cost, latency, authority, and risk.
TAG
Typed Agent Gateway
A typed delegation protocol that models every handoff between agents, tools, models, and humans as a structured object with explicit authority, scope, budgets, and evidence contracts.
DCS
Distributed Cognitive Systems
A research program studying cognition distributed across humans, models, agents, operators, tools, memory stores, machines, and control planes — and the authority structures that govern it.
Open questions
Published as open. These are the parts of the program the lab does not consider settled.
- 01
How is coordination cost measured well enough to compare two topologies?
- 02
Which decompositions survive contact with real work, and which only look clean on a diagram?
- 03
How does failure propagate across a delegation boundary, and what contains it?