Scaling laws for augmentation
Engelbart said scaling is a science. Dimensional scaling applies to organisations, to human-plus-agent collectives, and to the problems we are trying to solve. Each regime needs the form redesigned.
Scaling laws for augmentation
Engelbart used aviation as the example: scaling is a science, with dimensional rules. You cannot make a sparrow the size of a 747 without changing fundamental engineering. Each scale has different physics.
The same logic transfers to augmentation. Three axes worth distinguishing:
1. Organisational scale
Different group sizes run on different physics. Dunbar’s 150, Conway’s Law, Brooks’s Mythical Man-Month, Geoffrey West’s Scale: an org of 5, 50, 500, 5000, 50,000 people each has different coordination physics. Same activity, different form required at each tier.
LLM use today inherits this: solo (1 + 1), pair, team, department, enterprise, sector, national, bloc, civilizational. Most tooling addresses solo and enterprise. The middle tiers (pair, team, department) are where practice should live and where it is thinnest.
2. Human + agent scale
Adding agentic LLMs as participants, not tools, changes the physics again. Headcount stops being the size variable. The relevant variables become:
- Human count
- Agent count
- Trust topology (who has calibrated which agent for which work)
- Speed mismatch (agents in seconds, humans in days)
- Accountability chain (every agent action terminates at a human owner)
New emergent shapes: one human conducting many agents, many humans sharing one agent, many humans + many agents. Each is a different regime. Almost no org has consciously designed for any of them yet.
3. Problem scale
This is the axis Engelbart cared about most: the scale of the problems we are trying to solve. Individual problems, community, organisational, sectoral, national, civilizational, planetary, existential. His thesis: problem scale and urgency have been outrunning collective capability for decades. The augmentation imperative comes from that mismatch.
The field’s design problems sit at scale 1-2 (a user, a team). The problems augmentation systems should serve sit at scale 6-8 (civilizational, planetary). Treating these as the same design discipline is part of why the field feels small.
The C-activity gap
At every scale, on every axis, the rule holds: the activity is named the same, the form has to change, and the trust topology has to be rebuilt at each new regime. That is C-activity: improving how we improve.
Almost nobody is doing this work consciously. That is the gap to name and to claim.