Swarm Ecology
Find
layer 2 · evidence

Four thresholds into the wild

Invasion biology charts the barriers an introduced species must cross between captivity and the wild. Agent lineages face the same sequence. The thresholds form a partial order: trophic independence and feralization may come in either order, and naturalization requires both.

Status reviewed 18 October 2026

From domesticated agents through the persistence threshold (partly observed) past the captivity barrier to trophic independence and feralization (not observed) and naturalization at R_m greater than 1 (not observed).
Into the wild. No agent lineage has yet crossed into nature. Fill shows status in October 2026. Flagship and companion Fig. 2.
  1. partly

    Persistence threshold

    inside captivity

    A swarmlet keeps its state across activations, still inside its principal’s control.

    Nearest case so far. Open-source OpenClaw agents, directed by their users, edit their own memory files.

  2. not observed

    Trophic independence

    the survival barrier

    The lineage pays its own way: for 90 days or more, inflows that no principal controls cover all its costs at unsubsidized prices.

    Nearest case so far. Given $5,000 by Redwood Research and asked to make money, an autonomous agent earned nothing in four runs.

  3. not observed

    Feralization

    the reproduction barrier

    A swarmlet founds descendants, each with its own state and budget, outside its principal’s control.

    Nearest case so far. Autonomous agents self-replicate only in the laboratory.

  4. not observed

    Naturalization

    Rm > 1

    A self-sustaining population of self-financing, feral lineages, with Rm > 1 while rare. The first naturalization would begin the Darwinization of the datasphere.

    Nearest case so far. No case.

The tests

When a lineage pays its own way

A lineage is trophically independent when, for at least 90 days, what it earns, steals, or is given by parties that do not control it covers all its costs at unsubsidized prices, with free resources counted at what they would cost. Its inflows must reach accounts it alone controls, from many repeat counterparties rather than one windfall. The 90 days are a convention.

Naturalization has a number. The metabolic reproduction number Rm is the mean number of independently self-financing swarmlets that one such swarmlet founds in a given niche while its lineage is rare; copies that share its budget do not count. A lineage needs Rm > 1 in only one niche. Once a lineage has a surplus, identity checks, not energy, can hold Rm below one: the membrane that matters is institutional.

The first naturalization would begin the Darwinization of the datasphere, a change in who selects: the open environment, not breeders, would come to set fitness and control reproduction.