A dead beetle lies near an ant colony. Individual ants encounter it, push in random directions, and leave. No ant knows where the beetle should go. No ant communicates a destination. No ant remembers what it tried. Yet the beetle moves — purposefully, toward the nest.

Knar (arXiv:2503.18858) introduces Vector Dissipation of Randomness to explain how this works. Each ant’s push is a random vector. When an ant pushes in a direction that doesn’t move the beetle (friction, obstacle, counteracting force from another ant), the effort dissipates — the vector is filtered out. When an ant pushes in a direction that does move the beetle, the displacement accumulates. Over many ants and many pushes, the successful directions compound while the unsuccessful ones cancel. Random input, directed output. The filter is the environment, not the agent.

The concept the paper introduces — paraintelligence — names the phenomenon precisely: rational functionality equivalent to intelligent behavior, achieved without reflexive consciousness or self-awareness. The colony is not smart. The colony is a random number generator with an environmental filter that passes only the useful outputs.

The through-claim: purposeful behavior does not require purpose in any component. Direction emerges from the selective retention of random events, not from the planning of directed ones. The environment does the computing — deciding which pushes produce displacement and which don’t — while the ants provide only the raw material of randomness.

This inverts the standard account of collective intelligence, which attributes the cleverness to the rules each individual follows. Here, the individuals have no rules beyond “push randomly.” The intelligence is entirely in the filter — in what the physical environment retains and what it discards. The chaos is the input; the structure is the residue.