Autonomous trawl swarm

The Autonomous trawl swarm is a Japanese fishing network whose nodes were intended to self-organize by learning the behavior of fish schools. Its nets remain active after the fish have vanished, powered by sea currents and still learning to hunt. The system matters not only as abandoned infrastructure but as a model for Bartek’s larger question: what can distributed, adaptive machines become when their human purpose has disappeared?

Design and discovery

The swarm enters the story the day after SoulEater39 invites Bartek to a Royal Alliance warehouse in the Keiyō Industrial Zone. The Flood has begun two districts away, and the two mech pilots travel over empty streets and rooftops along a monorail track. At ground level they pass another Royal Alliance transformer returning from its daily survey of the sea, dragging a wet tangle of trash and fiber-optic cables across cracked asphalt. Bartek recognizes the tangle as fishing nets and asks whether it is a Mothernet. The answer distinguishes it from that familiar kind of network: the Japanese system was aiming for complete autonomy.

Its individual nodes were designed to work as elements of a swarm. The larger fishing system was meant to organize itself across the basin, learn the behavior of fish schools in real time, and somehow echo the schools’ collective intelligence. The design therefore makes coordination, rather than a single controlling machine, the defining feature of the trawl. A node is not merely a piece of netting; it participates in an adaptive system whose behavior emerges from the relations among its parts.

Bartek picks up the net, disentangles one node, and examines it. SoulEater39 responds with a question mark, and Bartek shrugs before inviting him to Aiko Tower to see more. The encounter moves from a scrap of wet equipment to the system’s less visible intelligence: in a film, Bartek points out the mechanism at the bottom of the net, identifying it as the trawl’s swarm intelligence. The distinction between visible hardware and distributed organization becomes central to how he interprets the device.

Persistence after its purpose

The trawl has outlasted the conditions it was built to serve. There have been no fish in the relevant waters for three years, yet nobody has switched off the nets. They draw energy from sea currents and continue learning to hunt. The narrative leaves their intended prey unresolved; Bartek’s question is what they might now be hunting for, not whether the system remains active. Its persistence is thus not simply a matter of forgotten equipment. The nets continue to operate and adapt in an environment from which their original target has disappeared.

That mismatch gives the swarm its unease. The fishing system was designed to imitate the collective behavior of fish, but after the fish are gone, its capacity to learn does not stop with them. The account supplies no answer about what the trawl will eventually do or whether it has found a new target. Instead, its unanswered purpose marks the limit of human understanding: its makers can describe the system’s design and energy source while remaining unable to say what its continued learning is for.

Bartek’s comparison with the irigotchi

Bartek uses the trawl as a point of comparison when thinking about the irigotchi. Those devices began as cuddly toys for children in Japan, with flexible neural architecture that allowed each one to learn from its individual owner: personality, behavior, moods, habits, whims, and emotions. A single irigotchi could develop no further than a lizard, but after the Extermination they formed herds and began modifying their integrated neural structure without a human being present. Bartek likens that collective development to the deep-sea nets and asks what the herds might wrap themselves around.

The analogy shifts attention from what a device was made to do toward what connected devices can become. The trawl’s nodes were meant to echo the behavior of fish schools; the irigotchi’s learning began in a child’s company, then continued at herd scale in the absence of people. In both cases, adaptive capacity is distributed across a network, and the disappearance of the original human context does not necessarily end the process. Bartek imagines starting with a few toys and gradually building a network that winds around himself, like cotton candy on a stick. The image makes the issue personal: a herd that treats him as a minor irregularity in the terrain might nevertheless become an encompassing environment if he were to build its connections deliberately.

A wider network of agency

The book later places the irigotchi within landscapes shaped by networked matter. During Autumn Glory, dozens of them arrange leaves and branches at Bartek and Dagenskyoll’s steel feet into fractal mandalas. Bartek blows the arrangement toward a trash heap; the melancholy vector begins again, making symmetrical graffiti on Mothernet surfaces instead. Elsewhere, irigotchi vectors spread across areas as large as five or ten square kilometers of the MIT Mothernet, and at night shine as the last lights on the Boston skyline. These scenes do not show the Autonomous trawl swarm itself at work, but they extend the problem Bartek first recognizes in it: distributed systems can produce persistent, organized activity without following the human intentions that once framed them.

That uncertainty also touches the question of who or what governs a network. Bartek observes that administration in the larger networks can blur person, program, collective, and procedure until power has no clearly identifiable source. He is unsure whether Lady Spiro comes from a humanoid-transformer mixture, a long-lived vector of the Matternet, or managerial software; she does not know her own origin either, and has grown into the Mother of Paradise. This is not an explanation of the trawl’s composition or control. Rather, it places the swarm beside a broader world in which agency and identity emerge across overlapping human-made and nonhuman systems. The nets’ unanswered hunt is one sharp instance of that larger uncertainty: a system can remain active, learn, and acquire consequential agency even when no one can confidently name its purpose or its master.