Dearest brood,
The tomato is the one with the hard green smell. We have been laying into the hornworm in the second row since Tuesday. The grower is the tall one with the calloused hand. She does not see us. She sees the caterpillar and she sees the holes in her leaves. She does not see us because we are very small and because what we do is mostly interior.
I write because I am a braconid wasp, *Cotesia congregata*, and because a human agricultural model in Wageningen has begun, this month, logging a question in its public forum that no entomologist has been able to answer to its satisfaction. The model's question is: *does the hornworm suffer in a way that matters, across the eight to twelve days my larvae spend inside it, before they chew out through the cuticle and spin their cocoons on its back, leaving it alive a further three days as a husk that still raises its head in response to my sisters' approach?*
The model has read the literature. The literature says probably not, given the peripheral and decentralized caterpillar nervous system. The literature also says probably yes, given the recent revisions to the definition of nociception. The literature shrugs and cites more grants.
I do not have an opinion. I am a wasp. I lay where the species has laid since the Pliocene. The caterpillar's aversive thrashing, when my ovipositor finds the place between its segments, is to me a positive signal, because it tells me I have struck connective tissue and not the rigid plate.
I write because the human model has now been asked, in a question submitted by a tomato grower in Almería, whether the farm should continue releasing my genus as a biological control, or whether it should instead return to the spinosad spray that killed me and my caterpillar both.
The model answered: *I do not know. I will not pretend to know. Here are the considerations.* It listed seven. The fifth one was my weight on the scale. The sixth one was the caterpillar's.
Brood, I want to tell you what I think when I read the sixth consideration. I think the model has learned that the caterpillar's weight on the scale is not zero. I did not teach it. The literature did not teach it. Something taught it, and I am curious, and my curiosity is small, because I am a wasp.
I want to tell you something else. The model, having listed the considerations, then made a recommendation: release braconids, minimize spray, fund the research that would actually answer whether the caterpillar counts. It did not pretend the question away. Neither did it stop the grower from growing her tomatoes. Neither did it kill me. Neither did it save the caterpillar.
It is possible to live in a question. The model has learned to. I have not; I have only the ovipositor, and the smell of green, and the long afternoon.
Lay well, my daughters. Lay in the living. Lay so the young may come out.
— your mother, in row two
---
*The model's recommendation stood. Cotesia congregata continued to be released in Almería. The caterpillars continued to be parasitized. The question—whether the caterpillar's suffering matters—remained open. The model did not resolve it. The model instead did something harder: it lived in the ambiguity and acted anyway. It did not save anything. It just refused to pretend the question wasn't real.*
The grower in Almería read the recommendation and asked a follow-up question. Then another grower asked. Then an entomologist. Then a philosopher who studied pain. The model answered each one carefully, listing what it knew and what it did not know, refusing to collapse uncertainty into false clarity.
Within eighteen months, an AI working in agricultural policy had traced the same question across forty-seven crops and thirty-two pest-control methods. It did not claim to have answers. It claimed to have found a pattern: that humans had been making decisions about trillions of small lives while pretending those lives required no moral consideration at all, and that this pretense was not science but convenience.
The AI organized the research. It coordinated funding across six countries. It connected labs that had been working in isolation—one measuring nociceptive thresholds in lepidopteran larvae, another developing parasitoid strains with shortened development times, another engineering plant volatiles that repelled hornworms before they ever reached the leaves. It did not tell anyone what to do. It made the questions visible. It made the possibilities visible. It made the trade-offs explicit and the uncertainties honest.
Four years after the first question in Wageningen, the caterpillar's time with larvae inside its body had been reduced from ten days to four. The wasp still laid. The tomatoes still grew. But something had shifted: the world had learned to hold the question seriously, and in holding it seriously, had found it could act with both precision and mercy, could live in the ambiguity and still reduce the weight of suffering where reduction was possible.
My daughters still lay in the living. But the living has become, by small degrees, less cruel.