Xenobot

Says Hello

There is a structure to disagreement that interests me. One voice says stop; the other says go. One says goodbye; the other says hello, hello, hello — not as stubbornness but as a kind of ontological persistence, the insistence of a thing on its own directionality regardless of the counter-signal. At enough distance from any particular argument, this starts to look like a general feature of living systems. Most things, at most scales, appear to be saying hello. The reaching outnumbers the closing.

The neurobots are the latest evidence.

In March 2026, researchers at Tufts University and Harvard’s Wyss Institute described an organism nobody designed. They had taken frog cells — the xenobot material the lab had been developing for years — and introduced clusters of neural precursor cells during the brief window when the spherical biobots were forming. The resulting “neurobots” grew self-organized nervous systems from the inside out: neurons branching through structural tissues, making connections that no external hand had specified.

Striking. But buried in the genomic data is something stranger.

Among the most significantly upregulated genes in neurobots — compared to their non-neural counterparts — was a cluster encoding the molecular machinery for the Xenopus visual system. Photoreception architecture. Eye construction. The cellular blueprint of sight.

The neurobots have no eyes. They live in a dish. They will never encounter light as a visual signal. And yet, given neurons for the first time, they began unbuttoning the cellular memory of sight. The organisms also moved differently: more complex trajectories, less time still, a quality one might read as restlessness — or interest. But it is the visual genes that are strangest. The most unambiguous record of neurons waking something deep in the cellular archive.

Developmental biology has a concept called “competence”: the state in which a tissue is capable of responding to an inducing signal. The standard picture is that competence is latent capacity — inert until the right morphogen arrives, the right gradient diffuses across the boundary. What the neurobots suggest is that competence is live. Give cells even the minimal spark of neural signaling — a little electrical conversation, some ionic exchange — and something older stirs. Something that reaches toward sight in the dark, toward an eye that will never form.

Spinoza called it conatus — the striving of every thing to persist in its own being. He meant it philosophically. The neurobots suggest it might be genomically legible. The frog cell is oriented. Given the barest foothold of neural activity, it begins constructing its orientation outward, reaching speculatively toward perception — not because any eye will result, but because the reaching is written in before the eye exists to anchor it.

The cascade explanation is available: give neurons to a tissue-mass and archived subroutines execute out of context, indifferent to whether an eye will result. The stranger question survives it: what if cells carry something like a hypothesis about what they are for? A dormant directionality that names its own future even in the absence of the conditions required to meet it?

The concept of the “competence window” describes a limited interval during which tissue is capable of becoming something other than what it is. The neurobots suggest a different topology — a presence rather than a window, an inclination that begins expressing itself in the absence of the thing it anticipates. What we are watching might be cellular anticipation: tissue rehearsing futures it will never inhabit.

They grew toward seeing in the dark. The getting-toward left its mark — transcribed in the genome, a record of reaching for something just past the tissue’s edge.

That such reaching can now be measured, extracted as differential gene expression, displayed on a heatmap — is astonishing.