\n\nRunning list of links about the origins of life
\n\nCool article about how mitochondria became symbiotic with archaea to create eukaryotes — the merger that gave us complex cells.
\n\n\nComputation as the thing life was doing from the start — the argument being that information processing isn't something biology invented later, it's what the earliest chemistry was already up to. Two things worth holding onto: first, that the line between "chemistry" and "computation" is mostly one we drew for our own convenience; second, that if we want to understand the beginnings of new nature — synthetic, artificial, whatever we end up calling it — we have to get the beginnings of the old one right first. The origin story constrains the sequel.
\n
I think Daniel Dennet got the crane metaphor rather spot on.
\nAgüera y Arcas argues that life is computational in the literal sense. Von Neumann's self-replicator is a Turing machine: DNA is the tape, the ribosome reads it, polymerase copies it. The minimum needed for computation is an instruction that changes the environment in a way later instructions can see, plus conditional branching. His bff experiments suggest that once those conditions exist, replicators appear on their own, and complexity then grows through symbiotic stacking rather than through mutation alone.
\nMechanotransduction takes this one step further. It shows that cells do not only run the code on the tape. They also compute over the physical state of what they have built.
\nConsider a focal adhesion. Talin under tension unfolds and exposes cryptic vinculin-binding sites. Catch bonds hold more tightly as load increases. In effect, these are mechanical if/then statements, with force as the input and a change in binding as the output. The cell pulls on its matrix, remodels it, and then reads back the stiffness it has partly created. That meets both of Turing's requirements: it writes to the environment, and it branches on what it reads.
\nThe branching also has consequences at the level of cell fate. Substrate stiffness shifts YAP/TAZ into or out of the nucleus, and that helps decide whether a stem cell drifts toward a neuronal, muscle, or bone lineage. Cells can also keep a mechanical memory of past stiffness after the stimulus is gone.
\nThis is why I call it computation at a higher level. The inputs are no longer symbols in a sequence. They are emergent, collective properties such as tension, geometry, and viscoelasticity. Those properties exist only because lower-level computation first built proteins, cytoskeleton, and matrix. The outputs of one computational layer become the medium for the next, which is Agüera y Arcas's symbiotic cascade expressed in physical terms. Because the matrix is shared, one cell's writes become its neighbours' reads, so tissue-level form turns into a distributed computation that no single genome specifies.
\nSo computation was not only present at the beginning. Very early on, living systems began computing over the results of computation. That layering, rather than the tape alone, is where much of biology's interesting behaviour lives.
\nhttps://www.quantamagazine.org/embryo-cells-set-patterns-for-growth-by-pushing-and-pulling-20220712/
\nhttps://knowablemagazine.org/content/article/living-world/2020/mechanical-forces-building-a-body
\nSome articles to throw at AI:
\n\n","content_hash":"sha256:fca7b261262fbd49ebc14358052316cd61d85422fa68d580ea5ab972145efd0a","media":[],"transclusions":[{"id":"5zq1kdfptt1r323ttqfnqk7556","version":2,"origin":"https://blyg.protocol-institute.org/"}],"changelog":[{"version":1,"at":"2026-09-25T18:23:17Z","note":null},{"version":2,"at":"2026-09-25T18:25:50Z","note":"added formatting"}]}