Where This Work Stands
Philosophy of technology, from outside. The ground, the tool, the angle.
I work on the philosophy of technology and of artificial intelligence from an external position: no affiliation, no academic reputation to defend. It is a limit and a freedom at once. The limit is obvious. The freedom is that I can follow an idea wherever it leads, without asking a discipline for permission.
The ground
Philosophy of technology and of AI
The philosophy of technology does not make technology: it reflects on the conceptual structure of what technology sets in motion. What changes, precisely, when a machine begins to rewrite itself? It is a question about the meaning of the facts, not one more fact.
The case that interests me — AI that modifies itself — sits at the crossing between this discipline and the philosophy of biology. There, people have long discussed when a process is still “evolutionary” and when it stops being so. That conversation existed before AI; AI makes it urgent.
The tool
Evolution as a matter of degree
I use a way of thinking about evolution that does not ask “is it Darwinian, yes or no?” but “how Darwinian is it?”. A process can move away from the paradigm case along precise directions: how blind the variation is, how slow and external the selection is, how far apart the one is from the other. It is the right instrument for saying not that AI “is not evolution”, but that it is an evolution that leaves the paradigm along one specific dimension.
“The useful question is not whether a process is Darwinian, but along which axis it stops being so.”
The angle
Compressione evolutiva
The dimension I isolate is the lag. For four billion years variation and selection were kept apart by time and by death: a variant was proposed, and only much later did selection say whether it held. Compressione evolutiva — evolutionary compression — is the collapse of that distance until the two coincide in a single step.
It is a different angle from the one that prevails in the neighbouring debate. The best-known analyses look at individuality (which new unit evolves) and at risk (what a system that escapes control will do to us). My cut is neither populational nor prudential: it is internal and temporal. I do not ask who evolves, nor what will happen. I ask what happens to the structure of the process when the lag that made it Darwinian disappears.
The posture
This is not a new theory of evolution, and it would be dishonest to sell it as one. It is an explanatory refinement: a distinction laid on top of a framework that is largely already correct. I am not looking for validations which, as an independent, would not arrive anyway; I am trying to say one precise thing, with the right references within reach, for whoever is curious enough to follow it. I go round the obstacle instead of breaking through it. It is the way I think best.
Who is writing
Massimiliano Fazzini — Estrinseco.
I write about AI, robotics and the future of work from a position outside academia. An
entrepreneur, not a researcher: the lack of a philosophical training is exactly what
allowed me to arrive here starting from somewhere else.
There is an irony I keep in view: compressione evolutiva holds that the novelty is not speed, but the collapse of the lag between trying an idea and finding out whether it holds. This work, done by thinking together with AI systems, was born in just that way — by reducing that lag. It is, on a small scale, an example of what it describes.