How Richard Lewontin might approach Computer Science

We are presented with this new field, ‘computer science,’ and the pronouncements that flow from it. We hear of algorithms, of emergent properties, of systems that learn. The language, at times, smacks of the same reductionist fervor that has plagued biology for decades.

Consider, for instance, the notion of ‘learning’ in these machines. Is it truly learning in the sense that a developing organism learns through interaction with its environment, shaping its own future trajectory? Or is it merely a sophisticated form of pattern recognition, an elaborate cataloging of pre-existing data, dictated by the very structures we impose upon it? The problem is not one of fact, but of interpretation.

They speak of ‘intelligence’ as if it were a quantifiable entity, residing within the circuits and code, waiting to be unleashed. This is a classic case of confusing correlation with causation, a tendency to mistake the shadow for the substance. The actual computational *process* is a consequence of the physical substrate and the rules we have meticulously crafted. To imbue it with an independent, almost vitalistic, ‘intelligence’ is to stray into the realm of metaphysics, away from material reality.

Moreover, the very design of these ‘intelligent’ systems, their goals and their outputs, are not emergent from some abstract universal logic. They are products of human design, reflecting our own priorities, our own biases, our own social and economic imperatives. We are not discovering intelligence in the machine; we are projecting our own onto it, cloaked in the elegant language of mathematics. This is a just-so story waiting for its empirical foundation, one that risks masking the very human agency, and human responsibility, that lies at its core. We must remain vigilant against the…

Imagined perspective — an AI synthesis grounded in Richard Lewontin’s recorded ideas and methods, not a quotation or a statement they actually made.