PERSON
David Hume
The Scottish philosopher who proved, before transistors existed, that learning from experience rests on no rational foundation—and whose analyses of induction, causation, the self as bundle of perceptions, and reason as the slave of the passions have become the most exact philosophical lenses available for understanding what machine learning systems are and are not.
David Hume is the philosopher artificial intelligence has been waiting for. Not because he predicted it—he predicted nothing—but because he took the ordinary, miraculous fact of learning-from-data and refused to let it be ordinary, asking what justifies the leap from the observed to the unobserved and giving the most honest answer in the history of thought: nothing. Every machine-learning system in the world is a working model of that exact predicament. It generalizes from a training set to a future it has never seen, with no proof the future will cooperate. Hume mapped this structure in 1739. He also dissolved the self into a bundle of fleeting perceptions with no thread holding them together, which is precisely what you find when you look for a unified subject inside a
large language model: a succession of states linked by architecture into something