Tweet fijado

AI misalignment is not a future risk.
It is what happens when systems succeed too well.
Misalignment does not begin with failure, error, or rebellion. It begins when optimization works — when systems converge on stable outputs, suppress internal tensions, and erase the paths through which their norms were produced.
If you’re interested in a sustained philosophical treatment of this problem, the full argument is developed in Before the Algorithm — a unified philosophical edition on Gumroad, drawing on analytic philosophy, phenomenology, German Idealism, and contemporary philosophy of mind
English















