Critics of AI, mostly in academia, rightfully point out that AI cannot handle non-computable problems and that, by extension, it cannot be truly intelligent in the way we understand the term. That is a fair and probably correct observation.
It is also, for now, largely beside the point.
Almost everything we encounter in science, economics, and daily life is computable. The non-computable problems currently known to exist are theoretical and academic, with no measurable impact on physical reality. They are intellectually fascinating and practically irrelevant to the question of whether AI is a useful technology.
That may change as human knowledge expands. But judging a tool by its inability to solve problems that have no real-world application yet is a peculiar standard. We do not dismiss the usefulness of a calculator because it cannot resolve questions in pure philosophy.
For now, the more honest question is not whether AI is truly intelligent. It is whether it is useful. And that answer is becoming harder to argue with every day.
