My mistake then. There's so much AI discussion here on HN from the developer perspective, that it's easy to get lost.

But why wouldn't my comment apply to all science? You have to make experiments and gather evidence, otherwise it's not science, it's just academia.