Sure but you can't say “the numbers are flat hence nothing is happening” when the numbers aren't in fact flat. Maybe it's trending up for completely unrelated reasons, as it's very hard in economics to derive the actual causal relationship, but a blanket refusal is unwarranted and should be counted as the unsubstantiated opinion it is.

Yeah, I would be impressed by an economist talented enough to discern a defensible signal on the impact of AI on real job growth. I think the next best thing we have is to observe the tendencies of supply/demand; if people were actively being replaced by AI, that large demand would be being enabled by some pretty wealthy companies providing AI laborers, white collar desk job perhaps; however the companies providing such a service still involve a lot of human-in-the-loop type stuff, like those startups offering "AI coworkers" through a Slack or Teams interface.

Those companies though are not nearly as successful as OpenAI/Anthropic last I checked, which are supplying a lot of consumer AI/ AI being used by humans rather than AI doing what humans do, independently. And I'm pretty sure the majority of AI token burn is not replacing humans; it's building software under the direction of humans.