It may be of interest to philosophers, but it has little impact on economic job replacement and how people will earn their living and all the downstream upheaval from that. At some point maybe philosophers will find AI-generated philosophical musings about the nature of AI to be better than what comes from their peers (if blinded).

It also has little impact on the dangerous use cases.

I think it matters a great deal that LLMs and other "AI" technology are stable within a tolerance that's acceptable.

You're jumping the gun talking about "job replacement". We have not thought about it enough from that engineering angle. It's still very early days. That engineering is going to require people. :-)