Another issue that I've seen you and others address elsewhere and that I want to inject in this conversation is that the public discussion of "AI" often conflates everything from the simplest ML to to NLP to LLMs to GenAI to AGI in terms of both benefits and harms.