Keith believes that LLMs are a red herring as they have an impoverished world view, however, he doesn't rule out machine consicousness. Saying it is likely that we will have to extend moral concern to AIs once we have convincing, self-sustaining, world-facing robots.
[link] [comments]