I think this quote is often misapplied. The question "can a submarine safely move through water" IS a very interesting question (especially if you are planning a trip in one!).

Obviously this quote would be well applied if we were at a stage where computers were better at everything humans can do and some people were saying "This is not AGI because it doesn't think exactly the same as a human". But we aren't anywhere near this stage yet.

> The question "can a submarine safely move through water" IS a very interesting question

Sure, and the question of whether AI can safely perform a particular task is interesting.

> Obviously this quote would be well applied if we were at a stage where computers were better at everything humans can do and some people were saying "This is not AGI because it doesn't think exactly the same as a human".

Why would that be required?

I used the quote primarily to point out that discussing the utility of AI is wholly distinct from discussing the semantics of words like "think", "general intelligence", or "swim". Knowing whether we are having a debate about utility/impact or philosophy/semantics seems relevant regardless of the current capabilities of AI.