A similar take would be that why use LLMs to query information since we can use normal search and still find the results. Defaults and usability matters.
That's probably the worst analogy one could make when trying to show that one system is superior to another when we all know and experience daily that LLMs hallucinate search results all the time and are completely unreliable.
My original point is that Kagi is NOT better than Google at search.
I think you missed the point - for masses the correctness matters less than average usability and efficiency of getting results that are good enough most of the time.
