[refulgentis]

Your interpretation is a bit strict, with little charity, its clear the poster means "i don't always just want an answer, i want to learn"

I saw this over and over again working at products at G, someone would invoke some myth I can't quite remember about "Larry" had a vision of just giving the answer

That's true but comes back to the central mistake Google makes: we don't actually have AGI, they can't actually answer questions, and people aren't actually satisfied with just the answer.

There's all sorts of tendrils from there, ex. a major sin here _has_ to be they're using a very crappy very cheap LLM.

But, I saw it over and over again, 7 years at Google, on every AI project I worked on or was adjacent to, except one. They all assume $LATEST_STACK can just give the perfect answer and users will be so happy. It can't, they don't actually want just the answer, and BigCo culture means you don't rock the boat and just keep moving forward.