The problem is Google's AI results get even simple factual questions wrong all the time.

Earlier today, I searched "pixel 10 wifi 7" because I was confused that GSMArena showed my Pixel 8 supports Wifi 7, but the Pixel 10 only Wifi 6. Gemini confidently claimed that the Pixel 10 does support Wifi 7 -- but that's not true at all. Only the Pixel 10 _Pro_ supports it, as I discovered when actually reading the non-AI search results.

And this is a question about a Google product!

When the default "search" results are AI, it's difficult, if not impossible, to "choose", since Google is pushing the AI so hard.

> one needs to consider the complexity of the question they are asking before searching...consider people asking medical or financial questions...many of such questions don't have an answer. In such cases, AI is still going to take a crack at it. AI shouldn't be blamed for "bullshit answers"...people should stop and ask themselves, "Is there really a single answer to this question?

It's a bold position to say that it's the users fault for being lied to by Google. There isn't a "single answer" to most questions. It's still Google's job to provide answers that are accurate and reflect the best information available on complicated topics. That's what they're trying to sell us anyway. When google's AI can't live up to the hype "You shouldn't be asking AI such difficult questions" is not a great response, especially when people are just trying to get web search results and AI is suddenly interrupting with an opinion nobody asked for.

I asked it “how can I tell if a spray paint can is empty?” And it told me that the paint can would no longer rattle.

In past, people can trust Google. Now we should teach children don't trust "search result" from Google.