What a generous euphemism. Depending on the topic, it's not unusual to see Google AI search referencing links that indicate the exact opposite of what it presents in its prose.

So many of these tools are presented as if they're doing something you'd intuitively expect of a human operator, like collating search results and then summarizing them in this case, but the actual operation of them is so alien to us that our intuitions don't apply and these presentations are all but fraudulent.

Because it's presentation cites references, and because those references sometimes reinforce the expression of the AI prose, we grant it trust, but it doesn't use those references in the way a human would and so the relevance of them being cited is radically different (and weaker) than we're accustomed to.

The closest example in human behavior might be the rushed, naive student who just pastes citations into an already written paper at the last minute, but even that's an anthropomorphism that misrepresents the alien ways in which the tool works.

I mean, just stop to think about it. You have a Youtube account, you post videos. Analytics shows traffic from Google but can't show you the keywords.

At one point, Google merged Youtube with Google+. Just did it outright without asking anybody.

Youtubers link their Youtube accounts with Google AdSense.

You can link your Youtube account with your Google Ads account if you want to advertise.

One of the latest features added to the Search Console was... an AI chatbot.

So Google showed us they can link anything with Youtube when it wants, and they can add new features to GSC if it's AI, but for some reason showing Google search terms for Youtube channels took 20 years.

I wish I knew a word like enshittification but for when something improves but at a rate so slow you don't even feel like celebrating it anymore. I think I'd use such word a lot...

People are probably annoyed that AI/LLMs are lying to them, so not everyone sees the main threat being Google trying to containerize and "walled garden" approach access to information.

> I have certainly found that when I ask a chat tool something, and then go to the human-written source webpage, it's not uncommon I get a different insight than what the AI implied.

Well, many people pointed out that AI lies. We can have two reasons for this:

1) deliberate 2) a feature of how AI works

I think it is mostly 1), but 2) could also factor in, because if ALL AI is lying, then probably something is wrong with how AI works right now. Since they prefer to hallucinate. But AI could be made stricter and not hallucinate, so I think 1) is the main issue.

I've replaced Google almost wholesale with ChatGPT and Claude.

[1] If you use Claude Code to assess a Netlify website (even your own), they'll block your IP address from the entire netlify.app IP range.

Do you have a source for this? So far every single LLM-like does a lot of hallucinating, often just small things like wording that implies stronger correlations than exist, or wording that implies relationships that don't exist, etc. but sometimes it's also just wrong facts.

Used to be you'd look at your server logs and see referrer headers of google.com/?q=search+terms

Then they broke that (deliberately, well before cross origin header concerns were a thing) so you'd have to sign up for their webmaster tools.

Google also did 100% make search worse on purpose as this great piece discusses: https://www.wheresyoured.at/the-men-who-killed-google/

Doctorow goes into greater detail on how Google purposely worsened Google search in his book Enshittification. It’s a very thorough chapter and it’s clear what they did. They were chasing “engagement hacks.”