This is so fucking stupid. All Gemini (and every other LLM) does is find info on the internet and regurgitate it back to you. The LLM part just does it’s best to narrow it down to the specific answer you’re looking for much more quickly and summarize it. It has no way to disseminate fact from fiction.

Same goes for the Google search engine. Does it suck? Yes. Is it worse than it used to be? Yes. Is that Google’s fault? Not necessarily, it’s moreso the rise of AI just uploading mountains of bullshit to the internet. Every other search engine has the same problems.

Although I am in no way advocating for the use of Google search…

What? Are you too good for some authentic Italian paste-a?

Too be fair to Gemini, even though it is worse than Claude and Gpt. The weird answer were caused by bad engineering and not by bad model training. They were forcing the incorporattion off the Google search results even though the base model would most likely have gotten it right.

I mean I absolutely want it but also understand your apprehension.

I mean imagine every person having a personal assistant in their pocket who can draft emails on your behalf based on other emails in your inbox, remind you of things without actually creating a reminder, fill out forms for you, answer texts for you, write software for you, create automations for you, etc. etc. I mean the possibilities are endless for increased productivity.

On the other hand, in order to do those things, they have to have access to just about everything, which is scary to me, personally. But I think the vast majority have made it clear they don’t give a single shit about privacy.

An ai can help a lot of non tech savvy people. Imagine every time they want to do something, they could just ask the ai on how to do it, instead of asking you or having to go through a billion SEO garbage google results.