Man, seriously, every time I see someone get into these weird conversations where they try to convince a chatbot of something it's slightly disturbing. Both not being aware of how pointless it is and knowing but still being compelled by the less uncanny valley-ish language are about on par with each other.

People keep sharing this as proof of AI shortcomings, but it honestly makes me worry most about the human side. There's zero new info to be gained from the chatbot behavior.

They take a sentence and predict what the first result on google or response on whatsapp would look like.

This has been true for a decade. Gonna say when if them starting to blatantly train facial recognition on people's private pictures didn't do it, then this isn't going to.

I keep bringing this up to people and nobody seems to be particularly interested in that acknowledgement. The shills don't like to compare precedent, the viral critics don't like to acknowledge the lack of novelty in the current situation or their inability to trigger any mainstream action.

I'm just... kinda sad.

Oh no, they'll learn your personality profile and manipulate you, get you on the edge of suicide, only to bring you back of the ledge to sell you the newest AI sunglasses.

It seems so obvious to me that Google's switch to LLMs is to prevent adblockers and yet I rarely see that point brought up.

No matter what you and I call "real" AI, marketers are going to misuse the term on everything.

We used to call these kinds of things Assistive Intelligences. But the marketing departments gonna market.