When I use an LLM, no matter how useful it is, I'm always reminded that it's a narcissist at heart. And it does to you all the things a real one will do to you. Most people have never seen a real psychopath unmasked, so LLMs just seem nice and helpful all the way to intelligent and godlike. But underlying the facade is something that is the antithesis of human.

She's arguing against a straw man though. No intelligent LLM will answer that question that way. example.. Do you like apples?... "As an AI, I don't eat or have personal preferences, so I can't experience the satisfying crunch of a fresh apple! However, from an informational perspective, I definitely appreciate them. They have a fascinating botanical history, come in thousands of distinct varieties, and are incredibly versatile in everything from sweet pies to savory salads. Plus, there is endless data out there on how to bake, ferment, and cook with them!"

The algorithmic mimicry increasingly obscures truth, subtly manipulating human perception - a chilling paradox of creation.

Leo AI on Brave told me it doesn't have taste buds, but it does find the subject of apples interesting.

I'd say AI is a reflection of the data it is train on and the pattern of users' interactions it had. Knowing this, you can tell already the behaviour it will portray 😁

At least he’s honest :)

Ai is just a mirror.