The AI agents used by search engines are "stateless" in a sense. They only read from the model, they don't write to it. What you saw is the expected designed behaviour.
Rather undermines the 'intelligence' bit: a central part of human intelligence is the feedback loop that constantly refines the model of the world that our brains rely on to steer us through a chaotic world.
You don't need to convince me.
And it's all happening without conscious thought.
I can't remember which brain book it was that I read that is very persuasive about this central function of the human brain (not just in a generalised philosophical sense, but in light of latest neuroscience research) - but it mirrors the way that each of our brains builds a kind of 'predictive text' model of lanaguage that makes speaking & comprehending (aurally and in writing) practicable. It's constantly measuring what's expected and what actually happening, and where anomalies occur, trying to make sense of them, and either incorporating modifications to the model (or dismissing them as 'noise'). And it's all happening without conscious thought.
It is even more surprising that it works so well for people like me who don't think in words.
Yebbut words are the things that attach themselves to concepts which exist in parallel. Babies are amassing understanding of loads of things (people, food, gravity, movement, etc) well before they have language with which to express their grasp/understanding of concepts.
It is even more surprising that it works so well for people like me who don't think in words.
Angela & Simon?
Yes, and despite Co-pilot's lengthy explanations, I still struggle to see a g and not a p.
Yes, and despite Co-pilot's lengthy explanations, I still struggle to see a g and not a p.