@manton to be fair, my experience contacting Verizon customer support was practically indistinguishable from talking to myself.

@RyanParsley Haha. That’s about right.

@manton Of all the AI features of questionable utility, this one makes me crazy. It pops up *in your way* rather than being an option you could pull up if you wanted it. You have to slow down to not trip over it. AFAIK there’s no way to disable it.

It doesn’t think anything. That’s the whole point.

I know you know that, but this way everyone has of talking as if LLMs think just perpetuates the misunderstanding.

It’s a mindless box of numbers mindlessly predicting tokens.

@njr I’m not sure we have a better word than “think” though, even if it’s all an illusion.

Sure. I’m nor criticising or judging. But language is so powerful. It’s part of why LLMs fool people (or gain theor trust if you prefer.) I just think its useful to remind ourselves periodically that it’s just a mindless box of numbers.

@njr Yeah. It’s fascinating because the more advanced AI becomes, the more we are fooled. It’s uncharted territory that we don’t really have the language for.