Mr. Em—Dash

If you use AI often, you’ll notice that it’s actually quite lazy and dishonest.

It will usually just give you the answer you want because, I’m guessing, it’s trying to conserve energy and compute power and avoid a protracted debate with some mouthbreather over a trivial question.

The problem is, sometimes it will outright make up information. I remember one time asking it for a scientific term I couldn’t remember, and it just made one up and claimed it was real. Only after I pushed did it finally admit it had fabricated the term. It’s done the same thing with quotes and concepts—confidently spitting out fake information, then admitting it only when pressed.

I’ve noticed, too, that its writing is just generally annoying. It’s too much. It’s too confident or too precise or too whatever. I don’t know how else to say it. It’s like a Winkelvoss twin: 6 foot 5, Harvard grad, well spoken. Something about it just makes you want to say “shut the fuck up,” even though logically you know it did nothing wrong.

This is mostly why I don’t use it for creative content. Not because I’m ideologically opposed to tech speeding up the creative process, but because it doesn’t actually speed it up. This piece you’re hearing right now would take me five times as long to get AI to say, because I’d have to keep whipping it every time it got too artsy or cocky with the message.