A guy on a recent Startalk really came up with some good analogies for AI and why, even though it’s becoming a pretty good intellectual problem solving tool, it’s still not particularly intelligent because it’s so inefficient. The analogy he gave on that was “If I asked <other guest on podcast> to get me a glass of water, and he got up and got a glass of water and then ran around the room 20 times before bringing it to me, sure he got the job done but we’d all wonder what the hell was wrong with him.”
I’ve been using AI to fix a few old and new gentoo linux configuration snafoos, and even though it has been overall very helpful, in that eventually, working together, we came up with solutions–man was it a winding road for most of them. Sometimes it just makes stuff up I’m pretty sure. Or it will lead you down a path trying something and then seven steps into it be like “oh that’s just for X scenario which has nothing to do with what you’re trying to do, try this instead.”
A guy on a recent Startalk really came up with some good analogies for AI and why, even though it’s becoming a pretty good intellectual problem solving tool, it’s still not particularly intelligent because it’s so inefficient. The analogy he gave on that was “If I asked <other guest on podcast> to get me a glass of water, and he got up and got a glass of water and then ran around the room 20 times before bringing it to me, sure he got the job done but we’d all wonder what the hell was wrong with him.”
I’ve been using AI to fix a few old and new gentoo linux configuration snafoos, and even though it has been overall very helpful, in that eventually, working together, we came up with solutions–man was it a winding road for most of them. Sometimes it just makes stuff up I’m pretty sure. Or it will lead you down a path trying something and then seven steps into it be like “oh that’s just for X scenario which has nothing to do with what you’re trying to do, try this instead.”