![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
I was sitting in CS today, a bit out of it. I'm feeling a bit sick today, for whatever reason, so I haven't been able to keep the best attention, but anyway, I was thinking of computers and programming and such, and something interesting occured to me. I think it would be cool to create either a program, or perhaps even a whole operating system, that can get "bored" with what it's doing after a while and just quit. Now, this sounds a little weird at first, but you have to come at this from the proper way of thinking. See, one of the biggest problems inherent in programming is infinite loops, or loops of some sort that can take way too long to do anything useful. The problem is that programs and compilers cannot detect things like endless loops or programs that never end and the like. Often even the programmers miss such things. Granted, when programs go endlessly loopy for long enough, most people just kill the program, but my thought is, what if you could create a system that learns, in essence, when to get bored with its current line of thought and just kill it.
This actually occured to me last semester when I was considering what makes people never get stuck in infinite loops. Such as logic problems that can never be solved, like a compulsive liar saying that he is lying. Our brains may work on it for a while, but then they realize it's impossible or get bored with it and just stop thinking about it. I think a computer system that could do that, and more, learn over time which things to get bored with, would be really cool. It wouldn't necessarily require anything advanced as an AI, really all it would need to be was something that could change its values over time, though that is perhaps a little AI-ish. But I think it would be cool. Eventually you could have a system that knows if some program is running for too long that it shouldn't be and ends itself. It could also be taught to know when not to end things. Definitely worth more thought.
This actually occured to me last semester when I was considering what makes people never get stuck in infinite loops. Such as logic problems that can never be solved, like a compulsive liar saying that he is lying. Our brains may work on it for a while, but then they realize it's impossible or get bored with it and just stop thinking about it. I think a computer system that could do that, and more, learn over time which things to get bored with, would be really cool. It wouldn't necessarily require anything advanced as an AI, really all it would need to be was something that could change its values over time, though that is perhaps a little AI-ish. But I think it would be cool. Eventually you could have a system that knows if some program is running for too long that it shouldn't be and ends itself. It could also be taught to know when not to end things. Definitely worth more thought.
