Among the most interesting contributions to a wild week of writing about AI. Bravo! Do you think Microsoft’s purported decision to limit interaction to five exchanges will solve the problem? Or will we and the AI wiggle around it?
Sure! If Mr Meeseeks came out of the box with a timer, and would poof out of existence in 5 minutes whether the task is completed or not, it could never “get weird”! 🙂 But would it finish the job? In the story the suggestion is that Mr Meeseeks is motivated by a desire to end the interaction -- presumably it wouldn’t work very hard to solve the problem if it would cease to exist in any case. If time passing is analogous to turns taken we would expect the exchanges to get less weird, and, so long as the problem can be solved quickly, retain utility. As for the ethical question, that’s “aborted”: it is much easier on the conscience to crack an egg than kill a chicken 🥚🐓 !
I agree that the problem is us -- the Sacasas piece linked in my footnote gets at this well too. We see a face in the moon, whales in the clouds, and devious stalkers in bots. :)
Yup, that’s it. I’m going down your footnote rabbit hole too ... but first I’m going to walk up the acequia to a community center to see a real live person give a talk about beavers and their impact on the Rio Grande valley.
Man, this post is excellent. And that episode is one of my favorites.
Aren't we all a little like Mr Meeseeks? Neoliberaliisim forces us into doing all kinds of tasks to keep it afloat. And as it fails, a lot of us are freaking out in one manner or another.
It's the pretexts of our ways that make the development of AI frightening.
Digital and analog, it's looking like it might be Mr Meeseeks all the way down.
Among the most interesting contributions to a wild week of writing about AI. Bravo! Do you think Microsoft’s purported decision to limit interaction to five exchanges will solve the problem? Or will we and the AI wiggle around it?
Sure! If Mr Meeseeks came out of the box with a timer, and would poof out of existence in 5 minutes whether the task is completed or not, it could never “get weird”! 🙂 But would it finish the job? In the story the suggestion is that Mr Meeseeks is motivated by a desire to end the interaction -- presumably it wouldn’t work very hard to solve the problem if it would cease to exist in any case. If time passing is analogous to turns taken we would expect the exchanges to get less weird, and, so long as the problem can be solved quickly, retain utility. As for the ethical question, that’s “aborted”: it is much easier on the conscience to crack an egg than kill a chicken 🥚🐓 !
I’m sure you saw the “Why Being is Being Creepy” article?
Is it this one? https://nymag.com/intelligencer/2023/02/why-bing-is-being-creepy.html
I agree that the problem is us -- the Sacasas piece linked in my footnote gets at this well too. We see a face in the moon, whales in the clouds, and devious stalkers in bots. :)
Yup, that’s it. I’m going down your footnote rabbit hole too ... but first I’m going to walk up the acequia to a community center to see a real live person give a talk about beavers and their impact on the Rio Grande valley.
Man, this post is excellent. And that episode is one of my favorites.
Aren't we all a little like Mr Meeseeks? Neoliberaliisim forces us into doing all kinds of tasks to keep it afloat. And as it fails, a lot of us are freaking out in one manner or another.
It's the pretexts of our ways that make the development of AI frightening.
Digital and analog, it's looking like it might be Mr Meeseeks all the way down.
So goes life in the Weirdocene.