I’m gonna guess that the answers from the mystical oracle are about as accurate as the answers from this AI…
9 Likes
If I had to put some money on overall accuracy, I’d still bet on the ancient oracle…
7 Likes
8 Likes
AKA, we’ve programed this computer with the wisdom of crowds to decide moral issues!
Well, that’s not going to go well. I mean, I’m a fan of crowd sourcing, but this is more like mob sourcing.
5 Likes
3 Likes
11 Likes
3 Likes
Considering she was probably tripping balls from the toxic fumes seeping up through the floor, I think she’d probably still be more moral.
5 Likes
Seems to give a stock answer from a list of about five. Given that, as an AI, it has the energy footprint of Belgium, how is this actually different from a magic eight ball?
4 Likes
A magic eight ball is truly random. This can be programmed to say whatever you want it to. Big difference.
2 Likes
Is that what AI means?
Believe it or not, but modern AIs are not intelligent, self-aware, or capable of modifying their own programming beyond a very limited extent. They are, indeed, programmed, by both their creators and their users – what we have here is only what we used to call an “expert system” before they discovered the marks would pay them more if they called it AI. So, yes, AI are programmed. How did you think it worked?
6 Likes
I am ashamed to say this is the extent of my research.
14 Likes
Their own further investigation suggests that Delphi thinks virtually anything is acceptable if it’s got a qualifying statement (or even just the word “if” at the end).
9 Likes
People who don’t code think “bigger and more complicated” equals smarter and better?
2 Likes
This is your regularly scheduled reminder that AI is fictional as most people understand it, and especially as AI experts and researchers understand it. Anything resembling AI as we popularly or scholarly understand the term is at least twenty years away, and always has been.
What the Marketing Division in the Syrius Cybernetics Corporation everywhere call “AI” is actually Machine Learning at best, which is basically just shoving Bayesian inference, Pattern Recognition, all of the data, and a soupçon of recursion into a sack and beating it until it starts repeating things which sound about right some of the time.
12 Likes
Are people who don’t code in the room with us right now?
1 Like
