Not really. AI has proved very effective at winning Chess, Jeopardy and Go, but it has done very little in the grand scheme to replace stuff that humans do. In the near- to medium-term it's mainly low-hanging fruit like search engine optimization that's at risk of replacement.
-
-
Replying to @anti_minotaur @Outsideness and
I hate these discussions we're expected to choose between saying AGI isn't possible and GI is perfectly measurable, reducible to game winning etc. the fact that AI that does these things isn't good at anything else means Nick is, concretely, wrong about what intelligence is.
4 replies 0 retweets 3 likes -
Replying to @baroquespiral @anti_minotaur and
either that, or the whole paradigm of "AI in a lab" is mistaken.
3 replies 1 retweet 6 likes -
Replying to @cyborg_nomade @baroquespiral and
which I take is Land's larger point: AI is simply Capital at work. things like AlphaZero are parts of an ongoing automation trend, not its final triumph.
2 replies 1 retweet 13 likes -
Replying to @cyborg_nomade @baroquespiral and
and I think something like AlphaZero could be applied to, say, corporate processes and/or trading in order to automate a huge chunk of the economy.
3 replies 1 retweet 10 likes -
Replying to @cyborg_nomade @baroquespiral and
Oh yes, let's apply board game logic to economy and beyond. What can go wrong, maybe just a few charlatans get caught, but the rest will be thriving.
3 replies 0 retweets 1 like -
Replying to @NegarestaniReza @baroquespiral and
absolutely see no reason why a corporate process couldn't be seen as a very simple board game
2 replies 0 retweets 1 like -
Replying to @cyborg_nomade @baroquespiral and
If you had a morsel of familiarity with theoretical computer science, you could see that every board game is a subset of a larger one. Computation is all about the range of interactions, not all of them can be treated within the game theoretic framework orscenarios of competition
5 replies 0 retweets 5 likes -
Replying to @NegarestaniReza @cyborg_nomade and
The whole point of alphago was to demonstrate superhuman performance in an uncomputably large state space.
1 reply 0 retweets 1 like -
Replying to @higherOrderNet @NegarestaniReza and
Also, see the universal approximation theorem.
1 reply 0 retweets 1 like
No you are making things up. What does superhuman ability even mean? Does a god who can smell a broad range of smells is superhuman? If yes, you should be more context-sensitive rather than than all over the place.
-
-
Replying to @NegarestaniReza @higherOrderNet and
You seem to misunderstand universal approximation theorem as truly universal. No, this only applies to certain functions which fall under polynomials constrained by Lebesgue measure. There are different types of functions which don't follow this pattern.
1 reply 0 retweets 1 like -
Replying to @NegarestaniReza @higherOrderNet and
I mean just because there is a word universal there, you shouldn't think it as Universal. that's 101 semantics.
1 reply 0 retweets 1 like - 5 more replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.