Latest A.I. can work things out without being taught.

[Shadochi]

Err, yes they can. Someone that taught themselves something is called an autodidact. I am not trying to say AI isn't impressive, I'm just saying that it'll get *really* impressive when they start to self-direct.

Yes, but they still needed someone earlier in their life to tech them how to read, learn and so on... Which they later used to learn on their own. Nobody was proficiently self taught without any prior knowledge about learning methods. The same is happening with AI (or is the goal) teach it how to learn and let it teach itself afterwards.

[Khallid]

https://www.economist.com/news/scien...gs-out-withoutI'd be interested in seeing A.I. tackle a complex card game like Poker or Magic the Gathering. Go players were amazed at how familiar, but alien the A.I. strategy is. I could easily see MTG's Modern format looking completely different if A.I. got a hold of the card database and did it's thing.

A.I. are very interesting and they can learn in weird ways that don't resemble people at all.

I remember reading a few months ago something about two A.I. computers that were linked and they made their own language to communicate with each other. Since no one could understand them, they had to be discarded, apparently.

Here's the article: https://www.forbes.com/sites/tonybra.../#7b8e02e2292c

[AeneasBK]

You mean just like humans can learn anything by themselves without any guidance? Oh wait, they can't...

Most of the non religious people would disagree with you, but maybe that's why you posted it.

[Shadochi]

Most of the non religious people would disagree with you, but maybe that's why you posted it.

Ummm, I wasn't talking about religious guidance, but more so learning guidance as in teaching you how to read, learn and so on.

[PC2]

Optimal Bluffing is actually very procedural and mathematical despite its depiction in popular culture as being "human". There's an optimal randomized percentage for every combination of cards-the only barrier to perfect play are the very large number of permutations making correct strategy difficult to calculate for humans. This is what computers have always been good at.

I'm not sure AI is actually making as much progress as people seem to think. Computers are really good at working through permutations quickly because they don't have to worry about leaving the iron on. They don't seem to be making that much progress towards replicating human learning: see youtube's idiot AI banning random and arbitary videos.

It's not making much progress except on low dimensional problems, often 2D flat games. Adding in depth, time(planning), or embodiment is too much for current hardware/software.

There is a lot of hype and it's very trendy to say it's going to cause significant net job loss sometime soon. I think these people are severely underestimating the human mind and how good it will be at moving into new job categories.

[AeneasBK]

Ummm, I wasn't talking about religious guidance, but more so learning guidance as in teaching you how to read, learn and so on.

So who taught the first people to use eg. written language? Everything humans know we taught ourselves, or mimic'd nature; the idea that we had to be taught it first is intrinsically religious...

EDIT: sorry get what you meant now.

[Shadochi]

So who taught the first people to use eg. written language? Everything humans know we taught ourselves, or mimic'd nature; the idea that we had to be taught it first is intrinsically religious...

True, but the first written language didn't happen in one generation, it evolved from picture-like languages where its more drawing and interpreting the drawing(which were quite primitive and lacked abstract concepts) to letter based languages.And that transfer took a long time.

Also, the first written languages were "invented" not "taught" since you can't really teach something that doesn't exist yet even in concept.

If we had an AI that spew out gibberish strings of random letters/numbers how could we know that it did/didn't crack all the secrets of the universe and is telling them in it's self-invented language since it doesn't know ours? That's why we teach them our languages and model them a bit around ourselves. We just can't wait for them to develop something that took us millions of years of evolution, and then we would have to learn their ways as well. There's no need to go and reinvent the wheel.
For me a truly independent AI is gonna be one that can reproduce (not clone) itself into not-completely-same offspring. Heck, add a physical form to that which they can build and we might need to start calling them "living" synthetic beings. But this is probably hundreds or thousands of years away.

Also, while my knowledge on this is by no means absolute, I did do research on early and modern written languages for my Worldbuilding. As for AI, I'm in University and focusing on AI and Machine Learning.

[RaoBurning]

Seeing this thing play Starcraft or Magic would be really neat. There are limitations and quirks to Magic, for instance, that don't exist in Chess and Go. Deck construction being the primary example. Even if the AI understood the rules, if it sucked at building a deck it wouldn't win against a moderately competent human opponent. Given the vast, vast number of cards to build from, teaching it (or letting it teach itself, as the superior case seems to be) to build from scratch and win might yield combinations that human players never even thought of.

I'd be super excited to see that play out.