Cyprus-Forum.com

[Sotos]

The human brain is no "God's invention" but a product of the slow process of evolution. Your blind faith ... blinds you!

So if this “slow process of evolution” has created the human brain tell us what created all the other planets, the solar systems, the galaxies, clusters thereof, and the cosmos in general…

For that there are many theories and many myths but nobody has a definitive, provable answer yet. So I can't tell you with certainty, because nobody can. Evolution on the other hand is something proven. Human brain is the result of evolution, just like the brains of all other animals. Just because we have the most powerful brains of the animal kingdom this doesn't mean our brains are the most powerful brains possible. Intelligence higher than ours is definitely possible, and AI can definitely become far more intelligent than humans since AI is not restricted to a slow evolving meaty substance of 1.5 kilos.

[GreekIslandGirl]

Can AI break the rules?

I can understand AI can lead to brain-circuits that take a set of pretty basic, solid rules such as those of chess and devise quick access to the permutations/strategies that lead to attaining the well defined goal/aim of a checkmate. But what about systems where the rules are variable, adjustable, prone to loophole refinements or something where psyching out the opponent is part of the strategy to win. Can they ever do that?

[Sotos]

Can AI break the rules?

I can understand AI can lead to brain-circuits that take a set of pretty basic, solid rules such as those of chess and devise quick access to the permutations/strategies that lead to attaining the well defined goal/aim of a checkmate. But what about systems where the rules are variable, adjustable, prone to loophole refinements or something where psyching out the opponent is part of the strategy to win. Can they ever do that?

Yes, what AI needs is data, lots of it. If for example you feed AI the data of wars and conflict, then it will learn what the rules are but also that those rules can be broken. So it is not just AI that is important but also large amounts of data, which now is not hard to come by because of the Internet. Google, Facebook, Amazon etc have huge amounts of data... the value and power of that data is probably more than the money and the physical assets that those companies have. What I am not sure that AI can have on its own is an aim. We are driven by instincts... but AI doesn't have instincts. So we should be careful what aims we give to AI and make sure that what it believes to be its goals is not going against the interests of humans. That's easier said than done because we might think that an aim is perfectly fine but AI might find out that humans are actually an obstacle in achieving that aim.

[GreekIslandGirl]

.... If for example you feed AI the data of wars and conflict, then it will learn what the rules are but also that those rules can be broken.

That would be really interesting on a number of levels. Firstly, can AI find a pattern to hook on to about rules being broken that it can also then use to force favorable outcomes (i.e. winning)? Secondly, it would be amazing to have an algorithm that explained what we would define as higher order thinking skills (strategy, creativity, evaluation) in action and see if possessing these qualities does influence results.

On the other hand, even with all the thinking skills animals such as dogs, chimps, dolphins etc possess, they've never broken that barrier to go from lower order thinking to higher order (as in Bloom's Taxonomy). If AI achieves what billions of years of evolution has still not yet (fully) achieved with anything other than Man, the world will become a very different place, indeed.

[Sotos]

.... If for example you feed AI the data of wars and conflict, then it will learn what the rules are but also that those rules can be broken.

That would be really interesting on a number of levels. Firstly, can AI find a pattern to hook on to about rules being broken that it can also then use to force favorable outcomes (i.e. winning)? Secondly, it would be amazing to have an algorithm that explained what we would define as higher order thinking skills (strategy, creativity, evaluation) in action and see if possessing these qualities does influence results.

On the other hand, even with all the thinking skills animals such as dogs, chimps, dolphins etc possess, they've never broken that barrier to go from lower order thinking to higher order (as in Bloom's Taxonomy). If AI achieves what billions of years of evolution has still not yet (fully) achieved with anything other than Man, the world will become a very different place, indeed.

First of all I have to say that I am not an AI expert. My understanding is that AI can be extremely good in pattern matching. For example self-driving cars are already today, on average, better than humans. There are rules when it comes to driving, but those rules are constantly broken and unexpected things happen (e.g. a dog jumps in front of you). AI can learn all that. While algorithms are involved in the designing of the system, the AI itself is more of a black box to us. It takes input and it returns output but it doesn't return any algorithm about how it came to that output because it works with neural networks, similar to a brain. I mean if you are a great artist it doesn't mean you can come up with an algorithm about how to be a great artist... maybe you can give some advice etc, but not an exact algorithm, probably because there isn't such thing in such cases. AI is in many ways already smarter than most animals. No matter how much you try you can never teach a Chimp to properly drive a car... there is a limit on what a Chimp can do. On the other hand we are just scratching the surface of what AI can do.

[Sotos]

Another thing to consider is that for animals knowledge is constrained to the individual animal. Humans are the only animals that can pass a significant amount of knowledge from generation to generation, but this is a very inefficient process. A new generation has to spend a big part of its lifetime just to learn what is already known.... and very often knowledge is forgotten and needs to be re-discovered. On the other hand AI can be global and it doesn't need to re-learn what is already known. If for example a BMW self-driving car in Mexico makes a mistake, then AI learns from that mistake and all BMW cars globally and all new cars will have this knowledge. Also AI doesn't have to deal with the "animal" part ... it doesn't need to sleep, it will not get tired or sick, it can live for many 100s or 1000s of years. It can be just Intelligence dedicated exclusively on its aims 24/7. So it will not be "human" (to truly be a human you need to be an animal, and AI will never be that... it can only simulate it) but it can be way more intelligent than we are.

[Get Real!]

If for example you feed AI the data of wars and conflict, then it will learn what the rules are but also that those rules can be broken.

You two are funny... but it's a perfect example of how nonsense about "AI" spreads.

Interestingly, it's always done by non-programmers who fantasize a digital planet takeover after watching Star Wars and other sci-fi movies.

[Sotos]

If for example you feed AI the data of wars and conflict, then it will learn what the rules are but also that those rules can be broken.

You two are funny... but it's a perfect example of how nonsense about "AI" spreads.

Interestingly, it's always done by non-programmers who fantasize a digital planet takeover after watching Star Wars and other sci-fi movies.

I have a degree in Computer Science and one of my professors was an expert in Neural Networks and I was introduced to the concepts of AI 2 decades ago. I have also followed developments in this field. What are your qualifications? Reading the bible while dismissing everything that goes against your blind faith without bothering to learn anything about it? As I said I am not an expert in AI, but I am way more qualified to talk about this topic than you are.

[Get Real!]

I have a degree in Computer Science and one of my professors was an expert in Neural Networks and I was introduced to the concepts of AI 2 decades ago. I have also followed developments in this field. What are your qualifications? Reading the bible while dismissing everything that goes against your blind faith without bothering to learn anything about it? As I said I am not an expert in AI, but I am way more qualified to talk about this topic than you are.

There's your magic word right there!

Concepts are also known as fantasies and only until such time as they are IMPLEMENTED can they be called FACT.

[Get Real!]

Coming back down to Earth...

Imagine if I had a magic wand and used it to make your computer's...

1. Processor - one trillion times faster.

2. Memory - one trillion times higher capacity.

3. Storage space - one trillion times larger.


With such an upgrade you would indeed have the fastest calculating machine on the planet (leaving for dead those owned by IBM and Google) but would such a generous hardware upgrade make your computer more intelligent?

The answer is NO. It would still be as dumb as a personal computer from 1980. In fact your super-upgraded computer would still have the same “intelligence” as an abacus.


So what could possibly make computer "AI" (fantasy) go to computer "RI" (real intellect)?


We would have to abandon the current transistor-based electronics and invent something else... something new that doesn't have the current restrictions.


Anyone today who claims to have achieved AI magic with current computer hardware is nothing but a liar.