14 replies to this topic

[darko100]

It's relative. What "smart" is to you. Can you calculate pi for 20,000 decimal numbers for seconds? nope. but computers can.

What you call dimension is learning. By letting computers change theirselves from latest actions you could gain something very smart , like a mini brain. There are many smart bots which use the computer power (calculating faster than human) but learning power too.

So:
-multitasking power ✓
-learning / not so much
-using learnt objects to recognize X not at all

So a small 3yrs kid might not be able to calculate pi , but he's still able to recognize shapes,colours,voice.
Computers need to learn the simplest things

[Quatrux]

Artificial intelligence always was an interesting subject, even though all the current artificial intelligence isn't really artificial intelligence, it's just a simple program, there's currently no computer which could work like a brain, a brain can work in parallel of million of neurons, what we have Today is dual core CPU's or Quad core, when we will have CPUs which could emulate the patterns of brain activity, than maybe we could talk about artificial intelligence For now it's only a program full of predicates, usually a knowledge base and rules, which comparing to our brain is really slow on current technology.

But from a programmer perspective, having the technology, I think it's possible to program a lot and an artificial intelligence might be smarter than current humans, but as we all know, we don't use all the power of our brain, so I guess we will always be "smarter" than a computer, you can't rely on a computer decision, but if it has all the knowledge gather, it can help you make a decision, most of doctors Today use different expert systems to make their job easier for any average doctor to make a conclusion much easier.

So, a computer can be programmed in a way to learn and make decisions, a computer can recognize images, sounds and objects and even smells or taste, but quite primitive, as I know only humans have real intelligence, a bug or a beattle for example has no intelligence, if we talk about real intelligence, but I think a program can be made which could act as a bug or a fish and live, as it's also quite primitive, but you would need to have the knowledge base and rules for it.

One of the harder things to implement isn't instincts, as it's quite easy, but the hardest thing and most powerful thing a human has is intuition/presentiment/feeling, lets say when a computer plays chess, it scans all possible moves and uses some kind of rules to not scan all possibilities, a human can sometimes rely on intuition and in most cases he will be right, but that requires knowledge.

Also, on this subject, can a computer feeling be the same as human? If programmers wrote a program which is a model of love and happiness, is that computer programmed happiness the same as real happiness? Most would say no, but practically a person is also a program, just really a good one, with one of the best nature algorithms which is looping since the beginning of time and our algorithms of computers are just spinning since the year ~1950, but we have a lot of knowledge, that is why I also think that intuition can be programmed, but it will really be a one hell of a program

Nevertheless, a computer can't have a soul.

In addition, I saw some sci-fi movies, where a subject like: can a computer/robot be in love with a human? or can a human love an intelligent robot? The future may tell, but I guess it can, due to Today some people can love a lot of strange things So why not a robot...

[SeaOfThought]

While I type in this keyboard I am reminded of how much of my "functioning" is automatic. I already know where the keys are, I already know the words, all I have to focus on is the general idea which, in this case is, I have my own computer inside my head. I don't know how it works, just as I don't know how my laptop works, but I know then why I try to communicate words get out of my mouth and when I try to write my fingers get to the correct keys without much effort (or hardly any conscious effort at all).

But not everything is automatic in me. There is consciousness, a sense of freedom, lights and sounds all around. It is this "light" of consciousness that allows me to distinguish between right and wrong answers, to see if anything is missing in a puzzle, if some goal "feels right". The computer in my head doesn't have nothing of that. At least as far as I can see it is simply programmed by me to do common tasks (like reading, writing, putting the gears in my car) without me having to think about it. It does not feel, it simply does.

Now, this difference between what I do automatically and what I do consciously may lead me to establish the same difference between automatic digital computers and conscious persons. After all, a computer is just a machine, programmed to do something. So it is just blindingly executing orders just as my brain is repeating a well-rehearsed movement, when walking or typing, that I taught it long ago.

But this analogy might not be good. Notice that certainly our automatic/unconscious procedures are essential to our lives, we couldn't operate in a complex world if we had to think constantly about keeping balance, breading, directing both eyes to that location and focusing, and so on. The fact that we don't feel anything when those kinds of processing takes place does not mean necessarily that no feeling is associated with them. Perhaps there are huge amounts of feelings and consciousness in my brain, associated with every small (automatic) function, and I just am spared of feeling them, due to some mechanism in my brain that allows the "director" part to be alienated from all the rest and concentrate on the big picture.

This hypothesis would be compatible with:

• every sufficiently complex automaton is conscious.
  
• we are nothing but complex automata (that describe themselves as conscious and free).
  
• digital computers are in fact conscious.

I for one don't believe this due to a whole different set of data. There is, I think, sufficient data on paranormal perception to strongly suggest that we are capable, sometimes, to anticipate future events or communicate feelings telepathically (even Turing agreed with this last one).

This ability seems difficult to reconcile with a fully deterministic description of our brain. If paranormal abilities turn out to be real, then the brain cannot be wholly described as processing unit having as input the 5 senses. There must be something else. And that is one thing digital computers would be missing, the ability to be influenced by something non-local.

Of course all of this is nothing but a bunch of conjectures. What seems clear today is that we don't know how far artificial computers will be able to go. In any case, I wouldn't be surprise if digital computers will teach creative drawing in the near future. True creativity cannot be taught anyway. Also, I hope that, in the long run, sufficiently complex computers can be built in order, not only to emulate intelligence, but to host consciousness. That would bring a whole new dimension of discovery. Just imagine: a life that could last for millions of years, we don't need to kill to live, we don't need words to communicate, we don't even need to stay fixed on a planet... it'd be a whole new horizon.

Edited by moderator, 02 April 2012 - 03:46 AM.

[mrdee]

The first reason why we are more intelligent than computers is the fact we invented them.
That being the case, we should know and understand how thy work and operate so we should be in control of them.

Just to elaborate: people sometimes speak about "intelligent machinery", while a computer in itslef is basically a stupid, ignorant thing.
All it can do by itself is add, subtract and compare, and if it has not been instructd in great detail by humans what to do, it will no be able to do anything.

A few examples:

Imagine a robot that has been programmed to understand voice commands, you tell it: "bring me the butter", it will pick up the butter and bring it to you without letting the butter go and/or putting it down in front of you.
The command you meant to give it (and which you assumed the robot would understand, as humans have no problem with the command) was, "bring me the butter and give it to me", but the robot recognised and therefore executed the command "bring" and did just that.

A second example: In som countries, traffic lights are referred to as "red lights", so, imagine a car with voice command abilities:you tell it: "drive up to the first red lights and turn left there", the car goes off, comes at the first set of traffic lights, but the lights are green, so it drives on.
Only the 4th set of traffic lights it comes to are red, and there it turns left, much too late of course.

The point being that a computer has to be told exactly and into the smallest detail what you want it to do, or it will not do it, or at least do it wrong.
Rpgsearcherz also mentioned about computers "learning", I wonder what exactly they learn, they just execute a command that is given to them, and if a human changes that command (eg. a line in the program), it will execute it as the new line commands it to, without understanding that change or what it is doing.

Granted, a computer "brain" can execute certain commands much faster than a human brain.

However, all the above can also take away a lot of blame from computers too.
We sometimes hear; "Due to a computer error your tax rebate will be two months late".
NONSENSE!!!
Computers do not make any errors, they do as they are told, and if an error comes out, it means the one who programmed the computer or the one who entered the data put the error in there.

If the program is correctly made and the data are entered correctly, the computer will also do what it is supposed to do correctly, ie. do the correct calculation(s) and output the expected data correctly.

Finally, thank goodness for what I said in the beginning of this post about us being in control of the machines we invented.
Imagine they had, or started having a life and intelligence of their own, I daren't think of the consequences.

[multfilm]

Humans are superior to computers because computers cannot identify CAPTCHAs! Of course, that's what it's all about! --err, by "it" I mean "differences between computers and humans in identifying CAPTCHAs", j/k.

But on a more serious note, computers cannot distinguish quality (i.e. expensive vs. cheap violin sound); computers cannot generalize (teach computer how to add and multiply, and have it generalize it... of course, humans know that it generalizes to exponentiation).