What do you think and why?

My friend and I have had this battle for quite a while. He claims that they can think and have been able to for years, while I say that they do not, but follow a set of pre-programmed instructions.

As an example, he heard that scientists built a robot and told it clime stairs. After making many failures, it eventually figured out how to climb them. But is this an original thought, or did it work out subroutines it had been programmed with originally to accomplish this task?

And what is thinking anyway? If one defines thinking as a decision based system using a set of rules one is given, then I say sure, computers can think. I mean hell, Linux can solve many problems by itself, and even Windoze can, (although I think it runs a few cans short of a six pack).

Oh well, now I am rambling. So what do you guys think?

What is it, Resident Evil?

Nope, I think the 13th Floor is more appropriate.

"thinking machine" is a contradiction by itself.

there is no one (in their right minds) working on actual AI it is widely considered impossible...but then again so was flying until that changed...current technology simply cannot do it nor do i believe any direct offshoots of current technology would be able to

A very detailed proof(a not-so-mathematical proof) of why it's
not possible is given in "shadows of the mind" by one of the
most intelligent brain of our time , Rogers Penrose....

for those who dont know,he was the discoverer of impossible
triangle....

I'm not going to answer the question of whether computers can think, but I am going to say that learning is not necessarily the same thing as thinking. The example of the 'machine' that could climb stairs is not particularly new... heck, there are electric wheelchairs (for the disabled) that can do that! There was recently (and still ongoing, I believe) a setup where you could download the (fairly) simple schematics to create your own 'robot' that communicated with others of its kind over the internet (obviously connecting to your computer and using software aswell) - the twist was that they didn't have a language to communicate with... they had to develop and learn a language. I can't remember where on the internet it is, but there's probably a link to it from New Scientist's page - that's where I read the article. It was pretty cool, actually - those in Japan that were running the 'program' had 'taught' their robots some Japanese words, and those running it in the USA had taught theirs some English words... when they were connected together over the internet, the robots started to develop their own language that used words they had already been taught from both languages, and they even developed their own words for new situations presented to them!

Erm, where was I? Oh, yeah! Thinking =/= Learning.

I'm going to stop drinking coffee for a while now...

Thanks to everyone who has participated.

I figure that computers cannot truely think the same way we do, but they do have the ability to learn, which can be construed as thinking.

Perhaps consciousness is what my friend is thinking about. Truely, no computer can have that.

If a program gathers results, then modifies itself based on those results. That could very well be a form of learning. Sure it may just be if-then-else statements, but after a while it "learns" by adding new variables and testing those variables, then gathering those results.

Isnt that what babies do? They test something to see what will happen then adjust to fit the outcomming results based on its experience gathered. AI could do the same thing but that would just be "simulated learning."

However there isnt such a thing as "Learning AI" yet. All present AI is still incrementing and decrementing weighting variables on alogorthims based on specific and predefined data.

There are robots/computers that can mimic those actions very well up to the point that they fool the human operators, but then the robots go into diagnostic testing mode and test all of their servos, motors, and sensors. That kills the momentary charade of AI robot/computer for the operators. Sometimes you just want to belive they are really "alive".

I think that if someday we can model the human brain exactly then computers will learn and think just as us. Of course I think feelings would play a big part in all this. But thse feelings would be something gotten from modeling our brain exactly.

relational? if so yes

seen AI? it seems we haven't a clue what thinking or machines are.

not many people say that there are project from academic world of fisic that study the way for build a computer that can think but only on the question if an fact it's true or false.
Not easy it's the problem to solve on true or false because a person can tell a lie but know what is true and what is false in the moment on which tell a lie. while a computer cannot destinguish the defference from what is true to what's false because i can tell a lie but for you a can tell the true you don't know if I tell a lie!! and the computer in front to this problem krash!.
logics programming for Robots are differents from computers.
The neural computers and fuzy logic get to the people an overview about what is a method for programming systems that can! think.
Computers can think ???

No !!

Regards

Some things to "think" about (heheh):

What is thinking? How we define that determines the answer to the question: Can machines think?

What is a virus? Some biologists say they are alive, some say no. If it's not alive, what is it? A machine made of protein? Whatever they are, they are very successful at what they do.

Now let us look at things computers have been able to do:
1. Discover and prove mathematical theorems
2. Provide specialized advice, such as medical diagnoses or legal advice.
3. Drive a car on an interstate.

my own two bits...

isn't it not the machine that "thinks" (if it could think), but the code? the machine without the code is worthless, and the code is still somthing that a programer writes. I think it comes back down to the learning thing. If you can write code that will 'learn' and write it'self, then we might be getting somwhere. But right now, because all code and action is binary, a simple testing of true or false, i think we have a ways to go in computer design before we get to "real" artificial inteligence.

did any of that make sense?

What part of a person does the thinking? It's not really the body that thinks, but their brain, right?

The way neurons work can be simplistically described as an electrochemical process, with chemicals being released once a certain electrical potential is reached. (sound like 0's and 1's?)

There doesn't seem to be a single point we can point to and say "there's where consciousness is". It's more likely that consciousness is a result of having all those simple cells together, communicating.

But there has been a lot of work with programs that mate, reproduce, mutate and evolve inside a computer. Whether this can be called "alive" or not, they allow scientists to model and study the way populations evolve. Some interesting conclusions from these studies are that parasites are very important for organisms to development, that mutation is not as important to evolution as previously thought, but sexual reproduction is.

Sorry if I'm excited about this topic, I just read some books and I'm still in "golly-gee-whiz-wow" mode.