Do you think computers will ever be able to think for themselves or have “real” intelligence?

[cupidkisses]

Yes

No

[ausmaster]

here's my opinion being a Software Engineer. Mashines are Definite Finite Automatas (DFAs), that means they are mashines that have only a finite amount of possibilities thst you can predict or measure. Even if you were to give it the most "intelligence" there is only a finite amount of possibilities you can give to mashines. Not like Nondeterministic Finite Automatas, where they have a properly that allows it to choose the next state at random so you could be uncertain what the next state would be. Outside of what computers are programmed to do, it can't really.. learn like humans do. Like of course there are Neural Networks but I dont think you could get a computer that will be as adaptable as humans will. To mimic the human brain, thats a very large and complex task. We might... but I am not certain that is within the bounds of reality.

[Soteris]

Why would we want to mimic human brains in the first place? I mean its cool and all as well as a good experiment to see if we can but humans are not the gold standard for intelligence, we can do better than a fundamentally flawed system.

[ausmaster]

Well yes, it is not the gold standard. Yet if we are not able to mimic it, we cannot go higher. If we want to achieve higher intelligence than the human brain can possess, you firstly mimic the brain to prove you can do it. Then you do the human brain + 1 to prove you can go higher than the hunan brain. We haven't done that yet. Take a super computer, it does calculations thst we can do in a shorter period of time. But it only plays a certain role.

[Soteris]

Problem with that is that they are fundamentally different systems. A computer is built far more modular and logical than neurons. By trying to mimic human brains we are basically losing efficiency by emulating a whole different process instead of focusing on the advantages of using a computer.

[ausmaster]

Well the advantage of using a computer is basically doing specific things of the human brain to help us. The usefulness of a computer goes away when its starts calculating items beyond what we can comprehend. The tipping point of a useful computer is basically the human brain. You can make a super fucking fast human brain perhaps but once you go beyond that point in terms of intelligence, humans won't be able to comprehend. Also if you're not able to program something you cannot comprehend, you cannot program it at all.

[HungLikeAHorsefly]

Some things to nit-pick about:
1. *Deterministic Finite Automata

2. NFA's still have a finite number of possibilities

3. Are you assuming the human brain has infinite possible states? If so, I disagree. It may be very complex, but being a physical machine it has an upper bound to what it can process.

4. Modern *machines* are DFA's, but software is not necessarily. You mentioned neural networks, but there are also side-affecting functions, some cryptographic functions, Bayesian algorithms, logic programming, anything that utilizes an entropy device, trampoline functions and anything whose behavior is dictated by I/O. Also, CPU's are internally deterministic, but not externally. In a sufficiently sophisticated operating system such as Windows or Linux, it is not possible to figure out exactly what instruction a CPU core will execute next.

[slatyb]

The thing about artificial intelligence is that as soon as we figure out how to do something, it no longer seems intelligent. We have computer programs that play go, chess, backgammon, poker, and other games better than any human ever can. We used to call this research "Artifical Intelligence" but now that we've solved it pretty well it doesn't seem so intelligent anymore. Same with speech recognition and image understanding. It's no longer artificial intelligence, it's just applied math and engineering. We have bipedal robots. Very difficult until it wasn't anymore.

[LittleSally]

Yes. For sure. All of their 'thinking' will be done based on actual real facts (unlike many humans)... so it's ultimately infallible (statistically speaking)... Or most probably, if you will...

[Anonymous]

No. They will always have to be programmed by humans to at least some degree.
It's called Artificial Intelligence because it's artificial.

[Mr_Smile]

How do we know our intelligence is real intelligence? For all we know, we are hyper advanced machines created by a race of aliens that have long since gone extinct.

If we define real intelligence as being able to make decisions by iteself and be able to learn and develop it's own programming (similar to how we learn), then they already exist. AI capable of learning, making decisions and developing it's own code already exists.

https://www.youtube.com/watch?v=R9OHn5ZF4Uo&t

[Soteris]

I think we need to grow up from the whole "fake intelligence" thing, there simply is no such thing. Our current computers are far more advanced than many living creatures. Of course there is still some ways to go for human levels of intelligence but not as far as people might think.

[ZanyInternetName]

Well, there's problems that, theoretically, humans will be able to solve but machines never will. Doesn't that mean that there's a difference?

[Soteris]

@ZanyInternetName We are fundamentally different mechanically so of course there are things one do better than the other. Computers are what we call Turing machines and they have limitations like for example paradoxes. A computer can never solve a paradox because it is unsolvable which makes sense. The problem is that while we know a lot about what computers can and can't do we are not that sure about humans. Are we also Turing machines? If humans are also working on the same principle then there is nothing we can do that computers can't. It might be slow and inefficient but its not impossible.

Also the thing with computers is that we can always add a new module to deal with for example non-Turing problems later.

[ZanyInternetName]

Humans are not Turing machines, otherwise we couldn't solve the halting problem, for instance. An old example, sure, but a good indicator for the limits of machines.
From the fundamental difference between humans and computers it follows that the way intelligence should be defined for computers must be different from how it is defined for humans. Since therefore there then has been added a different denotation to an already existing word it may be fair to call what it denotes "fake" insofar as that it is not denoted by the classical definition.

[Soteris]

@ZanyInternetName I have not seen anyone prove that humans can solve the halting problem. As for intelligence it is not really important how something is generated as long as the end result is the same or similar enough.

[johnmspaghetti]

In the truest sense, no. They can only work as programmed given inputs. Someday someone might figure out how to allow computers to program later generations without any guidance, but for now everything is guided by basic unchanging programmed rules.

AI is powerful, but only at making decisions for itself. Those decisions are still this far governed by basic programmed rules.

[Anpu23]

It depends on what you consider "real intelligence" computers out think us logically already. They have the ability to process extremely complicated mathematics at a level that no human could ever match. They are quite poor at intuitive thinking, and horrible at social understanding.

[chriscdi]

Well there was the 2 AIs that were communicating through a made up version of the English language that the AIs created and they understood each other wo they were shut down because no one was able to translate what they were saying, thats soumd really close to what you are looking for

[GreedyGiver651]

A true AI , does not exist yet. I believe it is possible in the future. I wouldn't really call them comuter at that point because, in theory, they have the same brain as we would, with less actual flesh and bone.

[JZ909]

Yes, but I don't think we're very close yet. Humans very strictly determine how computers "think" primarily using math and Boolean logic, because computers literally cannot function without these explicit instructions.

[ZanyInternetName]

Define thinking and real intelligence. Would a perfect simulation of a human brain be able to think and have real intelligence? Some would say yes, some would say no. Soooo…

[nbbn5]

Maybe but I mean still they will be programmed and that has its own frwbacks and limitations
Like their is good home and it lietens to commands and what we say and reply accordingly and it very intelligent lol but has its limitations

[nbbn5]

Drawbacks **
Google home ***
Listens****

[vald9inches]

It will take a lot of work but its possible to create artificial intelligence. how close it will be to human nature is a different debate but once it happens we will wittnese the Singularity which will change EVERYTHING!!!

[Paris13]

As of Now, They do.
They are slowly Replacing... Humans. xx

[JohnPhoton]

Well I am hopeful for it and that it will help us, and us help it kind of relationship.

[Iron_Man]

No mankind is always in charge if we didn't give them the time of day they would eventually stop working

[the_sinner]

If they ever start thinking for themselves or have real intelligence then they won't be considered as computers anymore.

They'd be equivalent to humans.

[AbdKilani]

Yeah
We are already having AI revolution , it is getting out of hand

[MarketData]

We have neural networks that can learn skills better than humans. Learning is the first step to true intelligence.

[ScaryCool]

Yes and I don't think it'll be a good thing.

[ZeussLightningBolt]

No. Computers will never be able to feel emotion

[shubhamchachapara]

If we use human mind in it

[F7777]

no way that is Just a fake news