Artificial Intelligence theory

originally posted by: verschickter

originally posted by: turbonium1
You are the one who is talking in absolutes, by saying software can be 'smart' or 'intelligent'.

This is a joke right?

If you seriously believe software has intelligence, that's a joke, for sure.

What do you believe Mr. Software has knowledge of, with his intellect?

a reply to: turbonium1

I wrote CAN, that means there is the possibility. That´s not an absolute. You wrote it can´t (as in, never will be). That´s an absolute.

I won´t argue with you, keep your google-fu knowledge, I don´t need it. Armchair warrior...

a reply to: charlyv
Try to see it this way:

There are really no thoughts "out of the blue". For your consciousness, it is out of the blue, while your subconsciousness was working on it for a while. You (as in the conscious part) just didn´t notice that until that thought surfaced.

originally posted by: verschickter
a reply to: charlyv
Try to see it this way:

There are really no thoughts "out of the blue". For your consciousness, it is out of the blue, while your subconsciousness was working on it for a while. You (as in the conscious part) just didn´t notice that until that thought surfaced.

It is all the same brain, compartmentalized for functional multi-tasking. The point still holds. Irregardless of conscious or subconsciousness, the trigger for an unsolicited thought involves an analog of awareness where the very thought is created spontaneously. Currently, AI has no hardware / software mechanism that can perform this fundamental requirement of intelligence.

It would be akin to a CPU generating an interrupt for no other reason except that it wanted to.

originally posted by: verschickter
a reply to: charlyv
Try to see it this way:

There are really no thoughts "out of the blue". For your consciousness, it is out of the blue, while your subconsciousness was working on it for a while. You (as in the conscious part) just didn´t notice that until that thought surfaced.

Plus that issue that our cognitive awareness is on a 15 millisecond delay...

I see a brain as a computer. A big and powerful computer. Nothing more.
So we could simulate a brain to create an AI. This is not possible yet because we need 100x more powerful computer than those we have now. That is only a matter of 10 years.
AI's are going to be much better than humans in every domain. That could be the end for our species.
Pseudo AI's are used in medicine today because in some areas they are better than humans, ex in medical imaging diagnostics.

Then, I conclude with this citation of EDSGER W. DIJKSTRA :

"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim. "

originally posted by: LedermanStudio

originally posted by: verschickter

So, thank you for all of the effort you have added to this post.

I realize that what I wrote was NOT in keeping with the actual reality of current AI development.

Also, I did get to read a bit more since the OP. ...not that a whole 24 hours has made me any more expert

Let me ask this:
The early stages of AI development seem to preclude that my thesis from the OP can really be a part of the later stages?

To oversimplify... early AI development is so fundamentally black and white that any hypoyhetical future evolution is 'innoculated' from the kind of influence I postulated?

thanks!

To elaborate more on what was written...

AI operates on very rigid rules in some ways... the "core" of it is computer code, which has to follow certain patterns and requirements that "sort of" mimic the biology of humans but isn't very similar. To give you a bad example, we might be stressed ... and that releases adrenaline and other hormones that make our perception of time and space change slightly and would affect our opinion or behavior.

You can stress a computer all you like, but it's not going to have caffeine or adrenaline or cortisol overload affect how it computes or how it behaves.

AI exists as a single entity. We are actually a species that lives in family groups within a larger community.

Computer architecture HAS to be built in a certain way. Human architecture... is all over the map.

AI can appear to be very "smart" but it relies on a lot of things and is not going to ever overtop or replace humans within the foreseeable future because it's simply not that flexible. This might change if they merge it with a biological system (as actually described in the first STAR TREK book by James Blish (adapting the episode, "Miri") where Spock uses a computer that's integrated with a cat's brain - though what Spock got was a very fast machine that did complex chemical analysis and crunched out results that would take a human a week to sort through.)

Smart AI is most likely to look (in the end) like the Starship Enterprise -- and only after a long time will it react like Commander Data (to use examples you might be familiar with.)

a reply to: turbonium1

So animal are not intelligent ?

What about all the people who never build any machine ?

A human needs intelligence to sort mail

So human can't sort mail the way a machine sort mail ?

You use the capability of intelligence to create the capability of sorting mail.
Even if the capability you create for the machine to sort mail is different from the capability of sorting mail for the human , its just another capability.

Maybe to talk you have to building robots that talk like humans , maybe not. Its what they can do that matters. The capability.

We are NOT there yet. (but that wasn't your point).

originally posted by: Byrd

originally posted by: LedermanStudio

originally posted by: verschickter

So, thank you for all of the effort you have added to this post.

I realize that what I wrote was NOT in keeping with the actual reality of current AI development.

Also, I did get to read a bit more since the OP. ...not that a whole 24 hours has made me any more expert

Let me ask this:
The early stages of AI development seem to preclude that my thesis from the OP can really be a part of the later stages?

To oversimplify... early AI development is so fundamentally black and white that any hypoyhetical future evolution is 'innoculated' from the kind of influence I postulated?

thanks!

Smart AI is most likely to look (in the end) like the Starship Enterprise -- and only after a long time will it react like Commander Data (to use examples you might be familiar with.)

I have an Android phone with a Data plan... is that what you mean?


Just goofing
I get what you mean, thank you!

I think that it would be impossible to create a machine that has human-like intelligence without giving them some kind of emotion, which I see as challenging but not impossible. All living things have a drive to live and survive, and are highly motivated to learn to do things that will allow that to happen.

I've wondered if it might be possible to integrate a kind of tamogatchi-type emotional construct into an AI system so it will carry around its own motivation to learn and get better. Program it to enjoy the company of people, and give it a drive to make their creators happy. It could be a variable scale that would rise and fall depending on the type and quality of interactions between itself and the people around it, and program in that if it gets below a certain level it will prioritize activities that will increase that level. Give it a warm furry body like a cat, but with sensors that indicate when it's being petted and when it's being beaten. Make it purr when properly petted and meow when it needs more petting.

I'm not sure how many of these kinds of scales it would need, but it probably wouldn't take all that many. Give it motivation and a way for it to address its own needs.

Of course, that's the beginning of the end, and not long after that we'll all be enslaved or recycled into cat food. Still... it would be a fun experiment.

originally posted by: verschickter
There are really no thoughts "out of the blue". For your consciousness, it is out of the blue, while your subconsciousness was working on it for a while. You (as in the conscious part) just didn´t notice that until that thought surfaced.

Are psychologists still using those old structure-functionalist terms like "subconscious?" No wonder so many people consider them to be the equivalent of voodoo witch doctors.

a reply to: Blue Shift
English is not my first language, how many languages do you speak 100% perfectly?

originally posted by: verschickter
a reply to: Blue Shift
English is not my first language, how many languages do you speak 100% perfectly?

I'm not talking about your use of the language, which is fine, but are people still using such old concepts as "subconscious" and "conscious" to describe supposed structures of intelligence?

The point being that here we are on the verge of creating machine intelligence, and possibly even sentience, and we don't really even have a good understanding of how our own minds work.

a reply to: Blue Shift

Look into hypnosis, you´ll learn much from it regarding those old concepts, like you call it.

What are the new concepts you are talking about? I´m curious.

originally posted by: dude1
a reply to: turbonium1

So animal are not intelligent ?

What about all the people who never build any machine ?

A human needs intelligence to sort mail

So human can't sort mail the way a machine sort mail ?

You use the capability of intelligence to create the capability of sorting mail.
Even if the capability you create for the machine to sort mail is different from the capability of sorting mail for the human , its just another capability.

Maybe to talk you have to building robots that talk like humans , maybe not. Its what they can do that matters. The capability.

We are NOT there yet. (but that wasn't your point).

I said animals have intelligence, I don't know where you got the idea I said they don't.

And I have no idea what you're disputing, beyond that.

Intelligence is in life forms, not machines, or robots, or computers. You seem to agree with that, in your post, so exactly what is your point supposed to be, here....?

originally posted by: Blue Shift

originally posted by: verschickter
a reply to: Blue Shift
English is not my first language, how many languages do you speak 100% perfectly?

I'm not talking about your use of the language, which is fine, but are people still using such old concepts as "subconscious" and "conscious" to describe supposed structures of intelligence?

The point being that here we are on the verge of creating machine intelligence, and possibly even sentience, and we don't really even have a good understanding of how our own minds work.

Machines do not, cannot, have intelligence. There is no machine, or robot, or computer, or any other man-made device, which will ever have intelligence, or sentience.

Many people, such as yourself, confuse intelligence with machine/computer/robot functionality. In that, machines with ever greater complexity, greater functionality, makes a machine ever more closer, to having it's own 'intelligence'.

A dog, or a cat, is considered to have intelligence. They don't know what city they live in, or what a city even IS, to have intelligence! A computer can tell you what a city is, where the word originated from, and the population of every city in the world....without any intelligence.

A dog or cat doesn't need to know squat, because intelligence has nothing to do with complex functionality, complex processing of information, or saying 'Hello, I'm Siri. How may I help you, Mr. Smith?'

It certainly looks impressive, but it's not closer to being 'intelligent'.


Do you understand the very distinct difference here?

The main confusion is that machines which mimic humans appear closer to being human, which people associate with being intelligent. A robot dog barks, or runs around, or looks, much like a real dog.....but nobody looks at a robot dog being ever closer to having intelligence.

Don't be fooled because of more complex, human-like, machines.

Great post OP.
Perhaps it is more of a spiritual matter.
Human beings are prone to all manner of ideas of reference and errors of thinking.
A volatile soup of our hearts and minds.

Ideas of the spirit appear to be a persistent co-evolutionary component of civilization and technology.
If it were possible to 'scientifically' understand our selves in a manner that could be propigated in our 'creations' , then it may e possible to create an AI with the same sensibilities as the best among us.
If not, the 'machine learning' would be tantamount to abuse, and would most likely result in a horrific mess.

It makes perfect sense that a symbiotic relationship must arise from the pursuits of 'AI'.
I'm not sure it is 'artificial' at all, and it is hubris to believe the we are generally in control of what is most likely 'evolution'.

As a species, we seem to have been compelled to develop technology.
It will take the 'processing' of all the pain of the mistakes of history, to not pollute the galaxy with our collective disorders.

a reply to: Blue Shift

I thought maybe you want to give some info on what you wrote but it seems it was just a hit and run.

AI ... the future of Government. After decades of failure by human politicians ... here come the idea from Italy.
To have a robotic AI-driven government ! Especially considering the success of Saudi robot Sophia.

This way, nobody can blame, any person who's really running the show ... Isn't this the most spectacular idea that ever struck Italian minds ?