The problem of designing an AI to understand what it sees

KellyJay

Why do you just keep repeatedly the same assertions over and over again without any evidence or reasoning to support them and that all AI experts (such as myself) know are are just plain wrong and have evidence are just plain wrong? (Example: you asserted that a computer can only do what its programmed to do and yet we gave clear examples of one's that don't). Exactly what do you hope to accomplish by doing this? Do you think we would mysteriously start to be convinced by those assertions merely by you endlessly repeating them? What you don't understand is that we are RATIONAL and therefore the ONLY way we could be convinced is via evidence and/or reasoning. For us, mere say-so accounts for nothing.

@humy said
KellyJay

Why do you just keep repeatedly the same assertions over and over again without any evidence or reasoning to support them and that all AI experts (such as myself) know are are just plain wrong and have evidence are just plain wrong? (Example: you asserted that a computer can only do what its programmed to do and yet we gave clear examples of one's that don't). Exactl ...[text shortened]... ay we could be convinced is via evidence and/or reasoning. For us, mere say-so accounts for nothing.

What a computer will do is only those things they are designed to do; even if they do the unexpected, it can only do those things because of how they were set up with the hardware-software limitations. It is made up of hardware, we install software, the software directs it, and it only does what it is set up to do. Nothing about that elevates the computer to have an awareness of the task it is performing.

The example I gave you ignored was playing chess, it can calculate moves by software to play chess the game of chess, but it is no different from being given code to forecast weather, play a song stored in memory, display a grocery list, it is all the same to the computer regardless of the how great or small the complex nature of the calculations have to be to do anything in it.

There is no understanding within the computer! There is nothing about a computer getting understanding regardless of how insignificantly simple the code once was, to incorporating a more highly complex a new software revision and steppings are; the code is still just telling it what to do, that is not intelligence.

@kellyjay said
What a computer will do is only those things they are designed to do

And what a human will do is only those things he has evolved (or designed by a 'god', if you believe such superstitious nonsense) to be able to do. And yet we wouldn't think a human was incapable of intelligence because of this.
And what some of the specific things an AI did was not what it was designed to do. And we would think this indicates it has intelligence.
You make no point.

@humy said
And what a human will do is only those things he has evolved (or designed by a 'god', if you believe such superstitious nonsense) to be able to do. And yet we wouldn't think a human was incapable of intelligence because of this.
And what some of the specific things an AI did was not what it was designed to do. And we would think this indicates it has intelligence.
You make no point.

I don't believe it is superstitious but an actual fact that we are designed by God.

Back to intelligence, you can pluck a few things off this list below, but it will be just like assigning human traits to animals, it does not mean that it is valid for the animals. Computers do not learn; they perform. Computers do not understand; they perform. Computers are not self-aware they perform. Computers do not know anything about problem-solving; they perform as designed if a problem is solved; it is because we directed it. Computers do not mull about the mysteries of the universe; they perform. Computers do not deny their creators. Computers do not perceive information.

We input data it, so we can do whatever calculations we desire to get, then we need to see the proper output by our design, this is all orchestrated by us as we want, or it is useless. That output cannot vary from our desires, or the computer is not functioning correctly, through a coding issue or some hardware fault, it has to obey the code we put in it.

A definition of Intelligence off the Web:

Intelligence has been defined in many different ways including one's capacity for logic, abstract thought, understanding, self-awareness, communication, learning, emotional knowledge, memory, planning, creativity and problem solving. It can be more generally described as the ability to perceive information, and retain it as knowledge to be applied towards adaptive behaviors within an environment.

Yet gain, you make the same assertions without any evidence or reasoning to support them.
But I guess that's just the religious way of thinking being expressed.

@humy said
Yet gain, you make the same assertions without any evidence or reasoning to support them.
But I guess that's just the religious way of thinking being expressed.

You are doing that, suggesting something is happening within a computer beyond what actually is. I guess that that is the nature of the material mind of man, cannot see things as they are, but must assign a truth to them that is not there.

@humy said
Yet gain, you make the same assertions without any evidence or reasoning to support them.
But I guess that's just the religious way of thinking being expressed.

- Richard Dawkins quotes from BrainyQuote.com "Biology is the study of complicated things that have the appearance of having been designed with a purpose." - Richard Dawkins

Assumptions are made by many about what is right in front of us; with you in this instance, it is that intelligence can be built into the computer, when it is no more than energy acting upon designed components doing what we designed to them to do, with nothing but a little more complexity separating what takes place in a watch, cell phone, calculator, to a desktop, or mainframe.

With Dawkins, he sees the complex nature of life and denies design with a purpose even though the nature of life and its designed qualities are so far beyond anything we can comprehend; it has to come from a mind that dwarfs our minds as the heaven do us while we are standing on the earth looking up.

@KellyJay
But getting back to AI, if one is made that can answer any question put to it by any human and respond in such a way as any human would, like any human genius can, how can you judge the underlying intelligent of that AI?
Of course it is not there yet but I have no doubt it will come to that in the future, based on what revelations we have already seen in the computer world, the advent of the supercomputer, the coming advent of quantum computers, the likely development of the two, classical computing and quantum computing being used side by side with systems a million times stronger than the super computers we have now, which even now calculate billions of trillions of calcs per second.
Think that capability times a million or more which is likely in the next 30 or so years, maybe even sooner.
Every year some new contender for the top dog of the supercomputer world shows up, the fastest in the world, for a while it was China but now back in the hands of the US, FOR NOW, it might be a team from India or Israel or Iran who beats them all next year, no way of telling but it is a clear computer ascendancy competition and it means a lot more than just who has the biggest, fastest and so forth.

@sonhouse said
@KellyJay
But getting back to AI, if one is made that can answer any question put to it by any human and respond in such a way as any human would, like any human genius can, how can you judge the underlying intelligent of that AI?
Of course it is not there yet but I have no doubt it will come to that in the future, based on what revelations we have already seen in the com ...[text shortened]... ascendancy competition and it means a lot more than just who has the biggest, fastest and so forth.

Any question put to it within the scope of the computer data base, or ability to calculate. This is still true of my watch, I look at it and it tells me the time, I go to my level it tells me what is level, and if my line is straight, or anything we design to give us answers or solutions. One of my favorite lines is I have an iron clad memory, I remember all, the down side is it only last a few milli-seconds so I take notes. When I refer to my notes are they intelligent, because they give me the answers I seek? It is always going to be hardware software, speed, but unless it is alive its a tool, not an independent life form that thinks on its own for its own reasons.

@KellyJay
What will you say when computers start making original scientific discoveries?
It's already happening in math, there is a math program that can make new math conjectures. It will be happening in physics too I think.

@sonhouse said
@KellyJay
What will you say when computers start making original scientific discoveries?
It's already happening in math, there is a math program that can make new math conjectures. It will be happening in physics too I think.

When they do anything they will remain what they are a tool we created to behave as we define them. We don’t credit microscopes, or telescopes with intelligence because they show us what we tell them to point at, I don’t credit my cellphone with intelligence when I use the GPS, or the calculator function or any other thing I need to know that it can give me the answers for.

@KellyJay
What do you credit Einstein with? What if you could converse on an intelligent level with an AI and it said, I just figured out how to go faster than the speed of light with zero energy required, and started a printout that proved it and people took that solution and we became interstellar voyagers because of that.
What would you say then?

@sonhouse said
@KellyJay
What do you credit Einstein with? What if you could converse on an intelligent level with an AI and it said, I just figured out how to go faster than the speed of light with zero energy required, and started a printout that proved it and people took that solution and we became interstellar voyagers because of that.
What would you say then?

It doesn't matter, as I pointed out to you that it comes up with answers to questions we want to know. I want to know the percentages of First Pass Yield of product as it goes through a factory, I can use EXCEL after I input the data and get an answer, that does not make EXCEL, or the CPU in my laptop intelligent. Coming up with more complex questions by running the data is still doing what we designed it to do, and if we let it sit unattended and giving it nothing to do, it would not ponder the mysteries of the universe or anything else.

@KellyJay
I am not talking about optimization, I am talking about genuine Nobel prize level inventions AI's might make in a hundred years or sooner.
If Einstein is intelligent enough to shake the foundations of physics a hundred years ago, and an AI did something equally groundbreaking, you would not call that intelligent? If such developments were shown to be beyond human capability by ANYONE on the planet? That would still not be intelligence in your book?

@sonhouse said
@KellyJay
I am not talking about optimization, I am talking about genuine Nobel prize level inventions AI's might make in a hundred years or sooner.
If Einstein is intelligent enough to shake the foundations of physics a hundred years ago, and an AI did something equally groundbreaking, you would not call that intelligent? If such developments were shown to be beyond human capability by ANYONE on the planet? That would still not be intelligence in your book?

I'd call that a great useful amount of time spent by hardware, and software engineering coupled with those people in science who were able to use a device made by man for ground breaking work.