Do you see goals, rules, and intelligence as a cluster of related ideas?

[PseuTonym]

I ask because I get the impression that there might be a popular misconception to the effect that the existence of a known framework of unambiguous rules that is provided to all participants in some activity is enough to remove any need for intelligence or ingenuity from the activity.

Perhaps we should take a step back before approaching the question “Do you see goals, rules, and intelligence as a cluster of related ideas?”

A more basic question: do you make a sharp distinction between achieving a goal and obeying some command?

Perhaps an example will help:

Here is a goal: write a computer program that will ask the user how many digits of pi are to be computed, and then compute them and display them.

Here is a command: given a positive integer written in ordinary base-ten notation, add one, and then write the result below the positive integer that was given.

In practice, it might require some intelligence to merely obey commands. For example, there is a tendency to interpret some commands in terms of some imagined goal of a person who is issuing the commands. It may require the application of some intelligence to recognize a distinction between a command itself, how an average person who is trying to obey the command might interpret it, and how the person who is issuing the command might interpret it.

However, in principle we can recognize that some commands (such as the lines of a computer program), are clear enough that they basically do not require intelligence to obey. After all, we do not think of a computer as being intelligent.

 

[Mary_Estelle]

Is this a test? I think I just flunked Computer Science 101. I think that you have a valid question, but if you could just bring it down a step or two it would be helpful.

 

[Tomdstone]

After all, we do not think of a computer as being intelligent.

It depends on the software used. There was a report of a test asking whether you are communicating with a 13 year old human or to a computer. The interface is by keyboard and monitor in either situation. Real human beings were not always able to determine the answer.
And there are chess programs which are able to defeat all but the top 1% of world chess players. Does it take intelligence to play a chess game?

 

[PseuTonym]

if you could just bring it down a step or two it would be helpful.

If you would quote something in particular that is problematic, then it would be easier for me to be helpful.

Maybe an example that seems to split hairs might be helpful to distinguish between pursuing a goal and obeying a command.

Consider the following list of items: 0, 1, 2, 3, 4, 5, 6

I claim that if I ask you how many items are in the list, then I am asking you to pursue a goal.

The reason it is a goal is that there are at least two different approaches to determining the number of items in the list.

First approach:
Notice that the list includes the numbers 1, 2, 3, 4, 5, 6, which count themselves. So now we convert the 0 to a 7, which gives us the total number of items in the list.

Second approach:
Add one to every number, giving a new list, with the 6 in the old list being converted to 7 in the new list, which gives us the total number of items in the list.

Of course, in this very simple situation it makes absolutely no difference which method is used. However, they are different methods, and I did not provide you with instructions that specified which method you are to use. Thus, I assigned to you a goal. I neglected to provide unambiguous instructions that could be considered commands.

 

[PseuTonym]

In practice, it might require some intelligence to merely obey commands. For example, …]

However, in principle we can recognize that some commands (such as the lines of a computer program), are clear enough that they basically do not require intelligence to obey. After all, we do not think of a computer as being intelligent.

It depends on the software used.

Are there computer languages that allow a line in a computer program to assign a goal to the computer, and are there computers that have been designed so that they have the ability to pursue such a goal, without any need for the software to specify explicit commands?

 

[Tomdstone]

Are there computer languages that allow a line in a computer program to assign a goal to the computer, and are there computers that have been designed so that they have the ability to pursue such a goal, without any need for the software to specify explicit commands?

The goal of a computer chess program is to win the game.

 

[Mary_Estelle]

You can ask someone to pursue a goal, but that does not mean that the goal will be pursed. I can pursue a goal that is unachievable, beyond the laws of physics or science.

 

[PseuTonym]

You can ask someone to pursue a goal, but that does not mean that the goal will be pursued.

I agree. Another possibility is that A wants B to pursue a goal, and presumably A wants B to achieve the goal, but A spreads information that has the known effect of influencing B to use a method of pursuing the goal that will in many cases prevent B from succeeding. That may sound implausible, but I have an example.

The SAT (scholastic aptitude test) changes from year to year. The organization that has the power and responsibility to decide what those changes will be has a history of telling students that the test is not coachable, and that the best way for them to prepare for the test is to simply minimize their number of misconceptions and maximize their level of achievement in their regular formal education classes.

However, a for-profit organization (such as Kaplan SAT prep) might have some statistical evidence that many students improved their SAT results via the coaching that it provides. In fact, such evidence might persuade the organization in control of updating the test to make a major overhaul of the test, and the organization might continue to claim that coaching is not an effective method to improve a student’s test results.

I can pursue a goal that is unachievable, beyond the laws of physics or science.

I agree. For example, you might attempt to design program that plays against anybody who wishes to participate in a game of tic tac toe, with your goal being that the program will always win and never draw. You cannot succeed.

You could create a chess-playing program, and your goal could be for the program to draw, and if the program is unable to draw, then your goal could be for the program to lose. Winning could be a last resort.

Whether or not your chess-playing program would itself have a goal is something that I do not know. Do you think that it would or could have a goal?

 

[Mary_Estelle]

No, a program will only respond to my goals. It is not true artificial intelligence yet and able of setting it’s own goals, It still needs what I feed into it even if it is a very sophisticate, complex “perfect” program. I really hope that artificial intelligence is never truly achieved. I want my or any computer to work for me, not think for me or make arbitrary decisions. Should ever such a program be achieved “It” might decide that humans are parasites and, have a good colonic cleansing and eliminate them from the earth. I wouldn’t like that. I suppose in a science fantasy genre’ computers could regenerate themselves, but to what end? It is like trying to find the “God Partical” or figure out “The String Theory.” While great discoveries are being made, I question “to what end?”
Just for the fun of it I will toss this in; “Adam and Eve ate fruit from The Tree of Knowledge” and what did that result in? Oh, I can just see it now; the earth overrun with little walking, talking, replicating naughty, nerdy computers! Anyway, it is all moot in my lifetime.

 

[PseuTonym]

Just for the fun of it I will toss this in; “Adam and Eve ate fruit from The Tree of Knowledge” and what did that result in? Oh, I can just see it now; the earth overrun with little walking, talking, replicating naughty, nerdy computers!

Are you thinking of a Tree of Knowledge where the knowledge we are talking about is not knowledge of good and evil …

… but instead is knowledge of how to create robots that do not have emotions or conscious experience but that do have goals and intelligence?

 

[Mary_Estelle]

OH, I just tossed in the Adam and Eve reference just for fun. BTW, how do you know that YOU are not a computer construct? Just tossing THAT in for fun, too.

P. S. Wen God created man He made Adam from a hunk of mud. When God created Eve he used a pile of calcium(Adam’s rib). So which had the most value; the hunk of mud or the calcium mineral? I am sure you will come up with something, but it is fun to aggravate-LOL!