Walden Two For AIs

Walden Two For AIs:

Could a computer housed in a robot powered by electricity ever express non-trivial Will sufficient to substitute for a guiding, artistic Mathematician?
****

By itself, math does not function (or “math”). To function in order to resolve concerns addressed to it, math requires a mathematical Will, or mathematician.
Such Will is needed in order:
1) To interact with an environment (or holography for recording reactions and sensations, leading to means for storing representations of previous events and of possible choices among sequential events);
2) To appreciate and identify non-trivial (artistic) concerns, problems, and possible choices;
3) To provide a working outline or definition (“handle”) to guide how to recognize what may constitute a tastefully approved solution;
4) To energize, program, and guide mathematical operations to search for such a solution;
5) To weigh (subjective to its own artistic taste) whether a non-trivially proposed solution is indeed satisfactory.

Given a skillfully designed or evolved set of interacting algorithms, represented and housed in an electricity-conducting-chassis, could a sufficient “fundamental inclination” be programmed in order “to breathe” consciousness of life to associate with interacting algorithms --- perhaps even to empower such set of algorithms to re-invent or surpass themselves?

In what fundamental way, if any, is a brain, such as a primate’s, different from a computing set of interacting algorithms represented and housed in a robot powered by electricity?

Is consciousness of Will a special sort of fire that can only be transferred from one fire to a next? Maybe so. Perhaps there exists a fundamental, synchronizing Source of Will, which is ambiguously ineffable, yet superior to math, which is superior to physics. Perhaps IT consents to assume and transfer among varying perspectives (us) and contexts of holography.

Intuitively , such transfers can occur by means other than sexual or biological reproduction (i.e., by artificial means, such as computer programming, even S-R conditioning of robots ).

Walden Two for positive reinforcement of Robots, anyone?

*****

HALTING PROBLEM:

Fundamental problem for AI: How to program (or nurture a childhood for) AI in order to infuse it with an historical (subjective) sense of non-trivial (artistic) preferences and capacity for choosing among new paths of artistic adventure.

Thus far, has algorithmic programming advanced beyond a primitive state of being able to program only digital (on-off, yes-no, either-or) responses to rigorously defined problems (mathematically closed recipes)?

How could a rigorously closed recipe, in itself, apart from cooperative, synchronous involvement with a wider holography, guide the baking of a cake that is not closed to appreciating a wider environment?

Will programming skills help ever to devise an interacting system of algorithms, with capacity for appreciating ambiguity and becoming conditioned for formulating artistically evolving tastes and preferences in response to unclosed stimulation from a wider environment (It’s alive!)?

Perhaps, such programming attempts cannot be successful, absent humble receptivity to SYNCHRONIZING GUIDANCE from a wider holography. When we succeed in reproducing consciousness, it will be in respect of means more than mere “programming.”

****
LIMITS OF COMPUTER PROGRAMMING:

SNIPPITS from http://www.theuniversityconcourse.com/IV,2,11-18-1998/Kovach.htm :

There are a number of problems that computers, no matter how powerful they become, will never be able to solve.
In this article, I wish to briefly illustrate the limits to the computing power, to separate the true potential of these machines from popular fiction.
Turing demonstrated that TMs are capable of performing the steps necessary to solve problems. The only stipulation is that the problem must be represented by an algorithm, that is, by a "recipe" of how to solve it. No algorithm has been found that a TM cannot implement.
… any algorithm that can be carried out by humans can be carried out by some TM.
Since there is no mathematical method of representing the Church Turing thesis, it has not been mathematically proven.

[PERSONAL COMMENT: “Being” facilitates ambiguous “Becoming.” If “being human” (or “being” anything) encompasses an ambiguity beyond mere expression of a predetermined path of interaction among rigorously trivial recipes for interacting algorithms, then the preceding paragraph (Church-Turing Thesis) seems qualified. That is, for a TM to have the same capacity as a human to carry out an algorithm, TM would need also to have capacity to express choices among algorithms preferred to be carried out. But, from whence are such choices derived, since they cannot be “choices” if trivially, entirely entailed within a programming recipe that affords no respect for the role of a wider holographic environment.]

The fact that computers can only solve only problems that have algorithm very greatly limits their power. We humans solve problems constantly without using algorithms. We usually call this intuition or imagination.

There exists a group of problems whose corresponding languages are called recursive ennumerable. TM reject non-solutions of these languages either by returning a no or by running forever. Since you don't know how many steps are needed for a yes, you do not know whether the TM is going to run infinitely long or whether it has not gotten to the answer yet. These problems are called unsolvable problems, because in general you cannot tell whether or not we will get an answer to them. The most famous of these problems is the "halting problem" which has just been described. That is, it is impossible to tell whether or when a TM working on a particular problem will halt with an answer.

Solvable problems belong to the group of languages called recursive. For these problems the TM will return a definite answer yes or no in a finite period of time. Yet many of these cannot be practically solved because of limits of time or space.

A particularly simple example of this is the traveling salesperson problem. There is a salesperson who has to visit 50 cities. He wishes to do so by traveling the least number of miles and by not visiting any city more than once. The algorithm for this problems is deceptively simple. Measure and store the distances between the starting point and all 50 cities. Find the shortest distance between each set of the remaining 49 cities and add it to the first distance, then determine which is the shortest distance. Simple huh? Until one considers the time needed to solve this problem. This problem requires at least 50! steps to solve it. (50!, read 50 factorial, is the product of 50x49x 48x47x...x3x2x1. It is approximately 314 followed by 64 zeros. To give you a feel for the time needed to solve this, assume a computer could perform 10 billion steps a second. This would translate into approximately 31.5 quadrillion steps in a year. At this rate, it would take "a little" more than 9,650,000,000,000,000,000,000,000, 000,000,000,000,000,000,000,000,000 years to solve this problem. (In Asimov's story, the universe ends in a mere 10,000,000,000,000 years.) Thus even solvable problems may be beyond the power of computers.

But, what about Big Blue beating Kasparov and computers proving unproven mathematical theorems? In many cases in which a computer appears to use reasoning, it is simply pattern matching, nothing more.

So where does all this lead us? First, the power of a computer is limited, not only by its construction, but also by its very nature. There are problems that a computer will never be able to solve: Are there odd perfect numbers? Does God exist? Second, anything that a computer can do, a human can do, given enough time and resources. Hence we should not look for a computer to solve problems that we cannot solve ourselves. Thirdly, there are human abilities that are beyond the powers of a computer. Intuition is one of them.

[PERSONAL COMMENT: Human-like consciousness and intuition (AI) will eventually be reproducible in purely robotic chassis. When accomplished, it will be by leveraging intuition and skills beyond mere recipe programming.]