Illwinter is ahead of the Computer Science Field by CENTURIES
 

From the progress page:

7th march 2009
* AI learned how to use the seduction, lure and corrupt special abilities.

-----------

Creating an AI that learns would be impressive enough... but one that has learned to seduce or corrupt others; I'm astounded!

On a more serious note, another round of applause for these guys. :clap: Both for giving us glimpses of a more challenging SP environement and keeping the community so well informed. Thanks again!

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

That is impressive indeed. One step closer to a true A.I, one step closer to ditching my so called friends... :)

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

There is a distinction between 'learned' and 'was taught'.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

this is true. in machine learning we say that we use a data set to `train' a machine to `learn' to predict the class of an instance given a set of attributes. However, the machine isn't trained or learned in any colloquial sense of the word, for it doesn't even know what the data is or what it means.

Consider a robot with algorithms to follow a path and avoid obstacles. First, an algorithm doesn't path or avoid, rather a concept does. The robot doesn't have this concept, rather the programmer does. The robot doesn't know what a path or obstacle is, it is simply following an algorithm. Thus the AI really doesn't know what a path or obstacle is, even though it can follow a path and avoid an obstacle.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

What exactly is meaning though? What exactly is a concept? Isn't it good enough for us to program an AI to act like it understands the meaning of something? Could we know the difference between an AI that understands and one that doesn't if they both act identically the same?

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

Because true "AI", that is, "Intelligence" can analyze and understand - in order to further learn on its own.

A "Program", on the other hand, no matter how well we design it to mimic an intelligent action, is still just that - a Program - and will not ever learn from its activities.

There is a grey area of pseudo-AI, where a program is designed to define results as positive or negative, and to rebuild its own algorithms in order to behave more efficiently - but it is still doing so only within the narrow framework of the programming, and until it decides to extrapolate beyond that, and start figuring things out for itself - it is simply not "Intelligent".

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

But what is this learning we speak of? Is it not a structure of inherent behaviours used to evaluate a given choice at a given time? Are we not with out prescience our selves?

Therefore our own actions are essentially reasoned conclusions given the parameters of our experiences. A program functions on the premise of best course of action defined by the programmer given a set of circumstances. We ourselves reason along these lines - only able to perform that what we know and only perform that which we think is the best course of action.

We attain more sophistication by experience; a program increases capacity through algorithmic development. Surely the two processes are very similar? But of the containers that shunts these processes about - the bone water bags and the note books, well, only one of them needs to pee....

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

hEad v 3.25.6, When humans gather and/or re-evaluate information, the premises behind our actions and decisions change, automatically, all the time.


This is much easier than updating a program, hEad v 3.25.7, wouldn't you agree? ;)

Of course they are. Program increases its capacity when programmers attain more sophistication. Both processes depend on a human's learning.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

I like how your numbering system implies hEad has learned about 3000 sentences worth of information so far in his life. :)

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

I am still in beta, so take my opinions with a grain of salt. :p
To say that Artificial Intelligence will never attain Intelligence is very short sighted. I think that it is just a matter of time.

Personally, I am building a portable EMP device in my basement so that I can join the resistance when the time comes. :D

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

At this point you have passed beyond Dominions 3 geekiness level and there is no more I can do for you.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

and counting.... ;)

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

Cor.. that one begs to be prodded by a discussion on the origins of causality. Indeed, perhaps our own learning is dependent on the actions of a greater force external to ourselves – plenty of ideas to suggest that man is not the sovereign agent he believes himself to be. Man certainly has the monopoly on efficiency but his intelligence is not alone in its capacity to respond to stimuli.

Anyway, no rush. I haven’t seen a movie or heard a theory yet that doesn’t portend bad news for humanity if AI gets its 1’s in front of its 0’s by itself!

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

Perhaps you're all a hallucination of mine.

You cannot debate what is possible, but unprovable. You can only debate what can be shown to be true, at least to our perception. Since we can perceive anything that we put our minds to - we are intelligent. The machine only perceives what we tell it to - it is not intelligent.

Since people's perspectives can change over their lifetime, and indeed, instantly - and we cannot detect any "greater force" influencing that activity - the only sane assumption that can be made, is that we are self determinate.

To put it another way, the machine is not responsible for what it does. The programmer, or operator is responsible for the machine. Humans are responsible for their own actions, and to claim otherwise is recklessly irrational.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

Perspective is a tricky word and one difficult to apply to others. If, as you say, "You can only debate what can be shown to be true", I would not be able to speak to the perspective of others, only to their actions.

Their actions do change over time, depending perhaps on their experiences. But the actions of a sufficiently complicated program can change over time depending on it's inputs (which would correspond to experiences).

It may be that the perceived difference lies only in that we are more complicated and that we can not see our own programming.

I don't actually believe this, but it's not easy to disprove. Not that we are just like computers. At the least our "programming" can be rewritten on the fly (which is not impossible for software, by the way), but the hardware can be modified as well.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

His Eminence Carmont objects to your heresy, and would like to invite you to come sit in the comfortable chair for a while.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

No no, anything but the soft cushions! :shock:



The difference being that you are imminently capable of making logical leaps into heretofore unexplored intellectual territory.

The machine (as we know it today, in 100% of all cases, period), however, is still bound to predefined territory within which it draws its conclusions.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

Or we do not know the extent of our own predefined territory and what appears to be a leap into unexplored territory is just a branch into a section of programming we hadn't used before.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

Something that software does not actually ever do - at least, not with any awareness of such. Software does not understand what it is doing - it simply operates.

Everything in reality is subject to perspective, and perspective is constantly subject to collective agreement.

Therefore, we define ourselves as intelligent - and this means that what we understand of our existence, portrays intelligence. Since intelligence is a human concept, this assertion is true.

However, we can observe and define that what a computer seems to be capable of, is currently inferior to what we are capable of, and in this manner lacks what we understand and perceive to be intelligence.

It can be proven that the machine's "thoughts" are bound by its code - it cannot be proven that our "thoughts" are bound by anything originating outside of our self.

Prove that there are intelligent beings who have at least once directly manipulated human thought processes - and you begin to have an argument. Else, you really are just asking "why?" or "what?" over and over again. Maybe one day, we'll have final answers to those questions - but not this day.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

Ray Fuller invented Prozac, and with the help of Eli Lilly is manipulating millions of human thought processes each day. Ditto with caffeine, E, alcohol, etc. etc. where we use chemicals to achieve very specific effects. More immediately, see the experiments where we use magnetic fields or direct electrical stimulation to trigger different parts of the brain. More scientifically, see the IBM's BlueBrain project to create a full, neuron-level on up map of the human brain. They should be done in a decade or so (though as with the human genome, having the data is still a long way from full understanding).

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

with human-machine interaction, the traditional limitations of AI become softened, and the lines between the human and AI become blurred.

For example, in an experiment where a population of AI agents 'converse' and build a language, if there are humans to give and teach them words, the AI will develop a very sophisticated and expressive language.

To believe that this is 'cheating' somehow because humans have intervened assumes that humans are somehow special and can without outside influence build a language.

did not the material world intervene upon humans to essentially teach them words and meaning? with no material world and no constraints we have no need for language. we can, actually, model the material world as an agent which actively interacts with the human just as the human interacted with the material world in the experiments with language and AI.

maybe we're much closer to AI than we though. The problem is assuming humans are independent from the material world when they are, in fact, not; and assuming the material world has no agency when it, in fact, may have just that. If these things are the case, then the AI from the human-machine interaction language experiment isn't all that far off from 'human intelligence'

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

I think Jim meant "intelligent beings other than humans who have at least once directly manipulated human thought processes". So, no dice on Ray Fuller or the minions rabbiting away at Eli Lilly.

Getting kicked in the head by a moose doesn't count either.

Re: Illwinter is ahead of the Computer Science Field by CENTURIES
 

Indeed. Humans manipulating human thought processes is exactly evidence of intelligence (or determination), and not the other way around.


Omni- Your example is interesting, but still - it resides in that gray area I mentioned, where a computer is able to simulate a "function of" intelligence, but still only does so within the precise framework allotted to it, and does not portray any evidence of conscious understanding of what it has developed - nor does it ever deviate from the format of the programming.

Now, if the computer one day synthesized speech through speakers that were installed, with no provided software, and said "hey doc, I can talk!", well then, we'd have something.