Androids and Gender--Resurrected 11/19!

But it has to evolve to that point.

The first thinking machines will look very robotic. More practicle and less aesthetic. Design comes after application usually.

When the differences are more noticeable, the likelyhood of fear and mistrust is greater.

quote:

Originally posted by Danial But it has to evolve to that point. The first thinking machines will look very robotic. More practicle and less aesthetic. Design comes after application usually. When the differences are more noticeable, the likelyhood of fear and mistrust is greater.

Oh I don't know about that. Honda's Asimo doesn't look all that scary to me. I think this may depend more on cultural values. The japanese love mechanical dolls and robots (no not that way pervert ) In America our usual xenophobia takes over, and we are much more ambivalent. Take a look herehttp://world.honda.com/ASIMO/

Evolution implies a long slow, gradual process. Robots and androids advance by intelligent design, which is a million times faster. The first robots DID look completely unhuman, but that was 20 years ago. Asimo is a humanoid robot which already exists and is capable of a fair amount of human interaction. From what I can tell its biggest problem is its power supply. What will Asimo look like in 10 more years?

quote:

Originally posted by R Trusedale Evolution implies a long slow, gradual process. Robots and androids advance by intelligent design, which is a million times faster. The first robots DID look completely unhuman, but that was 20 years ago. Asimo is a humanoid robot which already exists and is capable of a fair amount of human interaction. From what I can tell its biggest problem is its power supply. What will Asimo look like in 10 more years?

Robots will not necessarily evolve slowly, especially when they begin to tinker with themselves

Thanks for the link!

A link that might be of interest:

The Loebner Awards: The most human-seeming chatbot

I'm bringing this back up now that the excitement of the non-finale is settling down, LOL!

I promised I would say what was up with those pictures I posted of "three lovely ladies"

I think about 12 people answered me and half were right or close enough, and half had no idea.

Those are pictures taken from a website who sell specialty "adult play" dolls. (I'm trying to keep it clean here... )

I first found out about the site because someone sent it to me as something funny, but as I checked it out, I was completely flabbergasted. I can't give out the site here since it's 18-only, but please, take my word for it, there are some demos there and the dolls are quite advanced. (No, NOT demos like that. Ones that show skin and flexibility and that kind of thing)

I posted the pictures to make a point, though. If now, in 2003, we can make incredibly lifelike and "functional dolls", then it's not all that out there to assume that an android carefully crafted to look human could be made equally complete, thus sort of negating the "but it's a machine" argument.

Would we make female androids to eliminate prostitution? Would we get that Futurama thing going where the kids are shown films in health class that you don't get involved with your robot?

If it can be done, SOMEONE is going to do it, that site is evidence of that!

As always, a very thought-provoking post. You may have hit on something important here. In many developing countries AIDS has become a pandemic. The primary vector is heterosexual transmission. A "realistic enough" love doll might very well be the ticket to interrupt this vector, since it could be sterilized between customers. This would be the sexual equivalent of giving out needles to prevent the addicts needle-sharing vector.
I'm deliberately avoiding passing moral judgement. The first priority is to save lives, prevent orphanings, and stop the demolition of the healthcare infrastructures in the developing countries. The alternative is for the developed countries to do nothing, wait for economies to collapse, and then the nightmare shows up on our own doorsteps. (To some extent this already beginning to happen.)
These heterosexual "health stations" would be sited in the middle of all the local red light districts, no doubt upsetting the local talent greatly. Tough. They might also be used to test their customers for AIDs and thereby help determine the magnitude of the problem.

In the context of Big O, the whole issue is a no-brainer. R. Dorothy is female because the flesh-and-blood Dorothy Wayneright was female. R. Instro is male because Amadeus' flesh-and-blood son was male. We know that Instro has this boy's memories because we see one of them at the end of the episode, where the boy's face is reflected in the piano he's playing.

So in the context of Paradigm, the question is more: does anyone there know how to make a creditable artificial inteligence WITHOUT using human memories? The technology may simply not exist.

I'm resurrecting this thread because someone on one of my mailing lists made an excellent case for why a robot or android could not have everything pre-programmed. A chess-playing computer had been brought up at one point in this discussion, and I said that chess was actually relatively simple to program compared to other things.

Although the post doesn't touch on gender per se, a lot of other points were brought up in this thread, so I think it rightfully belongs here. Posted with permission!

quote:

Rene Horn wrote Go is many magnitudes more difficult than chess. With the most obvious algorithm to play any board game, the minimax algorithm, and the manner in which current (as of this year, 2003) computers are built, no computer could ever be built to make even the first move, even if all the matter of the universe were used to build one enormous super-computer. The amount of information that would need to be stored is greater than the highest estimated count of atoms in the universe (in fact, it's about 10^89 times greater.) The minimax algorithm is quite simple actually, and almost never used (except in games with a very small number of possible outcomes, like tic-tac-toe.) A computer that used it could easily play a perfect game all the time, but it's unrealistically demanding. Essentially, it looks at all possible states of the game, starting from the first move, and then branching out, and looks only at branches that eventually evaluate to a winning game. A similar algorithm, the name of which I cannot recall, "prunes" (cuts off) branches that show only losses as end states. Even if this algorithm were used for Go, not too many branches would be pruned. The game of Go itself has very simple rules, but is rather complex (as far as strategizing goes.) If you have fifteen minutes, go to the following link. The best algorithm today for Go is said to be able to beat the average Go player of about one year's experience. Personally, I can beat the algorithm, and I just started learning the game a couple days ago. It is not horrible, but it is not at all good. It is good at avoiding obvious traps, but it does no planning whatsoever, which is key to the game, especially in the beginning stages. I do not think that there can be a computer programmed on computers hitherto made to have anything approximating what we would call "intelligence." Computers, in the current context, require problems that have a rather determinable way of solving them. We, as humans, do not think deterministically. To go outside of the bounds of determinism would make solutions unreliable (since the solution would be based upon random chance) or non-existant (the algorithm would put the computer into a never-ending computation known as an infinite loop.) With the advent of quantum computers (which are expected in fifty to 100 years), that may change, and we may actually see something that resembles "intelligence," but I am skeptical since programming "intelligence" would require that we humans, as the creators, would need to fully understand "intelligence," and rarely agree upon what defines intelligence.

DISCUSS!

They are all dolls. That what it looks like to me.

A president of a high-tech start-up company once told me that he thought that people don't use logic much except as a tool for communications. You figure out how things work by whatever process is natural to you, then, as an entirely separate exercise, you restate everything in a way that "makes sense." But it's all lily-gilding. The real thinking has already been done, using different means. I think there's a lot of truth in this.

Rene Horn is right. We see and hear all this hoopla about how computers have beaten grand chess masters (Kasparov, etc.), but chess is way overated as a game of strategy and tactics. There are lots of possible moves, but the limit isn't as high as one might suspect.

On the other hand, Go has been taken and represented in multiple facets of life, from the concept of the stones and board representing earth and sky to the mathematical implications it carries (watch Pi, that movie shook me senseless for a while).

And although gender isn't touched on as Zola previously mentioned, there is still one distinction in the game (everything else being equal, and it is): Black or white, and who goes first (Black, officially). That's the only flaw in an otherwise "perfect" game, and ironically a flaw we had to introduce to balance the game since whoever goes first has an advantage...

Algorithms are both used by human and computer minds to solve problems, but they are extremely time-consuming and can arrive at an infinite loop. However, we prefer to use "shortcuts" in our way of thinking with what we call judgment or common sense. Of course, there's a price to pay, and that's why we get wrong answers in life even if we think they're right.

The field of statistics is a great example where the application of common sense fails in many circumstances, especially against our perceptions of chance and probability. Most computers work by algorithms only (if there are exceptions, I would like to hear about them ), so for us to perceive "their" minds as similar to ours is pretty far-fetched. But the SF genre would suck if we could not pretend such an idea, but we can't mistake that with our current reality.

So yeah, the day plausible A.I. arrives will be the same day the "perfect" game of Go is played, whether by a human or computer.

hmm well


then...

how would an Andriod reproduce? Could it? would it just pass on its uh memory or program to another body?

Oh man here we go again.

1. The argument that AI is impossible is demonstrably false. The human brain thinks, although not well, and the human brain was designed by the trail and error of evolution. There is no magic inside the human brain that makes thinking impossible to duplicate.

2. Computers that work by learning are called neural nets. Even these computers use algorithms, although the operation of these algorithms is too complicated for us to completely comprehend. This is true of human thinking as well. Every problem solving method is an algorthim at some level.

Based on 1 and 2, and knowing how technology has progressed in the past, it is likely that a true thinking machine, by virtue of conscious design instead of chance evolution, will actually think far better and faster than us natural thinkers. In the same manner that consciously designed jet planes fly far better and faster than birds, which were also designed by the blind chance of evolution.

As to when such a breakthrough arrives, we can only guess. It may take 100 years, or it may have already happened. If something out on the Internet was already thinking and self-aware, we couldn't detect it. (Remember, if it exists, it is smarter than you.)

quote:

Originally posted by C.R Foxhound how would an Andriod reproduce? Could it? would it just pass on its uh memory or program to another body?

Theoretically, an android wouldn't have to die because presumably one would just keep on upgrading and replacing parts as necessary. Would an android want to be able to die, I wonder?

Would an android want to reproduce? If s/he did, perhaps s/he could program his or her own characteristics into another android. Would s/he be interested in raising a human child? Its hard to say.

heh, i could only imagine if they pass court orders in the future allowing robotic couples to raise human children...

didnt somebody write a book on that?

quote:

Originally posted by Zola quote: Originally posted by C.R Foxhound how would an Andriod reproduce? Could it? would it just pass on its uh memory or program to another body? Theoretically, an android wouldn't have to die because presumably one would just keep on upgrading and replacing parts as necessary. Would an android want to be able to die, I wonder? Would an android want to reproduce? If s/he did, perhaps s/he could program his or her own characteristics into another android. Would s/he be interested in raising a human child? Its hard to say.

An android can reproduce any way you can imagine. (And maybe some ways that you cant ) Presumably the android has more options than a human being, because its mind was manufactured and so can be duplicated. We still don't know how to duplicate a human mind. (This last may not be true forever)

Zola, about android immortality through part replacement. The same thing could be said about humans. The key is when/whether *all* android/human parts are replaceable. Whether a human or android chooses to die simply depends on how happy they are.

Androids wishing to reproduce? In a loving relationship. it is quite possible. I believe it will be safest for humanity to make sure that all artificial intelligences love humanity. By love I mean the highest definition, where the loved one's happiness is essential to your own. Under this condition, an android and a human might very well choose to have offspring. The mechanics of it merely require the right technologies.

Wow thats very intresting.Good point.

Fifty years before we see workable quantum computers seems really pessimistic. Progress is already being made:

http://www.research.ibm.com/resources/ne...9_quantum.shtml

And in computer terms, fifty years is ETERNITY. Think of the state of computing in 1953 compared to now.

quote:

Originally posted by Zola Theoretically, an android wouldn't have to die because presumably one would just keep on upgrading and replacing parts as necessary. Would an android want to be able to die, I wonder? Would an android want to reproduce? If s/he did, perhaps s/he could program his or her own characteristics into another android. Would s/he be interested in raising a human child? Its hard to say.

Humans pride themselves in their brains but make most of their decisions with their glands. Thus, when discussing robots, people talk about artificial brains but forget about the equally important artificial adrenaline and testosterone.

Get the hormones right, and the desired behavior will follow!

quote:

Originally posted by A Clockwork Tomato Humans pride themselves in their brains but make most of their decisions with their glands. Thus, when discussing robots, people talk about artificial brains but forget about the equally important artificial adrenaline and testosterone. Get the hormones right, and the desired behavior will follow!

EXACTLY! If you gave an android gender and a feedback system with a couple of built in tendencies, you woul probably get something very similar to a human infant.