121 réponses à ce sujet

[Andrew_Waltfeld]

Collider wrote...

Individuality as we see it is what I meant. They are already have individual collectives, though IIRC only two. The impression I get from Legion is that they like sharing programs and don't understand or want to be individual.

As for asking if this unit has a soul, it's important to know that a "unit" is not one program. It's a multitude of programs. That's really all the Geth are, hardware and programs. So it's like asking if "this group of programs has a soul."

Actually according to legion, each program is an individual with it's own views and perspective which they all share with each other. Then they form an Concensus. otherwise during the legion loyality mission, you can not have an concensus if you consider each geth unit to be an collection of programs. Each program has an vote according to legion.

If you want to understand my theory on it, please see page 1, about 1/2-3/4 of the page down of this topic.

[Srau]

You are taking the science only road to try to explain life, i think by doing so you are narrowing the possible outcomes of your reflection.
I am absolutly not saying you should use theology but that you could at least consider a phylosophical point of view.
I sincerly hope we never ever creat an AI (a real sentient one i mean) or stumble on one because we would alienate it thinking it is nothing more than a toaster or worse, it could ironicaly think that life is only mechanical and try to clean up, ala reaper, failed bactrium evolution.

Modifié par Srau, 12 avril 2010 - 08:23 .

[abstractwhiz]

Vaenier wrote...

They are Made Out of Meat ^.=.^

Whoa, they made a video out of one of my favorite short stories ever! Nice!

[Nightwriter]

Srau wrote...

You are taking the science only road to try to explain life, i think by doing so you are narrowing the possible outcomes of your reflection.
I am absolutly not saying you should use theology but that you could at least consider a phylosophical point of view.
I sincerly hope we never ever creat an AI (a real sentient one i mean) or stumble on one because we would alienate it thinking it is nothing more than a toaster or worse, it could ironicaly think that life is only mechanical and try to clean up, ala reaper, failed bactrium evolution.

You're right, most of us are only thinking of this from a scientific perspective.

However, since I believe life is more than just the scientific, perhaps we should try looking at it another way.

[VettoRyouzou]

>..>

<..<



JOHNNY 5 ALIVE!



sorry had to do it go on with your debt.

[abstractwhiz]

I was once at a Marvin Minsky lecture on AI, and someone asked whether a true AI (like the Geth) could be considered alive. Minsky pointed out that nobody has any idea what the word 'life' really means. Biology studies instances of life, but if you make a generalization that covers all of them, which features are essential, and which are merely properties of the instance?

Take metabolism - all life that we know has it, but for all you know it's a quirk of Terran biology, and doesn't generalize. Reproduction seems reasonable (since every living system tries to make more of itself), but what if the organism doesn't die? Death might be irrelevant too - it's just something that happens here but may not be a necessary feature of living systems. Essentially, biology is experimental, in that it studies stuff that already exists. Making up a definition of life serves no purpose other than giving students something useless to memorize for an exam. <_<

There is an old idea called vitalism, which was what people believed for most of history. The idea was that there is some force (usually called the elan vital) that infuses matter with life. So we have it, but rocks don't. Unfortunately, there is no essential difference between the atoms of a rock and those of an animal. Living matter is the same as non-living matter - it's the configuration that makes the difference. Vitalism was pretty much tossed out of scientific discussions when we figured out that it wasn't really an explanation, just a curiosity-stopper.

The suggestion Minsky ultimately made was that we should stop using meaningless terms like 'life' (except maybe for literary or metaphorical purposes), which seems eminently reasonable to me, in a technical context. 

Modifié par abstractwhiz, 12 avril 2010 - 10:36 .

[Speakeasy13]

abstractwhiz wrote...

I was once at a Marvin Minsky lecture on AI, and someone asked whether a true AI (like the Geth) could be considered alive. Minsky pointed out that nobody has any idea what the word 'life' really means. Biology studies instances of life, but if you make a generalization that covers all of them, which features are essential, and which are merely properties of the instance?

Take metabolism - all life that we know has it, but for all you know it's a quirk of Terran biology, and doesn't generalize. Reproduction seems reasonable (since every living system tries to make more of itself), but what if the organism doesn't die? Death might be irrelevant too - it's just something that happens here but may not be a necessary feature of living systems. Essentially, biology is experimental, in that it studies stuff that already exists. Making up a definition of life serves no purpose other than giving students something useless to memorize for an exam. <_<

There is an old idea called vitalism, which was what people believed for most of history. The idea was that there is some force (usually called the elan vital) that infuses matter with life. So we have it, but rocks don't. Unfortunately, there is no essential difference between the atoms of a rock and those of an animal. Living matter is the same as non-living matter - it's the configuration that makes the difference. Vitalism was pretty much tossed out of scientific discussions when we figured out that it wasn't really an explanation, just a curiosity-stopper.

The suggestion Minsky ultimately made was that we should stop using meaningless terms like 'life' (except maybe for literary or metaphorical purposes), which seems eminently reasonable to me, in a technical context. 

I think that's a very valid point. Let's then shift the topic to "are the Geth sapient"? For instance, how many Geths do you think equals one human being? Mind you individually each Geth isn't smarter than dogs. Than how many of them would you save over a human being? If you must choose between, say, 2 people and 1500 Geths, what would be your choice?

I think essentially that question is equal to "how sophisticated does an AI have to be to considered sapient". Where do you draw the line?

[Xandurpein]

As with many other philosphical arguments it boils down to a definition of words as much as anything else. It's impossible to agree until you have agreed upon the definition of what 'life' is.

[abstractwhiz]

Shandepared wrote...

I suggest everybody in this thread look up the "Chinese Box" in relation to questions concerning artificial intelligence.

I think you're probably referring to the famous Chinese Room argument. Here's a nice place to read up on it, along with the various rebuttals that have sprung up since it was published.

Also, an extremely relevant idea is the Society of Mind. This is a model of human intelligence, and it relates spectacularly well to the geth. Post on that coming up. 

[abstractwhiz]

Some thoughts on stuff I've seen in the previous posts.

There's a lot being made of the fact that geth can download from one platform to another. I don't quite see the relevance of this. They have a very loose mind-body coupling, while organic species (at least in ME) have a very tightly coupled mind-body system. You can do this with an organic species too, with sufficient imagination. Something like the Trill from Star Trek. You can imagine them finding a host body - a nonsapient animal, or something manufactured using advanced biotech - and then inserting themselves into it. 

In fact, this could probably be done with humans, with sufficiently advanced technology. You might surgically remove the brain and whatever parts of the nervous system are necessary to preserve personality. Then put this in a specially prepared body, or a suit of armor, or what have you. It'll probably feel very strange at first, but you might be able to adjust. Of course this is just speculation at this point. 

I'll go one step further. I think the following is probably true - if you could make a sufficiently exact duplicate of the stuff that you were moving, you could just build that and put it in the target body. And now there will be two of you, ready to give nightmares to the legal community.

And now to make it even worse, don't put it in a body. Make a software model of the body and implant it in that. Now you have a digital version of yourself. 

The point here is that all versions are equally you. The essential part of your self is the information that describes it, and once you have that information, your self can be duplicated as easily as an MP3 or a game. (Yeah, I know I'm invoking metaphysical naturalism here, but I see no compelling reasons to believe otherwise.)

The geth are doing essentially the same thing when they make backups and download or duplicate programs. The only difference is that their form is much better suited to it, because of their loose mind-body coupling. The human form isn't, but it's only a matter of degree. 

Lastly, my favorite way to view the geth is as pieces of a single person. The geth are really one ENORMOUS mind shattered into pieces. If a human mind were shattered into its component processes (I'm assuming the Society of Mind model here), you could relate each one of those to individual geth programs - relatively stupid on their own, but intelligent in combination. Note that I'm talking about the mental processes themselves, not their physical realization. It also works well with what Legion tells you about the megastructure the geth are building - they're essentially trying to put the pieces back together. And given that this is one mind whose lower-level processes have human level intelligence (like Legion), it's going to be insanely smart. 

[Speakeasy13]

Shandepared wrote...

abstractwhiz wrote...



I think you're probably referring to the famous Chinese Room argument. Here's a nice place to read up on it, along with the various rebuttals that have sprung up since it was published.

Yes, that is what I meant. I find it very thought provoking. I'll remain skeptical of artificial intelligence until such a thing is actually created, assuming it is even possible.

A Chinese Box is a series of identical box of different sizes that contain one another. It's a completely different entity to the Chinese Room.

That said, I really don't think that's the center of the issue here, considering AIs have already been created in the fictional universe of Mass Effect. I wouldn't be debating whether the Geth are sapient beings of their own, rather than whether the Geth CAN be created, yet here they are.

It's something to think about, yet it means nothing in the context of this thread.

[abstractwhiz]

Shandepared wrote...

abstractwhiz wrote...



I think you're probably referring to the famous Chinese Room argument. Here's a nice place to read up on it, along with the various rebuttals that have sprung up since it was published.

Yes, that is what I meant. I find it very thought provoking. I'll remain skeptical of artificial intelligence until such a thing is actually created, assuming it is even possible.

I think it's eminently possible in principle, simply because we exist. In the worst case, you just need to simulate a human mind (or maybe all the way down to a full-fledged body, if necessary) and you'll have an 'artificial' intelligence. The word 'simulate' really doesn't do it justice. It's like saying that a digital synthesizer only simulates real musical instruments, but of course that makes no difference to the music. Music played on software instruments is still music.  

Modifié par abstractwhiz, 12 avril 2010 - 11:20 .

[Nightwriter]

The Turing Test leaves me dubious. I assume the test involves the computer talking to a human online without the human knowing it's a computer at all.

I'd have to see successful examples of this test with my eyes - I can't imagine talking to a computer in online chat and it being able to pass as a human being to me. The responses I would expect a human being to provide seem too random, complex, wide-ranging and individualized for a computer to handle.

[abstractwhiz]

Nightwriter wrote...

The Turing Test leaves me dubious. I assume the test involves the computer talking to a human online without the human knowing it's a computer at all.

I'd have to see successful examples of this test with my eyes - I can't imagine talking to a computer in online chat and it being able to pass as a human being to me. The responses I would expect a human being to provide seem too random, complex, wide-ranging and individualized for a computer to handle.

I think we all know a few people who would probably fail the Turing Test. <_<

I understand the usual problem is actually quite the opposite. People fall for all sorts of minor tricks, or bend over backwards to be 'fair'. 

[Nightwriter]

abstractwhiz wrote...

Nightwriter wrote...

The Turing Test leaves me dubious. I assume the test involves the computer talking to a human online without the human knowing it's a computer at all.

I'd have to see successful examples of this test with my eyes - I can't imagine talking to a computer in online chat and it being able to pass as a human being to me. The responses I would expect a human being to provide seem too random, complex, wide-ranging and individualized for a computer to handle.

I think we all know a few people who would probably fail the Turing Test. <_<

I understand the usual problem is actually quite the opposite. People fall for all sorts of minor tricks, or bend over backwards to be 'fair'. 

What exactly do you mean by bending over backward to be fair? Out of curiosity.

[abstractwhiz]

Nightwriter wrote...

abstractwhiz wrote...

Nightwriter wrote...

The Turing Test leaves me dubious. I assume the test involves the computer talking to a human online without the human knowing it's a computer at all.

I'd have to see successful examples of this test with my eyes - I can't imagine talking to a computer in online chat and it being able to pass as a human being to me. The responses I would expect a human being to provide seem too random, complex, wide-ranging and individualized for a computer to handle.

I think we all know a few people who would probably fail the Turing Test. <_<

I understand the usual problem is actually quite the opposite. People fall for all sorts of minor tricks, or bend over backwards to be 'fair'. 

What exactly do you mean by bending over backward to be fair? Out of curiosity.

I mean that they unconsciously reject certain lines of questioning as being too hard, or something analogous to that. But I should have mentioned that this happens in one of those big Turing test events they have every year (the Loebner Prize), and it's mostly just people writing chatbots. I read about this recently, but I can't find the link.

Modifié par abstractwhiz, 12 avril 2010 - 11:44 .

[Nightwriter]

abstractwhiz wrote...

I mean that they unconsciously reject certain lines of questioning as being too hard, or something analogous to that. But I should have mentioned that this happens in one of those big Turing test events they have every year (the Loebner Prize), and it's mostly just people writing chatbots. I read about this recently, but I can't find the link.

Ha! That's cheating!

"Questions that are too hard..."

Doesn't that totally ruin the test? Either it can pass for a human and answer the questions that a human can, or it can't. I mean, unless the question is, "Explain the cultural human perspective and philosophical meaning behind Ayn Rand's Atlas Shrugged  and describe the emotional impact it had upon you as an individual" or something.

Modifié par Nightwriter, 12 avril 2010 - 11:54 .

[Xandurpein]

abstractwhiz wrote...

Nightwriter wrote...

The Turing Test leaves me dubious. I assume the test involves the computer talking to a human online without the human knowing it's a computer at all.

I'd have to see successful examples of this test with my eyes - I can't imagine talking to a computer in online chat and it being able to pass as a human being to me. The responses I would expect a human being to provide seem too random, complex, wide-ranging and individualized for a computer to handle.

I think we all know a few people who would probably fail the Turing Test. <_<

I understand the usual problem is actually quite the opposite. People fall for all sorts of minor tricks, or bend over backwards to be 'fair'. 

It's like when you ask one group of humans to produce a sequence 0 and 1 by pure chance/coin flipping and another group to write down a series of 0 and 1 trying to make it seem random. You can almost always spot the the real random because they have longer strings of just 0000 or 1111. Humans try to be fair and even things out more than pure chance.

[abstractwhiz]

I think the reason it happens is because everyone wants *some* progress, and there's this fear that if you hold it up to overly exacting standards, the whole contest will be an epic failure. Better to go easy on them and scale it up as time goes on, or something like that. Keep in mind that most AI research is like this anyway - they pick a small subproblem and try to solve it. Sadly the problem is so damn hard that even the subproblems tend to leave us reeling.

[HeyBlade789]

The geth are, near true life, they dont experience emotion, but they can think for yhemselves, so i believe them to be more alive than machine

[Inverness Moon]

abstractwhiz wrote...

The suggestion Minsky ultimately made was that we should stop using meaningless terms like 'life' (except maybe for literary or metaphorical purposes), which seems eminently reasonable to me, in a technical context. 

I agree. I've mentioned before in this thread that I believe that there are multiple definitions to life, and attempting to apply organic definitions and/or philosophical definitions is a waste of time.