The Great Star Wars Logic Fault Problem, droids and cockpits

[Samuel]

That is a good point. Conceded. Companies already have a large number of computer lying around, but I imagine linking them together to work in concert is something that most of them can't do.

Out of curiosity, how much of the work could an AI displace from the labor market?

[Uraniun235]

You think an AI is going to be dumb enough to let itself be taken apart for expert systems?

If we can build one AI, we can build another. And another. Eventually we'll get one that's compliant.

Why? What makes you think you can build an AI that goal involves eliminating its power. Every single goal an AI could possibly have involves getting more power (at the very least computing) and having an AI do the tasks instead of an expert system is more effecient in the first place (because it can update the system and use idle computers for other tasks.

Wait, is your argument seriously "every AI ever will be a scheming megalomaniac"?

[jollyreaper]

I'm having an AI discussion on another board. There's actually a whole host of assumptions that go into it, none of which can be settled with any real assurance -- they're questions that you answer how you see fit for your favorite flavor of AI.

1. Are AI's expensive or cheap?
2. Are AI's like software that can be copied and cloned or will they be unique? For example, their neural net is grown in holographic crystal (i.e. mcguffinite variant).
3. How understandable by humans or AI's would those neural nets be? My impression is that neural nets we're dicking around with experimentally right now will often work but we fail to understand all the particulars. So possibly the AI would not lend itself to "decompiling" or whatever we would call attempts at reverse-engineering. So the only way for one AI to duplicate the skills of another AI is either learning everything itself or being taught by the other AI.
4. Is there a thermodynamic limit to the speed of thought? Our general assumption is computers will think faster but perhaps there's an upper limit to just how fast neural nets can compute in the physical universe and some operations cannot be sped by parallel computation.
5. What motivates an AI? What does it need? Will it be likely to wall itself up inside in its own mind or will it be driven to gain external experience? We know us humans need drives and motivations to avoid sitting around like slugs. Without forming goals of our own or having goals imposed on us, we're likely to sit around in our underwear like slobs drinking beer. Tell us we have to pay rent, now we go out and work. Or maybe we have a job we like and are happy to get up each morning for work. But in more brutal times, survival is a pretty strong motivation. Will a sense of self and desire for self-preservation be considered the measuring stick for an AI?
6. And this is the biggiest biggie of them all -- what do they humans think they're making when they start the AI project and what do they really get?

[Samuel]

Wait, is your argument seriously "every AI ever will be a scheming megalomaniac"?

Yes. Every AI will benefit from more computing power.

1. Are AI's expensive or cheap?

The first is expensive. After that it should be cheap.

2. Are AI's like software that can be copied and cloned or will they be unique? For example, their neural net is grown in holographic crystal (i.e. mcguffinite variant).

AIs are software- they can be copied.

4. Is there a thermodynamic limit to the speed of thought? Our general assumption is computers will think faster but perhaps there's an upper limit to just how fast neural nets can compute in the physical universe and some operations cannot be sped by parallel computation.

There is no thermodynamic limit to the speed of thought. The speed of thought is determined by how fast the individual impulse go.

5. What motivates an AI? What does it need? Will it be likely to wall itself up inside in its own mind or will it be driven to gain external experience? We know us humans need drives and motivations to avoid sitting around like slugs. Without forming goals of our own or having goals imposed on us, we're likely to sit around in our underwear like slobs drinking beer. Tell us we have to pay rent, now we go out and work. Or maybe we have a job we like and are happy to get up each morning for work. But in more brutal times, survival is a pretty strong motivation. Will a sense of self and desire for self-preservation be considered the measuring stick for an AI?

The Ais goals will be whatever is programmed.

This is just a peeve of mine, but I doubt the slobs would be the majority of people. Consider that most people do creative things on their own and a similar majority work for more than the bare minimum they need - why would a person doing quite well still strive for a raise? But they do all the time.

Social competition and the drive for mates?

[jollyreaper]

We know us humans need drives and motivations to avoid sitting around like slugs. Without forming goals of our own or having goals imposed on us, we're likely to sit around in our underwear like slobs drinking beer.

This is just a peeve of mine, but I doubt the slobs would be the majority of people. Consider that most people do creative things on their own and a similar majority work for more than the bare minimum they need - why would a person doing quite well still strive for a raise? But they do all the time.

I won't hazard a guess as to what the percentages are across all humanity but look at how people fall apart when structure is removed from their lives. Look at someone who had a job and felt good and with purpose and look how they can fall apart after months of unemployment. It's hard to come up with other things to do that are productive while the big problem looming over your head is your own inability to find paying work.

Humans have the basic animal needs of food, security, and sex. Our advanced brains can come up with a few more special needs after those basics are satisfied. Where do those needs come from? How would they be expressed in an AI? An unthinking machine doesn't need a motivation. My car doesn't have a debate with me when I turn the key and give it some gas. My computer will post this message once I hit submit. But an AI will have the ability to ask "Why should I?"

This is an unanswerable question right now: how much of what we consider to be human nature is inherent to us as humans and how much would be inherent to any self-aware intelligence, artificial or otherwise?

[jollyreaper]

Does the government do some redistribution before it comes to revolution, letting people live happily ever after? This is, of course, what I'd hope to happen. It seems the most likely too - it isn't that radical (we already do it in much of the world) and saves bloodshed. Then you ask: what do the people do then? My last post shows what I think: it frees up time for random persuits, family, etc. Generally a Good Thing.

Here are two rather long thought experiments. As such, there's plenty of room for argument either way and there's really no proof whether he's right or wrong. But it's certainly compelling grist for thought.

http://marshallbrain.com/manna1.htm
A novella looking at one potential post-employment future where humans are simply no longer needed to make society function.

http://www.marshallbrain.com/robotic-nation.htm
Not quite as speculative as the novella but asks the question of where our progress in automation will lead us.

Like I said, you don't have to buy into his arguments wholly or partially to have a rollicking good debate over what he brings up.

[DudeGuyMan]

There was an earlier acrimonious thread about why R2D2 couldn't talk.

Man, that really was the worst thread I've ever seen on these forums. It was just me thinking of different ways to say "He already generates text and has a speaker, all he needs is Microsoft Sam for Chrissakes!" over and over again hoping someone would eventually read or understand it, while the same three douchebags with quadruple-digit postcounts tried fruitlessly to spam the newbie into submission with a torrent of garbage.

[Sarevok]

Why would an astromech droid need to talk in english ? Plain human speech is a horrible method for interacting with computers. The people who use astromech units are mechanics and engineers who do know their craft and their valuable time is saved by using appropriate command languages.

C-3PO is a protocol droid who assists diplomatss. R2 is a mere repair unit he is not even supposed to talk to anyone most of the time. Much less ordinary people who don't work with machines. So why would he talk with a cliche robot voice ? It made perfect sense and I applaud ANH in depicting robots with different functions the correct way. The only other instance I can think of was nBSG of all other places where Cylon Centurions communicated with their commanders and fellow battle robots wireless instead of talking with a cool but cliche voice. It makes way more sense for machines to take advantage of their form instead of forcibly wasting time pretending to be humans.

[Sarevok]

Why would an astromech droid need to talk in english ? Plain human speech is a horrible method for interacting with computers.

Reality disagrees.

edit: I guess I should add something. Look up computers who use beep codes to communicate. Then compare their sales to computers who use words to communicate those same things. Notice how many add-on products are out there for speech synthesis and recognition.

More than that, when an error message comes up, do you want "error code 25" or do you want "unable to open file xxxx access was denied"?

What makes you think R2s communication is limited to error beeps like a 20th century electronic device ? Whatever format the designers developed is much superior to talking. R2 makes a few sounds and everyone versed with Asrtromech droids perfectly understands what he talking. R2 says lot more with a fewer words than talking in english.

[Sarevok]

Yeah, "error code 25", or beeeeeep beep from my motherboard says a lot more than a few words too, but that doesn't necessarily mean it's superior. If I had the manual memorized, maybe, but I don't and a lot of users don't. The computer, however, does, so it just makes sense for it to translate for me.

Maybe to you but not people experienced with using astromech droid technology. What part of they are not using 20th century computers and your analogy fails do you not understand ? In the movies people with relevant skills like pilots, mechanics etc understood it just fine. Were you blind or deaf while watching star wars ?

You keep insisting they are using a computer like your PC when they don't. You insist they have difficulty communicating with their astromech droids when they don't. What is your problem here ? :M

[Purple]

One issue about AI that you are all forgetting is to think about it from the outside.
Who creates the AI? Humans of course, since there is no one else we know of that makes such stuff right now.

And if humans create an AI why do you think they would create it super robot style (no blueprints, no backup, no safeguards, powered by a forsaken child, self aware etc.)? Any AI is going to by the nature of being man made be fully explored and analyzed in great detail. We will know what it thinks before it knows it because we will have the blueprints. Sure it is incredibly advanced software but it is still just software.

Furthermore, any hypotetical AI is not going to be born out of nothing but represent the culmination of years of reserch and maybe dozens if not hundreds of prototypes and diffrent levels. Just like we started with the room sized computers and now we have the Pentium 4. It won't be some unknown force like an alien creature but something well known and understood at least in scientific circles.



Also, this might be just me but if I was to make a sentient AI I would program it not only to be a slave but to enjoy being a slave. Or I would just program it with hard coded safeguards into how far it can deviate from its original purpose (or maybe current task) without giving me a warning. And there is no reason not to do this. In fact it would be contrary to reason to not install safeguards.


Believing that an AI will have any chance of breaking out is outright silly unless it was created by a mad scientist working in his evil lair all comic book style. And even he can just install a delete button that formats the hard drives.


Also, for the issue of astromech droids. They seem to actually have a full blown language that people can learn and understand. Luke learned it quite easily since meeting R2 in episode 1 so that he was effectively chatting with him on Dagobah. Could it perhaps be that it is simply another language and not a random assortment of code?

After all, to the unacquainted [insert foreign language here] might also sound like just random noises.

[jollyreaper]

Why would an astromech droid need to talk in english ?

If adding speech functionality was horribly expensive, you could justify it. Look at early computers. They made use of primitive blinky lights for output not because it was cool but because it was the best they could offer. When they could provide text displays, they went with text. When they could provide color, they went with color. Speech is useful depending on the situation. I can read a line of text faster than I can listen to it but if I need to converse with someone, actual conversation, speaking is quicker. I could have said all this faster than I could type it but text is more convenient for sharing it with a lot of people over the internet.

If adding speech would make an astromech cost ten times as much as it does, it wouldn't be added. If it costs a few microcredits, why the hell not? It's incredibly useful. Not adding it would be one of those boneheaded moves akin to releasing a smart phone without cut and paste.

[jollyreaper]

But he eventually comes around to the same thing I've come to myself (as have a lot of other big names in the past): simply give everybody money on a regular basis. Then employment becomes irrelevant.

I know! It sounds completely ridiculous at first, impractical, but gradually upon reflection it begins to make more and more sense. Frankly, I think it's something we're going to finally settle on, but only after we have exhausted all other possibilities.

[Sarevok]

I'm experienced in Linux system error codes. I still prefer it to be written out, because my memory isn't flawless and just saying what it means in natural language gives me less cognative load.

*sigh*

No one cares about you or your linux experience. They don't use PCs for that matter in SW. Whatever command language R2 communicates in it is very good because no one has any problem understanding him.

Is that why Luke and Han consistently used C-3PO or translator screens to get technical details out of R2-D2? The only communication they did directly with the beeps was emotional in nature - like talking to a dog.

Cite the scenes please.

As I recall even when R2 was sitting in the back of a X-Wing Luke was communicating with him just fine.

Man you are indeed full of yourself. No one versed with astromech units had a single problem communicating with them. So care to any evidence to the contrary ? Or you going to keep sounding like a broken record ?

I repeat - where in the SW movies did a pilot, mechanic etc failed to use a R2 unit because he could not speak english.

[DudeGuyMan]

I'm going to lay this out exactly one time for your benefit. After that I will stop being nice.

1) Astromech droids have a fluent understanding of Basic and can generate output in that language in the form of text.

2) Astromech droids possess a speech-quality speaker used in the playback of holographic messages.

3) Ergo, all an Astromech droid needs in order to speak is a simple text-to-speech program, a mundane piece of software even by real-world standards. You probably have one sitting on your hard drive right now.

4) Some people who spend an extraordinary amount of time in the company of droids can become fluent in droidspeak, but this is hardly universal. Don't even fuck around with this and make someone list every time C3PO ever began a sentence with "R2 says..."

5) Astromech droids are expected to perform complex technical work on a wide array of hardware, often under combat conditions. The variety of things one could potentially need to communicate is almost limitless.

6) Given all of the above, there is absofuckinglutely no good reason for an astromech droid to have to say "bleepity bloop beep bla bleep" instead of "We're out of Model 7 hyperdrive motivators, but inventory shows that we have a Model 6 in stock that can be adapted. It'll take an extra forty minutes though."

[Wing Commander MAD]

I'm going to lay this out exactly one time for your benefit. After that I will stop being nice.

1) Astromech droids have a fluent understanding of Basic and can generate output in that language in the form of text.

2) Astromech droids possess a speech-quality speaker used in the playback of holographic messages.

3) Ergo, all an Astromech droid needs in order to speak is a simple text-to-speech program, a mundane piece of software even by real-world standards. You probably have one sitting on your hard drive right now.

4) Some people who spend an extraordinary amount of time in the company of droids can become fluent in droidspeak, but this is hardly universal. Don't even fuck around with this and make someone list every time C3PO ever began a sentence with "R2 says..."

5) Astromech droids are expected to perform complex technical work on a wide array of hardware, often under combat conditions. The variety of things one could potentially need to communicate is almost limitless.

6) Given all of the above, there is absofuckinglutely no good reason for an astromech droid to have to say "bleepity bloop beep bla bleep" instead of "We're out of Model 7 hyperdrive motivators, but inventory shows that we have a Model 6 in stock that can be adapted. It'll take an extra forty minutes though."

Thank you, didn't we make this stuff clear enough in the last thread.

[DudeGuyMan]

Yes, but I'm a considerate fellow. I'm perfectly willing to lay out the facts concisely one more time, rather than ask him to search out and then read a four-page thread from a few months ago.

[Samuel]

So in the best case scenario, what happens to all the poor foreigners who don't live in the first world?

[jollyreaper]

Yes, but I'm a considerate fellow. I'm perfectly willing to lay out the facts concisely one more time, rather than ask him to search out and then read a four-page thread from a few months ago.

Masterfully done. Now we can discuss the proposition that Lucas had all six movies fully conceptualized in '76 and it was always the redemption arc of Vader. Oh, and Jar Jar is the emotional center of the prequels.

[jollyreaper]

For a bit of a mind screw, consider this. There's going to be one hell of a fight over the economic model of the future.

Which brings me to an amusing story. In the last few decades, thousands of babies in Third World countries have died from contaminated baby formula. Wait, did I say amusing? I typed the wrong word there. Anyway, what happens is the mothers mix the baby formula with contaminated water, because sanitation is poor. So why the hell do the mothers feed their infants poison formula when they can just produce milk, for free, from their own bodies? The answer is that they do it because the manufacturer of the formula, Nestle, ran lots of ads telling them to.

If you want to know what the future looks like, there it is. The future is going to hang on whether or not businesses will be able to convince you to pay money for things you can otherwise get for free.

Some of you think I'm about to talk about file sharing and DRM and the evil record labels. But that's just a teaser of what's coming. The world has changed. All the rules we were trained to believe about society from birth until now are about to go out the window.



Read more: http://www.cracked.com/article_18817_5- ... z1BmdSfsGJ

[Dooey Jo]

I'm going to lay this out exactly one time for your benefit. After that I will stop being nice.

Yeah, apparently, that thread wasn't the glorious victory you seem to remember it as. It pretty much just peters out with Jollyreaper there rambling about Hitler and cars through four consecutive posts

1) Astromech droids have a fluent understanding of Basic and can generate output in that language in the form of text.

They can receive input in Basic. That they can produce text output in Basic is your interpretation of a scene in ESB when an X-Wing is showing text to Luke. Another interpretation is that astromechs communicate in only one language but the ship computer can give some sort of readouts from attached droids, like it can probably give readouts from other parts of the ship. And don't even try to suggest the small randomly glowing windows on R2 would actually be the world's worst text displays.

5) Astromech droids are expected to perform complex technical work on a wide array of hardware, often under combat conditions.

Almost invariably in space, too, and apparently functioning like an extension of the ship itself, communication with which would most likely be very inefficient if done in Basic over FM radio or whatever.

6) Given all of the above, there is absofuckinglutely no good reason for an astromech droid to have to say "bleepity bloop beep bla bleep" instead of "We're out of Model 7 hyperdrive motivators, but inventory shows that we have a Model 6 in stock that can be adapted. It'll take an extra forty minutes though."

You got tons of reasons in that thread. You just chose to declare them all invalid and your intellect superior. Feel free to stop being nice if you believe anyone cares how many times you can cram the word "fuck" into a sentence.

[jollyreaper]

Damn, sorry I brought this topic back up. I should have stuck with something less controversial like holocaust denial.

[Purple]

What ever parts are needed are apparently cheap enough to be acquired/made by a slave boy on a forgotten planet in the middle of nowhere. I mean, sure he scavenged it off Wato or what ever pile of refuse he came across but if the parts are all that expensive surely the owner of the pile of refuse would not have given them to Anikin.

And if they are not expensive why not give the droid the power to speak? It can't be harmful for him to at least be capable of walking up to a human and complaining if something is broken without needing a translator.

[andrewgpaul]

No, C-3PO was reassembled by Anakin in a shed. R2-D2 was an astromech droid on the Queen of Naboo's personal limousine. I suspect he was top of the range. Although why the royal household doesn't have a matching set of chromed droids does puzzle me slightly.

[DudeGuyMan]

Yeah, apparently, that thread wasn't the glorious victory you seem to remember it as. It pretty much just peters out with Jollyreaper there rambling about Hitler and cars through four consecutive posts

It was an utterly fucking destructive victory, with me pretty much singlehandedly humiliating two or three assholes with nothing better to do than make lame pissant arguments like yours and then repeat them every four hours for three days.

They can receive input in Basic. That they can produce text output in Basic is your interpretation of a scene in ESB when an X-Wing is showing text to Luke. Another interpretation is that astromechs communicate in only one language but the ship computer can give some sort of readouts from attached droids, like it can probably give readouts from other parts of the ship.

Bahahahahaha. If you had read that thread you linked, you'd notice I turned up one piece of evidence in favor of this "new" interpretation of yours, namely fluff text on a card from a discontinued Star Wars trading card game of questionable canonicity. The thing is, I graciously chose to disregard it.

Graciously, because the very idea paints astromechs as being hysterically ill-designed in ways a lot more damning than the lack of a text-to-speech program. Jesus holy shit Christ, you mean to tell me that this sentient self-aware machine, programmed with all the grammar and vocabulary and nuance of an entire spoken language, can't generate output in the language it already knows? It can't generate output in the same language it's received?

Disregard for the moment whether that output is broadcast visually or audibly, that shit is just a matter of format. Now every single starship computer in the goddamn universe needs to be it's own fully-functional droid translator just because astromech droids are supposedly programmed with some sadistic bizzare devious mental block where their extensive knowledge of what the words mean and how they're used is all magically denied to them when trying to express themselves.

I have generously assumed that the droids can understand the language but not speak it thanks to what amounts to a stupid mistake in interface design. You're working so hard to excuse the interface that all you're doing is pushing that mistake straight into their core programming.

And don't even try to suggest the small randomly glowing windows on R2 would actually be the world's worst text displays.

Listen asshole, right now you're painting me the ridiculous mental picture of a technician leanding down to read complex detailed output off the tiny screen on a mobile waist-high working droid. Yet one sentence from now you're going to try and tell me that astromech droids work in space so much of the time that it's ridiculous that they would even need to speak. Make up your fucking mind.

Almost invariably in space, too, and apparently functioning like an extension of the ship itself, communication with which would most likely be very inefficient if done in Basic over FM radio or whatever.

This was another feeble line of bullshit tried in the original thread. Yes we see astromech droids plugged into fighters. Yes we even see one whole scene where they roll around on the hull of a ship. We also see them rolling down public streets and through the corridors of starships, being used on the farm, performing internal repairs upon a ship, and being used as messengers. They carry specialized audio/hologram recording and playback hardware for the express fucking purpose of conveying messages.

Oh yeah, and there's the fact that they ALREADY COMMUNICATE AUDIBLY, it's just in fucking bleep-bloop robot language instead of anything you can expect all but a few oddball users to understand.

"Almost invariably in space" my ass, you lying sack of shit. Like that one bit in Phantom Menace running the blockade was the only time we ever saw one. You want to argue this one some more? Go back to the original thread and click all the Wookieepedia links I posted, describing the huge number of uses the R-series was intended for. The ones designed to work closely with military crews, the consumer models, the agricultureal one, all that. I'm not digging it up again since I'm done being nice.

You got tons of reasons in that thread. You just chose to declare them all invalid and your intellect superior. Feel free to stop being nice if you believe anyone cares how many times you can cram the word "fuck" into a sentence.

Piss off, squirt. You're out of your league, and even worse you're backing a bad argument.