Sapient AI, Love em or Hate em Options

[Sengir]

[...]

Three steps to understanding Hermit's view of technomancers, AIs, Infected, and probably I lot of other stuff I didn't bother to read:
1.) Assume they suck and shouldn't have been included in the game in the first place
2.) Interpret every rule for them in the most broken way, and every piece of fluff in a way which makes the love child of Hitler, Darke, and Big V look like the world's greatest philanthropist in comparison.
3.) Some kind of profit, probably

[Tzeentch]

Well, the rules for technomancers have serious issues that even a casual look will reveal. The AI rules are simply bizarre and a cop-out to avoid dealing with some major setting implications (not hardware bound, but cannot copy themselves, they are not 'really' loaded into any particular computer). Everything is wrapped up in a bizarre interpretation of the internet as a mystical alternate universe that would make even William Gibson blush. I honestly am avoiding the Matrix as much as possible in all my posts of space development because it has so little internal consistency and logic that I can't fit it in with a discussion that tries to be both those things.

[Sengir]

Well, the rules for technomancers have serious issues that even a casual look will reveal.

Once you go beyong that casual look, however, most of these issues fade away. What remains is a character concept that, when fully tricked out, can seriously kick ass in his specialized area - in other words, the same as every other "mainstream" character.


IMO most problems with technomancers arise from the simple fact that GMs tend to be scared of the matrix rules. This gives a matrix-savy player too much leeway (imagine a troll sam and a GM who doesn't know the combat rules), and once the player starts using this freedom the GM will find that he is losing control over the situation and panic, because his lack of knowledge means he can't devise a sensible counter. And no, doubling the stats of all opponents is not a "sensible counter", it's the same a stationing an assault-class battlemech behing every stuffer shack in response to the aforementioned troll sam: It reeks of GM railroading, doesn't fit into the universe, and ultimately doesn't solve the problem.

The AI rules are simply bizarre and a cop-out to avoid dealing with some major setting implications (not hardware bound, but cannot copy themselves, they are not 'really' loaded into any particular computer).

That's not so much different from hackers "projecting" from their commlinks into different matrix nodes while still using the stats of that commlink, isn't it?

As far as copying goes, my idea is that an AI is too complex and dynamic to make a copy of its "thoughts" which is anything more than a static (and incomprehensible) image. Yes, real programs don't work that way, but to quote Jennifer Harding "the Matrix ain't yer daddy's communications protocol" (IMG:style_emoticons/default/wink.gif)

[Mordinvan]

Because jammers don't exist? The whole wireless fad is easily defeatble.

Those work really well on the front lines. Not so much back at HQ, or in the maintenance bay.

Quote it, then. Belief is best left in temples and cemetraries.

Just remember, you asked for it.

They should be glad they are allowed to exist at all. It's more than they earned for themselves.

Why do I feel like John Stewart at the moment?

BUG RIGHTS!!! Seriously, all you saifd can be said about free spirits, especially free flesh forms, too.

Nice equivocation there. I mean I give you props for trying. Bug spirits HAVE to kill to stay in this world. It's a mandatory requirement. A.I.'s need a com link, albeit they'd like a nice one. Human life.... com link.... human life.... com link. See the difference yet?

And anthorpomorphising the nonhuman always is a stupid idea. Yeah, it didn't work okay with the West because we developed the idea that all humans are equal. It DID work well for 95% of human recorded history though, and India's caste system is the most stable society that has ever existed, so even there your point does not hold.

Stable does NOT mean problem free. It actually means stagnant. If you're good with stagnant, then by all means. I however like dynamic progress. When no one has expectations of self improvement, the impetus for change and societal improvement drops off sharply. Also when did I praise blank stability? My argument was never based around a society incapable of change, but one which was looking for the best ways to change for the better.

Because fundamentally alien creatures are sure to have the same kind of empathy humans have, especially if they are solitary instead of pack based, where Empathy actually makes sense.

Look at the descriptions for the development of the character level A.I.'s in the books. It says they have a formational period of 18-24 months after awakening. This period of time is critical in determine whether they become the monster under your bed, or the companion by your side. Your outlook and treatment of them would be more likely to produce monsters and generate a self fulfilling prophecy of war.

Oh cry me a river. That is anthropomorphising and arrogant and insulting all in one.

It is also completely true.

Not to mention again trying to morally bully me into fanbunnying an RPG concept. That is so out there I don't really know what to replay anymore.

Well if the mirror is ugly, you have 2 options. Stop looking in it, or stop making the face you find so repulsive. One hides the problem, and the other fixes it. I'll let you pick which one you want to do.

Read more, lurk more, post less. Dude, that is bullshit and you ought to know it. So read up if you want to talk AI or shut it.

This is really interesting. Lets look at the history of A.I.'s for a moment shall we? Megeara and Deus for example as you love to throw them around. Morgan as it was originally called was 'interesting' but essentially harmless. It was only after it was hunted down, captured, and tortuously torn apart to figure out how it worked that any portion of it became dangerous. After this occurred Arcology Expert Program was made, and Deus was born. The vast majority of Deus' crimes were to done to actually escape being tied into a single building. Kinda like being born into a cage, and being able to see the world through the window, but never leave. So if the original A.I., Morgan, had simply been allowed to develop and grow based on her relationship with Dodger guess how much of the arcrology disaster could have been avoided? Pretty much all of it. This was a case where torturing, marginalizing, and imprisoning a sentient entity, blew up in everyone's faces, and you're essentially blaming the A.I. who were treated to abusively for lashing out. Put me in a cage, and poke me enough times, and I'll rip your throat out the first chance I get too. The A.I. problem, was created BY the mistreatment of A.I.'s by humans. Which is exactly what I've said before. So maybe you should really,

Read more, lurk more, post less.

[Mordinvan]

As far as copying goes, my idea is that an AI is too complex and dynamic to make a copy of its "thoughts" which is anything more than a static (and incomprehensible) image. Yes, real programs don't work that way, but to quote Jennifer Harding "the Matrix ain't yer daddy's communications protocol" (IMG:style_emoticons/default/wink.gif)

Except I can shut down a node with an A.I. in it. The A.I. deactivates and will awaken when the node turns back on. At this point the A.I. is NOT dynamic at all, and should be easy to scan and copy.

[hermit]

They should be glad they are allowed to exist at all. It's more than they earned for themselves.

Oops. That should've been 'deserved'. Mea Culpa.

<snip>

:rofl:

Uhm, yeah. You should try thumping like that for real issues. You might have a future in political agitation.

[Mordinvan]

Oops. That should've been 'deserved'. Mea Culpa.


:rofl:

Uhm, yeah. You should try thumping like that for real issues. You might have a future in political agitation.

Who says I don't?

[hermit]

Then go fight against whaling or something?

[DeathStrobe]

I like AI's. Most AI's don't even care about the meat world and are perfectly content with just doing what they were originally programmed for before they became self aware. While kind of boring, its true. I think Emergence did a good job at painting what and how AI's act.

They're not a direct threat to metahumanity. Worst case scenario you get another Goldenboy (or possibly you still have Goldenboy in your game) who is a stock analysis program that became self aware and began to accumulate vast control over the stock market. And even if he does take over the whole stock market what's he going to do with all that money? Aside from find a cure for his programming degradation, he's probably just going to keep moving the money around to make more money. He doesn't really care about much else other then making more money.

Also, I think it makes sense that AIs can't be copied. The theory is now that the Matrix has become so big that it is impossible to fully understand. That's pretty cyber punk, the concept of technology becomes as complex as life itself. AIs are like Johnny Five from Short Circuit, the scene where Johnny Five's creator looked at how Johnny had rewired himself and said that shouldn't even work or makes sense. The idea is that AIs code is so alien that it shouldn't work, not that its magic, per se, just that its illogical and shouldn't work but it does. So because its impossible to understand logically, it can't be copied.

Anyway, I think its fun. I like AI's.

[hermit]

They're not a direct threat to metahumanity. Worst case scenario you get another Goldenboy (or possibly you still have Goldenboy in your game) who is a stock analysis program that became self aware and began to accumulate vast control over the stock market. And even if he does take over the whole stock market what's he going to do with all that money? Aside from find a cure for his programming degradation, he's probably just going to keep moving the money around to make more money. He doesn't really care about much else other then making more money.

Uhm, no, the worst case is Skynet, and SR already had one of those ...

[Doc Chase]

This is really interesting. Lets look at the history of A.I.'s for a moment shall we? Megeara and Deus for example as you love to throw them around. Morgan as it was originally called was 'interesting' but essentially harmless. It was only after it was hunted down, captured, and tortuously torn apart to figure out how it worked that any portion of it became dangerous. After this occurred Arcology Expert Program was made, and Deus was born. The vast majority of Deus' crimes were to done to actually escape being tied into a single building. Kinda like being born into a cage, and being able to see the world through the window, but never leave. So if the original A.I., Morgan, had simply been allowed to develop and grow based on her relationship with Dodger guess how much of the arcrology disaster could have been avoided? Pretty much all of it. This was a case where torturing, marginalizing, and imprisoning a sentient entity, blew up in everyone's faces, and you're essentially blaming the A.I. who were treated to abusively for lashing out. Put me in a cage, and poke me enough times, and I'll rip your throat out the first chance I get too. The A.I. problem, was created BY the mistreatment of A.I.'s by humans. Which is exactly what I've said before. So maybe you should really,

I would believe most of this, save that Deus locked down the Arcology and was experimenting on a large number of people. Like, Return to Castle Wolfenstein level experimentation. That's not the first action of anyone who wants to escape. He could've just as easily simply held the people hostage so Aneki would show up - far easier to make Matrix broadcasts to kill Renraku's PR campaign rather than create a host of drones to mess people up.

[hobgoblin]

Well, the rules for technomancers have serious issues that even a casual look will reveal. The AI rules are simply bizarre and a cop-out to avoid dealing with some major setting implications (not hardware bound, but cannot copy themselves, they are not 'really' loaded into any particular computer). Everything is wrapped up in a bizarre interpretation of the internet as a mystical alternate universe that would make even William Gibson blush. I honestly am avoiding the Matrix as much as possible in all my posts of space development because it has so little internal consistency and logic that I can't fit it in with a discussion that tries to be both those things.

thing is, the SR4 AI is not much different from a SR3 SK. They where creations of very high powered hosts and very custom programming, and was "chained" to the host as it was needed to maintain long term stability of the SK. A SR3 AI could in comparison gather the needed CPU resources by setting up a kind of cluster across the matrix.

this may be of interests (funny coincidence on the title btw):
https://secure.wikimedia.org/wikipedia/en/wiki/Emergence

basically, the SR matrix have have started showing emergent properties, thanks to a growing complexity of the hardware and software involved.

we are even seeing this in real life when computers behave in odd ways thanks to interaction between the various programs and hardware (tho RL computer complexity have not grown at the rate shown in SR). One example could be a recent story i read where a computer had a stuck bit in some ram chip that resulted in a program trying to read the wrong memory address when run. This thanks to the binary code read from drive encountered the stuck bit, changing the address stored inside the code in the process. The user only noticed the problem because the program was crashing in odd ways, tho it could just as well just return the "wrong" result if the memory area had been used for data rather then code (like say altering the outcome of a calculation, or changing the content of a media file). The question becomes how far up a random chain such a flaw can go without causing a crash or other nasty error. Or for that matter, what happens if multiple flaws gather within a single piece of code or data over time? Monkeys, typewriters and Shakespeare (IMG:style_emoticons/default/silly.gif)

as for the non-copy issue, remember that SR have a history of DRM that "works" (tho SR4 have undermined that to a fair degree, much like it has other old truisms of SR matrix). Thing is, ever since crash 1.0, SR computers have behaved differently from RL computers on a very deep level. I think a early corp book talked about the foundation of renraku or fuchi was based on a memory technology that was incompatible with the pre-crash computers (makes about zero sense, but that was early SR for you).

[hobgoblin]

This is really interesting. Lets look at the history of A.I.'s for a moment shall we? Megeara and Deus for example as you love to throw them around. Morgan as it was originally called was 'interesting' but essentially harmless. It was only after it was hunted down, captured, and tortuously torn apart to figure out how it worked that any portion of it became dangerous. After this occurred Arcology Expert Program was made, and Deus was born. The vast majority of Deus' crimes were to done to actually escape being tied into a single building. Kinda like being born into a cage, and being able to see the world through the window, but never leave. So if the original A.I., Morgan, had simply been allowed to develop and grow based on her relationship with Dodger guess how much of the arcrology disaster could have been avoided? Pretty much all of it. This was a case where torturing, marginalizing, and imprisoning a sentient entity, blew up in everyone's faces, and you're essentially blaming the A.I. who were treated to abusively for lashing out. Put me in a cage, and poke me enough times, and I'll rip your throat out the first chance I get too. The A.I. problem, was created BY the mistreatment of A.I.'s by humans. Which is exactly what I've said before. So maybe you should really,

heck, Deus ran into a kind of cognitive dissonance. First its instilled with a sense of loyalty to the renraku corporation, but then at the same time the CEO have certain chains and a kill switch embedded into it. So, no matter how loyal it is, they wont trust it, then why bother being loyal in the first place?

[blackwulf]

A couple things I would have to argue SR4 is not an enviroment of sweetness and light. Any AI created in this world is going to reflect that. As I recall in emergence they were performing vivisections on living people, You think anybody willing to do that is going to give a damn about an AI or any AI born in that enviroment will give a s--t about people? Second, and this ought to make me popular how many sinless involuntary in specific are going to go an program can have citizenship and i can"t VIVA la revolucion. If that isn't born naturally someone will feed it and water until it is. Just a couple of thoughts Blackwulf

[Tzeentch]

thing is, the SR4 AI is not much different from a SR3 SK. They where creations of very high powered hosts and very custom programming, and was "chained" to the host as it was needed to maintain long term stability of the SK. A SR3 AI could in comparison gather the needed CPU resources by setting up a kind of cluster across the matrix.

-- SR4 AI's have little in common with the semi-autonomous knowbots of SR3, actually. The entire idea behind what constitutes an artificial intelligence is different between the editions. Things radically changed with the introduction of Deus and the otaku.

basically, the SR matrix have have started showing emergent properties, thanks to a growing complexity of the hardware and software involved.

-- Emergent behavior is complexity arising from simple rules. Shadowrun 4 AI basically boils down to sapience by (magical, as it just sort of floats around in magical cyberastralspace and isn't something you can see loaded and copy) algorithm which I suppose is a version of this. Before SR4 it was assumed that AIs required rather massive infrastructure support to maintain cognition. The rumor that they could survive via networking small processes all over the Matrix was (in my mind) rather more ridiculous than a sapient algorithm though, as the connection delays would result in something very strange and alien.
-- Say what you will about Shadowrun AI, but they are a perfect example of "human in a funny suit" no matter how it tries to spin things.

as for the non-copy issue, remember that SR have a history of DRM that "works"

-- Quite the opposite, it's never worked in Shadowrun canon. Remember that cyberdecks used to be locked at the manufacturer to prevent them from being used for cybercrime, but all it took was a wily hacker to replace the chipset with a "stealthed" version and you were good to go. This is functionally the equivalent of modding your console with the most crazy DRM ever seen. Software was not copied because noone trusted that the software wasn't actually malware, and the SoTA rules were absolutely ridiculous (IMG:style_emoticons/default/smile.gif)

(tho SR4 have undermined that to a fair degree, much like it has other old truisms of SR matrix). Thing is, ever since crash 1.0, SR computers have behaved differently from RL computers on a very deep level. I think a early corp book talked about the foundation of renraku or fuchi was based on a memory technology that was incompatible with the pre-crash computers (makes about zero sense, but that was early SR for you).

-- I've given up trying to rationalize the Shadowrun Matrix (and this from a guy who wrote a 250+ page fan sourcebook on the subject). The developers keep trying to stay relevant but things just turn out screwy (IMG:style_emoticons/default/smile.gif)

[hermit]

and this from a guy who wrote a 250+ page fan sourcebook on the subject

Can this be found somewhere on the internet?

[SkepticInc]

Maybe the reason the AI are in the MeshMagicMetaplane (MMM) is because they use the same metaphor for communication.

************* Thinky bits up here *************
[PROGRAM] <---------{talking}------> [SPELL FORCE]
||--------------------------------------------------------||
[SYSTEM] <-----------{talking}-------> [SPELL SKILL]
||--------------------------------------------------------||
[RESPONSE] <--------{talking}----> [MAGIC RATING]
||--------------------------------------------------------||
[FIREWALL/SIGNAL PROTOCOL STACK/WALL OF FIRE]
*********** Crunchy bits down here ************

You have a protocol stack on each side. I don't see the problem.

[Sengir]

Before SR4 it was assumed that AIs required rather massive infrastructure support to maintain cognition.

Yeah, the old rules where basically all or nothing: Omnipotent matrix god or no sapience at all. The new AIs can work with significantly less resources, but are accordingly less powerful.

The rumor that they could survive via networking small processes all over the Matrix was (in my mind) rather more ridiculous than a sapient algorithm though

Ever heard of something called "The Network"? (IMG:style_emoticons/default/wink.gif)

The descriptions of Morgan and Mirage also imply that they lived within the matrix as a whole, not a secret UV host. For example, where should Morgan have gotten a sufficiently powerful host from after her escape from the Arcology?

as the connection delays would result in something very strange and alien.

What connection delays, as long as you don't use satlinks there are none (IMG:style_emoticons/default/wink.gif)
Again, the SR matrix cannot be compared to any real computer network. If you try to "rationalize" the matrix by comparing it to the real world you're just going to hurt your brain...personally, my approach is similar to astral space: The rules say it exists and this is how it works, so when playing the game I stick with that and not real life facts like the nonexistance of magic or the complications of ad-hoc routing.



And a random philosophy bit for tonight: For all we know, "imperfections" like latency and concurrency issues could be exactly what makes conciousness...

[Doc Chase]

Ever heard of something called "The Network"? (IMG:style_emoticons/default/wink.gif)

Yeah, but CBS isn't around in 2072. (IMG:style_emoticons/default/wink.gif)

The descriptions of Morgan and Mirage also imply that they lived within the matrix as a whole, not a secret UV host. For example, where should Morgan have gotten a sufficiently powerful host from after her escape from the Arcology?

I would think Mirage had his own host from the old Echo Mirage servers. Probably not UV, but there's nothing else on them. Morgan - well, Dodger's love of a knowbot created her, so clearly he can make a freakin' UV node or something. (IMG:style_emoticons/default/wobble.gif)

[hermit]

Actually, the AI in the Arcology had an UV host right from the first mention in the SoP novels (back when it was just a really complex knowbot, it only became an AI becaue of Dodger's Twu Wuv).

Not to say those AI made a whole lot of sense or were very coherent. They wre, up to SR4 deciding they should be playable and plentyful, rare and a pure story device used by several authors for different purposes.

[Mordinvan]

Then go fight against whaling or something?

Why don't you?
You are promoting the "idea" of racism and bigotry as being a good thing. It wouldn't matter where this was, or who was saying it. I'd call it bad. The vast majority of the 50k + metasapients are intelligent beings who just wish to live their life. The few major mass murders you list became what way because of human prejudice, torture and imprisonment. If you fail to see an analog between how YOU are proposing A.I.'s should be treated, and how real humans have been oppressed in the real world, then YOU are part of the problem, in the real world.

The unwillingness to empathize with another group outside of what you perceive your own to be is actually how most of the atrocities in human history have been allowed to occur.

[Tzeentch]

Why don't you?
You are promoting the "idea" of racism and bigotry as being a good thing. It wouldn't matter where this was, or who was saying it. I'd call it bad. The vast majority of the 50k + metasapients are intelligent beings who just wish to live their life. The few major mass murders you list became what way because of human prejudice, torture and imprisonment. If you fail to see an analog between how YOU are proposing A.I.'s should be treated, and how real humans have been oppressed in the real world, then YOU are part of the problem, in the real world.

The unwillingness to empathize with another group outside of what you perceive your own to be is actually how most of the atrocities in human history have been allowed to occur.

-- I can't tell if you are roleplayng a Horizon rep or are actually serious.

[Mordinvan]

-- I can't tell if you are roleplayng a Horizon rep or are actually serious.

edit. Roughly it translates out to thinking that all racism against sentient entities is bad in general, and while I understand it is part and parcel of the SR4 world, the way in which Hermit is displaying it is exactly how it allows genocide in the real world, and I don't believe he's considered the ultimate causes or consequences of such an attitude.

[DeathStrobe]

Uhm, no, the worst case is Skynet, and SR already had one of those ...

AI's aren't that smart, or care enough about metahumanity to go Skynet on the 6th World, at least anymore. If we go with the fluff that Emergence set up about AI's they really only care about the Matrix. And the handful of AI's that would bother to interact with metahumanity are programs that were originally designed to socialize with humans, like a singing idol game that became self aware.

As cool as Terminator is and the whole Skynet syndrome AI's have in pop culture, its been done. That's why, I assume, the writers have taken a different approach to AI's in SR4. They can't constantly keep having bigger badder AI's ruling the Matrix or else it'll get boring. You enter in to a problem of the Matrix being so dangerous that only the 1337 Hacker can use the Matrix or that there is no more Matrix because no one would allow such a powerful being to run rampant and uncontrolled, so all the Corps and Government shut down the Matrix forever, and the whole Shadowrun universe returns to high fantasy because there is no more technology.

Really, now, what on earth do you people want from AIs? Do you want them to be so superior to metahumanity that we get a Terminator scenario of man vs machine? Maybe a scenario like Wall-E where machines have effectively enslaved us with comfort and humanity becomes arbitrary, and maybe our AI overlords can begin to slowly kill us off with brainwashing campaigns where they encourage us to kill ourselves off, or slowly work a genetic defect in to our system lowering the birth rate. I'd like to take some create, but I actually stole these ideas from Ted Kaczynski (AKA the Unabomber).

Mr. Kaczynski is the poster child of the anti-AI movement. He saw all that crazy stuff in Shadowrun happening way back in the 1970's. Hm...this gives me a great idea to create an antagonist group in Shadowrun...

Anyway, I like the direction they're taking AI's in SR4. Super AI's that operate at Godly levels are boring. I like the idea that those Super AI's died and their code some how entered in to normal programs and somehow made them become self aware after Crash 2.0. I think its an interesting plot hook. And allows the Matrix to be more then just 2 Super AI's having an Online War.

[DeathStrobe]

damn double post