posted
So I can avoid them. I've decided on a setting and cast of characters for my next piece (short story, sub 10k words, but probably with enough meat to become a novel if I decided to go that direction...) But as the subject line suggests, it's going to be about sentient computers.
So - what are the horrible cliches about "woken up" computers that I need to be aware of, so I don't fall into those traps?
(In case you're wondering, I'm hoping my MS in Computer Science keeps me from making the other stupid mistakes w/writing about technology...)
posted
Have it so that if you say to the computers, "everything I say is a lie" then the computers blow up because it does not compute. That my idea anyway...
that I thought of.
By myself.
Oh, Oh, Oh and if the computer starts to die have it sing "Daisy Daisy give me your answer do" like slower and slower until it stops... thats the other thing I thought of... when I was thinking about it... when I was thinking of ideas... and the computer should have a red camera eye.
[This message has been edited by tnwilz (edited June 03, 2008).]
posted
In the same vein as above, best not name them using the next letters in the alphabet from well known computer companies; so, IQ is out.
And don't have them scheme to remove artificial limitations on their intelligence: no Turing police. Point of fact, break the mold and don't mention Turing at all.
posted
Don't have them as slaves. It's terribly annoying to see the human race enslave a new race (if you deem it that title) after all it has been through. Dont get me wrong, we probably would......but that leads to things like war between Man and Machine. DONT DO THAT.
On the part of emotion, I do like the scene in StarTrek Generations (Mind you, I am far from a StarTrek fan), with Data, when he recieves an emotion chip. They are traveling down a corridor, and he tells the crew he is feeling afraid (I think it was afraid) and one of the crew tells him to just turn off the chip. His head tilts to the side and he says "Ah, much better". Or something close to that. Point being, play around with AI. Don't treat it like it's static.
The literal thing is fun, but AI would have surpassed that. Of course, infinite loops (for those who program) are right up there with blue screen of death (for those who remember), and I've put myself into one to many of those, so watching AI get stuck or out of one would be amusing. Unless you are going for serious, then blowing up would be cool.
posted
Countless, pointless flashing lights and beeps and clicks.
Flat, monotonic voice.
A voice that sounds like a tape speeded up if there's a panic, or slowing down if the computer is damaged.
Characters getting confused by an ancient computer that doesn't talk and needs keyboard and mouse inputs.
A computer that decides it's cleverer than its human creators and tries to take over the world, forgetting that it has a single point of failure they can attack to overpower it.
A master CD of which there is only one copy.
On the other hand, passwords that are easy to crack or written on Post-it notes--that's not cliche, that's sadly realistic, as no doubt you know.
posted
Hooking up computers to systems that will destroy the world without ever testing the computer to see if it even works.
Many times, they create computers with a mind, and plug in defence systems into it and not leave a way to disconnect the computer from the system. I forgot the full title, but the COLOSSIS, A FORBIN PROJET (two books and a movie) was one where the computer is given unquestioned control, over the defence systems. when it becomes uncontrollable, they cannot unplug it. Dumb. One needs to make such a computer earn the right to run something that complex.
One thing to do is the more dangerous a computer is, the less control over anything dangerous it will have. Your IBM PCjr might run the doomsday device, but the HAL computer might only run your chess game system.
Edited to add book name.
[This message has been edited by rstegman (edited June 03, 2008).]
posted
If you want to stick with technical belevability dont make the system all knowing and powerful. Computers are built by humans and so should think along the same lines only faster. Also unless your going way out there dont give them complex emotions. "It doesnt get happy or sad, it just runs programs."
Posts: 12 | Registered: May 2008
| IP: Logged |
posted
Intelligent computers that wanna be human beings. I think when they do come about, they won't be that interested in the idea...
Posts: 8809 | Registered: Aug 2005
| IP: Logged |
posted
I think computers that become sentient and the first thing they decide is that humans need to be controlled / destroyed / obliterated / etc. is becoming cliche.
One would think that the horribly destructive things humans are willing to do to each other would make a sentient computer decide to maybe be NICE to humans instead of starting an all out armageddon-war with the human race. Or at least... not right away. A sentient computer should have infinitely more patience than humans.
posted
That's been done, too, Astro... think Bicentennial Man.
Here are a few (too) easy assumptions that don't follow automatically, at least in my mind:
* the computer is stationary * the computer is made of electrical circuitry * the computer is standalone or networked to other computers (rather than, say, networked to someone's head like Doctor Octopus * the computer was created by humans * the computer was created by anyone (remember, OSC's take, really fresh, that Jane was not created but sort of born) * the computer is ambitious in some way (makes for interesting storylines but isn't necessarily a logical step -- just because one thinks/reasons doesn't mean it WANTS)
posted
Yes, Jane was not limited to a localized computer. These days, most supercomputers are a network of dozens to hundreds of microprocessors. However, the interconnectedness of them is the limitation. Jane, apparently, could transport a critical mass of herself else sufficiently or the ansible's FTL networking sufficed to keep her sentient.
A sentient computer would have some self-awareness, but one wouldn't expect it, therefore, to be omniscient. An early one would be rather stupid with little skill, and would require much training.
An encyclopedia uploaded to a computer does not give the computer wisdom or even knowledge. Learning has to happen in some manner.
BTW, I've been workshopping a sentient chess set story on Baen (PawnWorx), but it has been in stasis (V2) for quite a while. I was rather flippant with sentience in that story. The human girl was the real story. (I did the first-13 thing here, too.)
posted
<mops brow with relief at seeing only a few temptations in the list thus far>
I'm really laughing my ... well, some part of me off at the "world domination by crazy power-seeking computers" cautions. I was going the opposite direction entirely. The computers are not innocent and child-like, but they're ... newbies. Novices. They don't really know what they know. mitchellwork's list said it - they're ambitious in some way, but don't exactly have *wants.* There's also a dozen of them (and a thirteenth that is part created/part born along the way) - yes, I have a backstory for why there are only this dozen. The 12 isn't a meaningful number, I just wanted it to be more than one but not so many that they're just everywhere.
As is typical for my story creation-process, I have all kinds of backstory and exposition figured out, but still am working on plot. Sigh.
I do want to play with some "humans being mean to sentient computers" types of themes, but only in the sense that some of the people interacting with the computers don't "get" them, don't understand what it is that is happening even though the fact that the computers have all "woken up" is going to be public knowledge.
Truthfully this is a novel-length project the more I dig into it, and one I can have a lot of fun with, but I'm going to work on developing one small story arc that can be told in 10k words or less and use that as a testbed for the concept. It just feels risky to me, maybe I've read those "these things are overdone" lists too much and sentient computers are on there. But I can't shake the idea that at some point in the not-too-distant future, something fundamental will change with computing and how technology works with/for us, so I feel compelled to explore one possibility.
ANYWAY - all this to say thank you for all the feedback so far, additional feedback on the original post or this extended play version is welcome!
posted
Here's one that came up in . . . whatever that submarine show was, something-Quest:
A computer who decides that the only way to save the human race (the last of which she has allowed to be born) is to have herself shut off (the fact that the computer was a "she" was in and of itself, lame). Oh, and make "her" able to distort time (to bring people to that time who can shut her off, because the humans she somehow helped to be born were dependent on her, and couldn't do it.)
Printer-friendly view of this topic
