Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
I've noticed that most sci-fi is dominated by strong male characters designed to appeal to men, with the occasional strong female character designed to appeal to...men. (Throw in a token minority and an inexplicably human alien and you've got your starship a crew!) Why the sci-fi community is then surprised that it is so male-dominated is beyond me. This has led to a self-fulfilling prophecy of sorts, where sci-fi (and fantasy, outside of "paranormal romance") is presumed to appeal to men almost exclusively by default, leading to more and more male-oriented works being created, which (when Stripperific Space Bitches inevitably falls to take with female demographics) leads more content creators and consumers to presume that sci-fi is inherently masculine, which leads to more male-oriented works, &c. This cycle colors everybody's perceptions of the genre, and is loaded with stereotypes and unfortunate implications all around (is the reason sci-fi isn't more popular with women because science itself is intrinsically masculine? Really?).
I want to take that cycle, fuck it, and then throw a wrench in it. I think all art worth doing should have a reason to be, should attempt to put a ding in the universe in its own small way, and I think it's a shame that young girls might be turned off from liking sci-fi (or worse, the sciences) because all they see are Scantily Clad Space Hooker X or Cyberpunk Razorgirl in Tight Leather X. I'm not a great writer, not even a writer, really, and I'll probably never be the most clever or original. But if I could make science just that much less of a sausage party, I think I'd be okay with whatever criticisms I got. We all need heroes to inspire us, and there's a lack of intelligent, fleshed-out, complicated heroines in sci-fi. (This is why Contact is my favorite sci-fi movie. Send hate mail here.) And so...I'm writing sci-fi? Specifically, a five-part miniseries, with the hope of it becoming a substantially less mini series if things go well (the episode titles spell out P-I-L-O-T ). Both the genre and the medium are new to me, but I reasoned the best way to get a story out of the bookstore's sci-fi ghetto is to not make it a book, and I've been dying to experiment with serial storytelling for some time. ("No, not experiments with serial storytelling!" *explosion* LOST.) It's been going okay so far (well, not for the crew!), the conflict is natural and plentiful, and the miniseries format gives me ample opportunity to introduce the characters at the small cost of ensuring that the show will never see production. I'm finding the territory not entirely unfamiliar, as the current sci-fi genre seems be prone to many of the weaknesses of fantasy, namely sameness and over-emphasis of setting and showing off the writer's research over things like characters and plot. (The trick, of course, is to write something that's captivating as a down-to-Earth character drama first, but it's much easier to hide behind the literary equivalent of special effects.) But there's challenges involved in writing a story that's set in Our World but still fantastical. Like, math. I was hoping one of you could help me out on that regard. Okay, so, the best thing about writing sci-fi as opposed to fantasy is obviously robots, and of course I had to have one. (Well, more than one, but one "main" one.) I won't bore you (any more than I have ) with the details, but he's based on three processors, with a combined 18 petaFLOPS of processing power (a multiple of nine, significant? In one of my stories? No way!). The most powerful of his processors runs at 12 petaFLOPS. That is, of course, far more powerful than even any massive supercomputing cluster today, and it's one processor that has to fit inside the chest cavity of a bulky but humanoid robot and run on batteries without exploding. What I want to know is, with reasonable assumptions made regarding Moore's law and the like, about when would such a processor be available (using Earth technology)? I can't quite figure out the maths myself. I get that it's probably going to be a very vague estimate but I just want to make sure it's sometime later than next week but before the heat death of the universe. Are we talking, like, 50 years, or 500, or...? I'm trying to avoid a situation like where The Computer on a 23rd century starship has "over a million transistors!" or something, but I don't want to be implausible either. It's all backstory but it's Symbolically Significant and I'd like the series to be more grounded, so I don't want to just use meaningless units (or unit-less numbers). Thanks for reading, or skimming, or doing whatever it is people do instead of reading these days. (Watching TV?) P.S. The working title of the series is "Still," though people who follow me on Twitter may know the real title. I'll try to avoid talking about it too much. and i guess i've known it all along / the truth is, you have to be soft to be strong |
quote |
I shot the sherrif.
|
I would guess that you're more likely to hit quantum computing which would provide more 'power' for a robot, than a traditional present day processor reaching 12 petaflops of processing power, but I'm sure one of our more technically inclined members would be able to chime in with better, actual data.
Google is your frenemy. Caveat Emptor - Latin for tough titty I tend to interpret things in the way that's most hilarious to me |
quote |
Not a tame lion...
Join Date: May 2004
Location: Narnia
|
I think the question should be answered by a girl
|
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
Quote:
I wanted to avoid using technology that would be too alien or require too big of a breakthrough to have a scientific basis, i.e., magic. "The robot's 'brain' was a giant glowing neurocrystal!" sounds suspiciously similar to "The doll's 'brain' was a giant glowing Crystal of the Gods!" That was my reasoning behind my "like what we have now, but better" thinking. But I'm not at all a computer person so I have no idea if that makes any sense. Maybe the Intel Quentium is right around the corner, I dunno. and i guess i've known it all along / the truth is, you have to be soft to be strong |
|
quote |
Veteran Member
Join Date: Jan 2005
|
1.21 Gigawatts.
|
quote |
Veteran Member
Join Date: Jan 2005
|
Ok, so find out how many FLOPS a human brain does, and that will be roughly your answer. Apparently, Kurzweil said it's 20 quadrillion FLOPS (20 teraFLOPS, I think). I'd go with that.
|
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
Ha ha
Quote:
And the robot isn't really designed to emulate a human, because that would sort of be a waste of a character, no? It'd be like the inexplicably human aliens mentioned above. The ship's autopilot runs at 90 petaFLOPS, but it's a large array of processors, so that's different. and i guess i've known it all along / the truth is, you have to be soft to be strong |
|
quote |
Veteran Member
Join Date: Jan 2005
|
robo, I made a mistake. I meant to write 20 petaFLOPS.
The #1 supercomputer in the world right now does a little over 2 petaFLOPS peak. It has over 200,000 CPUs. If the brain does 20 petaFLOPS, that's like 2,000,000 cores. Intense, dude. You may want to bump that up a few orders of magnitude. Say 20 exaFLOPS. |
quote |
Veteran Member
Join Date: Jun 2006
Location: Florida
|
I like the realisim approach. I don't have an answer for you (math is above my pay grade) but I'll pitch it to my NASA friends tomorrow and see what they have to say.
|
quote |
Veteran Member
Join Date: Jan 2005
|
robo, if you do write a scene where FLOPS are used, don't dare explain to the audience what that acronym stands for.
|
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
So, wait. My totally-stabbing-in-the-dark symbolically-minded choice of the robot's processing power (18 petaFLOPS total) just happened to almost match Kurzweil says the human brain runs? Interesting!
Quote:
The most powerful single processor that's mentioned in the story, IBM's Blue Will, is currently set to run at 1000 petaFLOPS. But that's later. and i guess i've known it all along / the truth is, you have to be soft to be strong |
|
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
Quote:
Quote:
I have no intention on launching into a lengthy explanation on what FLOPS actually are, don't worry. That'd be boring, and I'm not one of those writers who writes just to show off their research. and i guess i've known it all along / the truth is, you have to be soft to be strong |
||
quote |
Veteran Member
Join Date: Jan 2005
|
Quote:
|
|
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
Yes, but A) not switching units makes it easier for non-technically-minded viewers to compare (1000 is obviously more than 100, but an exa- is less obviously more than 100 peta-) and B) as the most powerful single processor in the world, the Blue Will would only be compared to petaFLOPS-class processors, and could thus comfortably share their already-familiar units, just as the first GB drives were dual-branded 1000 MB. Also, C) Symbolism.
(I knew someone would post that, and I was ready. ) But I'm still curious what you meant when you said I should "bump that up"? Were you referring to the robot's CPU(s), or the autopilot's? And why? and i guess i've known it all along / the truth is, you have to be soft to be strong |
quote |
Not a tame lion...
Join Date: May 2004
Location: Narnia
|
Wikipedia says that a Core i7 can do 107 GigaFLOPS, so about 10^11 FLOPS.
If we make the assumption that personal computers have been doubling in power every year since their introduction then we can say that computing power is increasing at an exponential rate. So, looking at 10^11 = 2^x and solving for x should tell us how many years it took to get to our current 10^11 FLOPS. 10^11 = 2^x x = 36.5412... So, roughly 36 years after the introduction of the personal computer which sounds about right. Assuming your robot isn't anything special and is using off the shelf processors, then we can simply replace 10^11 with 10^15 petaFLOPS and solve for x again. 10^15 = 2^x x = 49.8289... So, in approximately 15 years our personal computers should be capable of petaFLOPS. Thanks Wolfram Alpha Last edited by AsLan^ : 2010-06-09 at 00:05. Reason: Complete rewrite! |
quote |
Not a tame lion...
Join Date: May 2004
Location: Narnia
|
Not sure if you caught it but I rewrote my answer above, I forgot that the 50 years was from the first PC, not from now.
15 years in the future might be a little close for science fiction, and due to the nature of exponential growth, even bumping it up to the exaFLOPS level only takes another 10 years! I'm sure you'll think of something |
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
Quote:
Quote:
Quote:
and i guess i've known it all along / the truth is, you have to be soft to be strong |
|||
quote |
Veteran Member
Join Date: May 2004
|
No petting flops! No petting flops! Only exxon flops! WHY THE PETTING FLOPS.
|
quote |
Not a tame lion...
Join Date: May 2004
Location: Narnia
|
Quote:
Quote:
Also, there's a couple of problems with Moore's law. 1. It's not really a law, just an observation by Moore that seems to be holding up so far. 2. It has to do with transistor density on dies rather than computer speed. So, is computer speed increasing at an exponential pace? So far yes, but when will it end? If you know the answer to that then I'm sure Intel has a job for you In fact, it seems Intel can now arbitrarily add cores so there shouldn't be any limit to the number of FLOPS a chip can do but what meaningful work can it do? This turns the problem into a software problem. That is, each added core linearly increases the amount of operations a CPU can perform (2 cores are twice as fast as 1 core and so on) but our ability to create software to take advantage of this remains limited. In fact, if we look at your robot intelligence as a software problem rather than a hardware problem I think the time scales would become much larger, and advances in software definitely have not followed an exponential growth pattern. Think about what computers could do in 1993 and think about what they can do today, not really that much different - there are still programs actively being used today that were created the 1970s! Your robot could very well be processing 12 petaFLOPS in the next 30 years or so, but will humans have created the necessary software in that same time? Hopefully we create true AI in my lifetime but nobody even knows if it's a solvable problem |
||
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
Quote:
Quote:
Quote:
Quote:
and i guess i've known it all along / the truth is, you have to be soft to be strong |
||||
quote |
Selfish Heathen
Join Date: May 2004
Location: Zone of Pain
|
|
quote |
Selfish Heathen
Join Date: May 2004
Location: Zone of Pain
|
As for performance scales, consider this:
The ENIAC in 1947 consumed 174,000 watts of power, weighed 27 tons, covered 63 m^2, and performed at roughly 350 flops. Today, the Intel Core i7 consumes about 100 watts of power when under full load, weighs a few grams, covers a few square cm, and (as mentioned) can theoretically perform ~100,000,000,000 flops. The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting. |
quote |
Not a tame lion...
Join Date: May 2004
Location: Narnia
|
Quote:
Quote:
|
||
quote |
Selfish Heathen
Join Date: May 2004
Location: Zone of Pain
|
And before anyone points out that ENIAC wasn't transistor-based and that Moore's Law doesn't necessarily apply, consider that over longer periods of time one shouldn't be constrained to a specific form of processing technology. Just as we went from vacuum tubes to transistors, we'll eventually move from microprocessors to quantum processors or some other technology that will advance us leaps and bounds.
The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting. |
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
@Brad: So you're saying that 12 petaFLOPS isn't that big of a deal, then? It has to be a bit more than 15 years out, I mean the show might actually be airing then and that would be awkward. 30 years out would be okay, but I might bump everything up an SI prefix just to be safe. Hmm.
and i guess i've known it all along / the truth is, you have to be soft to be strong |
quote |
Selfish Heathen
Join Date: May 2004
Location: Zone of Pain
|
Heck, no, not a big deal.
Shoot, just think about small time scales. Anyone else remember in '99 when Jobs was touting how the then-bleeding-edge G4 processor was classified as a "supercomputer" because it cranked in 1-4 gigaflops? 1 to 4 gigaflops! A mere decade later, the i7 performs 25-100 times that many flops. I dunno about you, but that kind of advancement in just a short span of one's lifetime is pretty mind-boggling. The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting. |
quote |
Formerly Roboman, still
awesome Join Date: Jul 2004
Location: Portland, OR
|
Quote:
and i guess i've known it all along / the truth is, you have to be soft to be strong |
|
quote |
Selfish Heathen
Join Date: May 2004
Location: Zone of Pain
|
True, good point. That would take a bit of time. My brain is still waking up and wasn't quite putting the scales into context with all you'd described here.
The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting. |
quote |
Not a tame lion...
Join Date: May 2004
Location: Narnia
|
Quote:
Quote:
Quote:
More than a decade sure, but probably not more than two decades. |
|||
quote |
Posting Rules | Navigation |
|
Thread Tools | |
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
iPad: Computer for a new "casual user" market? Redefining "personal computing"? | ezkcdude | Apple Products | 626 | 2010-04-03 13:51 |
Disney/Pixar announce releases through 2012, add "Newt," "Cars 2" | Robo | AppleOutsider | 28 | 2008-05-17 11:39 |
Kernel Panic, "System Failure" + "Corrupt stack" | SpecMode | Genius Bar | 2 | 2006-01-16 13:42 |
"Date Modified" selection in "Open" windows, buggy? | nathanaaron02 | Apple Products | 2 | 2005-03-09 18:10 |
Magnetic "Support the Troops" ribbons/bumperstickers miss the point. ("idiotic") | Paul | AppleOutsider | 2 | 2004-10-31 01:26 |