A processor that's faster than a speeding bullet, and more deadly!!
December 2, 2009 6:25 PM
Let's say you had the fastest super-computer the world has ever seen, and you were a villain who wanted to take over the World with it. How might your super-computer make that happen for you?
(For a story I'm writing.)
Please note: it could be either a super-computer, or a very powerful fleet of spy satellites. The important thing is, this tool can process information a thousand times faster than the best tech any other country has. That's gotta count for something in the evil genius community, right??
Given this technological advantage, please help me brainstorm ways in which a very bad dude could destabilize the West, or control the world enough to basically become a super power, because of this technology.
How might this computer (or satellite fleet ) in a short amount of time infiltrate and control:
--the stock market
--the defense system
--world trade
--whatever else you can think of.
I'm looking for something a little more sophisticated than the diamond-coated satellites of "Diamonds Are Forever" that direct giant laser beams, but more exciting than "create lines of code that would shut down the internet."
Thanks in advance!
(For a story I'm writing.)
Please note: it could be either a super-computer, or a very powerful fleet of spy satellites. The important thing is, this tool can process information a thousand times faster than the best tech any other country has. That's gotta count for something in the evil genius community, right??
Given this technological advantage, please help me brainstorm ways in which a very bad dude could destabilize the West, or control the world enough to basically become a super power, because of this technology.
How might this computer (or satellite fleet ) in a short amount of time infiltrate and control:
--the stock market
--the defense system
--world trade
--whatever else you can think of.
I'm looking for something a little more sophisticated than the diamond-coated satellites of "Diamonds Are Forever" that direct giant laser beams, but more exciting than "create lines of code that would shut down the internet."
Thanks in advance!
Although the strength of the computer itself is in the realm of sci fi, I don't want to make it TOO sci fi, so I'm staying away from artificial intelligence. More interested in it as a tool for terrorists to do something specific.
posted by np312 at 6:36 PM on December 2, 2009
posted by np312 at 6:36 PM on December 2, 2009
Breaking heavy encryption? Lots of secure networks, ripe for plunder or infiltration? Intellipedia would be an interesting plot-point.
posted by Paragon at 6:37 PM on December 2, 2009
posted by Paragon at 6:37 PM on December 2, 2009
A factor of a thousand isn't really going to get you into the realm of effective decryption, which is where most of the supervillain action would be. In technical terms it would only get you through about 10 additional bits of keyspace.
Stock-market wise, a fast system is certainly of value, at least to a point. There are companies that sell special-purpose boxes for analyzing trades in real-time. The real trick isn't the speed of the computer, though, it's the intelligence of the trading algorithm. Maybe you could say your supervillain had devised a really powerful algorithm that was too slow for anything less than his or her supercomputer.
posted by jedicus at 6:44 PM on December 2, 2009
Stock-market wise, a fast system is certainly of value, at least to a point. There are companies that sell special-purpose boxes for analyzing trades in real-time. The real trick isn't the speed of the computer, though, it's the intelligence of the trading algorithm. Maybe you could say your supervillain had devised a really powerful algorithm that was too slow for anything less than his or her supercomputer.
posted by jedicus at 6:44 PM on December 2, 2009
Maybe the super computer gets on ask.mefi and poses a question about how it might take over the world and a cadre of super nerds help out thinking they are helping some aspiring scifi writer?
posted by ian1977 at 6:45 PM on December 2, 2009
posted by ian1977 at 6:45 PM on December 2, 2009
Ok, sorry to be a bit of a spoiler, but it might be worth saying that if it is 1000 times more powerful that what the most powerful machine around was you'd have a computer that was about 10 years ahead of its time.
Check this timeline of machines here or this graph here
If you made it a million times you'd get about 20 years.
If you did, figuring out ways to trade might be worth doing.
posted by sien at 6:48 PM on December 2, 2009
Check this timeline of machines here or this graph here
If you made it a million times you'd get about 20 years.
If you did, figuring out ways to trade might be worth doing.
posted by sien at 6:48 PM on December 2, 2009
Heavy-duty supercomputers are useful in cryptography (at least in fiction). Here's an abstract from December 2000 that talks about cryptosystems as a double-edged sword in electronic communication. Might be some interesting jumping-off points in here if you could get the whole article.
posted by Quietgal at 6:56 PM on December 2, 2009
posted by Quietgal at 6:56 PM on December 2, 2009
Taking over the world creates a lot of enemies in return for a lot of headaches.
Use it to play the stock market, amass wealth. Take over a country, preferably an African one. The world doesn't seem to care much about what happens there and it's geographically far from the superpowers so they won't feel threatened. Somolia perhaps?. Build your army, secure your borders and resources. Use some of your wealth to get a few nukes. Set off a test to let everyone know you have them.
Now you've got a secure country where you can do whatever you want and no one wants to mess with you. If you get bored, there's plenty of other countries to invade, but I'd recommend not doing that. You want power, not entanglements with other countries.
posted by Brandon Blatcher at 7:09 PM on December 2, 2009
Use it to play the stock market, amass wealth. Take over a country, preferably an African one. The world doesn't seem to care much about what happens there and it's geographically far from the superpowers so they won't feel threatened. Somolia perhaps?. Build your army, secure your borders and resources. Use some of your wealth to get a few nukes. Set off a test to let everyone know you have them.
Now you've got a secure country where you can do whatever you want and no one wants to mess with you. If you get bored, there's plenty of other countries to invade, but I'd recommend not doing that. You want power, not entanglements with other countries.
posted by Brandon Blatcher at 7:09 PM on December 2, 2009
As has been mentioned upthread, breaking any sort of reasonable cryptography is still out of the question (unless you assume that it's a quantum computer, at which point all bets are off)
Before anything else happens, you're going to need some cash - a few ideas:
- Make a ton of money on the stock market. Should be easy enough to come up with some advanced models that return well.
- Apply it to the problem of protein folding or drug design, come up with some pharmaceuticals, then sell them at a ridiculous profit.
After you have some cash:
- Use it to calculate trajectories of incoming ICBMs, and build a reliable missile-defense system. This could have seriously destabilizing effects by eliminating the doctrine of mutually-assured destruction.
posted by chrisamiller at 7:16 PM on December 2, 2009
Before anything else happens, you're going to need some cash - a few ideas:
- Make a ton of money on the stock market. Should be easy enough to come up with some advanced models that return well.
- Apply it to the problem of protein folding or drug design, come up with some pharmaceuticals, then sell them at a ridiculous profit.
After you have some cash:
- Use it to calculate trajectories of incoming ICBMs, and build a reliable missile-defense system. This could have seriously destabilizing effects by eliminating the doctrine of mutually-assured destruction.
posted by chrisamiller at 7:16 PM on December 2, 2009
Crypto is the answer. The entire basis for current encryption is that it takes [Hitchhikers Guide the universe is really really big number] of calculations using prime numbers to break encryption. The process is straightforward but good encryption would take thousands of years to break brute force with current computers. Break strong encryption and every bank, military network, and spy agency is open for you to browse.
posted by Babblesort at 7:21 PM on December 2, 2009
posted by Babblesort at 7:21 PM on December 2, 2009
Also, I think you're missing out on a major plot point: Why do they want to take over the world? Doing it for money or power doesn't want because there's plenty of ways to get those without literally taking over the world (see any international businessperson). No, they need a major reason, something beyond the desire for wealth or power. They need believe they're doing this for the betterment of the world. The best villains are usually the ones that absolutely believe they're right and are willing to go any lengths to achieve their ends.
Give the villain a mission and have the super computer be what it ought to be: a tool to achieve that goal.
posted by Brandon Blatcher at 7:26 PM on December 2, 2009
Give the villain a mission and have the super computer be what it ought to be: a tool to achieve that goal.
posted by Brandon Blatcher at 7:26 PM on December 2, 2009
The power grid. Get it to shut down seemingly random areas of the world. Without the grid there is no civilization.
posted by ~Sushma~ at 7:30 PM on December 2, 2009
posted by ~Sushma~ at 7:30 PM on December 2, 2009
You might want to make your fictional computer qualitatively different from conventional computers. As others have pointed out, a factor of a thousand isn't much.
The basis of modern cryptography is the assumption that P != NP. If your fictional supervillain had a magical computer that could solve NP-complete problems in linear time (meaning: when the input encryption key is twice as big, it only takes twice as long to break the encryption)
Quantum computers have some of these properties, so you might want to start there.
posted by qxntpqbbbqxl at 7:37 PM on December 2, 2009
The basis of modern cryptography is the assumption that P != NP. If your fictional supervillain had a magical computer that could solve NP-complete problems in linear time (meaning: when the input encryption key is twice as big, it only takes twice as long to break the encryption)
Quantum computers have some of these properties, so you might want to start there.
posted by qxntpqbbbqxl at 7:37 PM on December 2, 2009
... and I didn't really answer your question. With a magic computer, you could break encryption. Watch the movie Sneakers to get an idea of the ramifications of this.
Essentially, your super villain could break through the security of any system connected to a network: Banks, governments, nuclear power plants, military, etc.
posted by qxntpqbbbqxl at 7:39 PM on December 2, 2009
Essentially, your super villain could break through the security of any system connected to a network: Banks, governments, nuclear power plants, military, etc.
posted by qxntpqbbbqxl at 7:39 PM on December 2, 2009
You might consider going in a slightly different direction; Moore's Law basically states that computing power doubles every two years or so (well, literally it says you can fit twice the number of transistors in the same unit of area on a chip), so a 1000x increase is as others have noted is only a 10-years-ahead-of-time advance.
Some possibilities:
1) Your villain invents some kind of practical Quantum Computer and uses it to quickly solve some problem in BQP; the red-flag choice is integer factorization which is the hard problem that cryptosystems like RSA rely on.
2) Along similar lines to (1), your villain invents some kind of exponentially-faster computer. I'm thinking of a some kind of literal nondeterministic Turing machine here. This computer can solve NP problems in polynomial time, enabling your bad guy to solve any NP problem quickly. This is farther fetched than quantum computing (which is still kind of out there).
3) Your villain creates some kind of back-door through his unique access/genius. Imagine, for example, if Bill Gates could magically log onto any Windows machine anywhere and be allowed to do whatever he wanted to it. Presumably he would be privy to all sorts of secrets, could manipulate markets, etc. This thing could be a virus or trojan.
It sounds to me from the tone of your post that you might do well to do some more research into the implications of computational complexity. Hopefully this post will help generate some leads, but your story will sound truer if you know what it is you're talking about, and really the only way to get that is to actually know something.
posted by axiom at 7:50 PM on December 2, 2009
Some possibilities:
1) Your villain invents some kind of practical Quantum Computer and uses it to quickly solve some problem in BQP; the red-flag choice is integer factorization which is the hard problem that cryptosystems like RSA rely on.
2) Along similar lines to (1), your villain invents some kind of exponentially-faster computer. I'm thinking of a some kind of literal nondeterministic Turing machine here. This computer can solve NP problems in polynomial time, enabling your bad guy to solve any NP problem quickly. This is farther fetched than quantum computing (which is still kind of out there).
3) Your villain creates some kind of back-door through his unique access/genius. Imagine, for example, if Bill Gates could magically log onto any Windows machine anywhere and be allowed to do whatever he wanted to it. Presumably he would be privy to all sorts of secrets, could manipulate markets, etc. This thing could be a virus or trojan.
It sounds to me from the tone of your post that you might do well to do some more research into the implications of computational complexity. Hopefully this post will help generate some leads, but your story will sound truer if you know what it is you're talking about, and really the only way to get that is to actually know something.
posted by axiom at 7:50 PM on December 2, 2009
I meant to add, there is some argument to be made that Moore's Law won't continue to be true (for one thing, the size of transistors obviously has a lower bound dictated by the size of their constituent molecules/atoms). Picking a different tack than simply faster might generate more interesting results.
posted by axiom at 7:52 PM on December 2, 2009
posted by axiom at 7:52 PM on December 2, 2009
As for "a thousand times faster" just change it to "fastest.computer.ever." I wasn't being specific, obviously, about how fast it actually is.
I agree I need to do more research into what supercomputers do and can be used for. But that's the next step. The current step is figuring out if this tool could actually cause major chaos in the world, and how. If not, I'll do something else.
However no one has addressed yet the satellite idea. Any thoughts on that one?
posted by np312 at 8:03 PM on December 2, 2009
I agree I need to do more research into what supercomputers do and can be used for. But that's the next step. The current step is figuring out if this tool could actually cause major chaos in the world, and how. If not, I'll do something else.
However no one has addressed yet the satellite idea. Any thoughts on that one?
posted by np312 at 8:03 PM on December 2, 2009
On the crypto idea, all it would take is to brute-force the private key of one of the large root CAs (most of which are 2048 bit RSA) and you could then sign your own fake SSL certificates, letting you impersonate practically any https: site without the browser presenting any warning to the user. Of course you would still have to engineer some way to hijack the traffic, but luckily DNSSEC will never be implemented so hijacking DNS would be one route.
posted by Rhomboid at 8:04 PM on December 2, 2009
posted by Rhomboid at 8:04 PM on December 2, 2009
Encryption seems like the obvious answer, but how about molecular dynamics instead? As people have said, cracking modern encryption isn't going to be possible with just a few orders of magnitude improvement on computer speed, because the problem "what key will decrypt this message" needs an exact solution, requires brute force algorithms, and will have only one solution in an exponentially large search space. But the problem "what viral DNA will produce proteins that are good at bypassing a cell's defenses" can have inexact solutions, is amenable to non-brute force optimization algorithms, and may have many solutions which meet a supervillain's requirements.
Of course, it's not easy to beat nature's own solutions to that second question, because viruses themselves are a genetic algorithm for solving it. The catch is that natural selection will tend to eliminate viruses that kill the host too quickly to spread to other hosts; artificial selection has no such limitation.
posted by roystgnr at 8:12 PM on December 2, 2009
Of course, it's not easy to beat nature's own solutions to that second question, because viruses themselves are a genetic algorithm for solving it. The catch is that natural selection will tend to eliminate viruses that kill the host too quickly to spread to other hosts; artificial selection has no such limitation.
posted by roystgnr at 8:12 PM on December 2, 2009
The process is straightforward but good encryption would take thousands of years to break brute force with current computers.
No. Not thousands of years. Trillions of years. Gajillions, in fact.
A number I heard once was that if every molecule of the universe was converted to pure computronium, the heat death of the universe would come before a 1024-bit RSA key was brute-forced. My back of the envelope calculations agree with that.
A computer merely a thousand times stronger than current TOPS machines wouldn't even put a dent in breaking any reasonable encryption. Given a large enough budget, you could build such a computer tomorrow. In fact, SETI @home or the Folding @home projects arguably have that much oomph available to them at the moment. 1000 is not a very big number in computation.
OP: qxntpqbbbqxl is right. For your story to have legs, you want a computer that is fundamentally different from a contemporary computer in some way. And quantum computers are the way most people see that going.
posted by Netzapper at 8:24 PM on December 2, 2009
No. Not thousands of years. Trillions of years. Gajillions, in fact.
A number I heard once was that if every molecule of the universe was converted to pure computronium, the heat death of the universe would come before a 1024-bit RSA key was brute-forced. My back of the envelope calculations agree with that.
A computer merely a thousand times stronger than current TOPS machines wouldn't even put a dent in breaking any reasonable encryption. Given a large enough budget, you could build such a computer tomorrow. In fact, SETI @home or the Folding @home projects arguably have that much oomph available to them at the moment. 1000 is not a very big number in computation.
OP: qxntpqbbbqxl is right. For your story to have legs, you want a computer that is fundamentally different from a contemporary computer in some way. And quantum computers are the way most people see that going.
posted by Netzapper at 8:24 PM on December 2, 2009
As for "a thousand times faster" just change it to "fastest.computer.ever." I wasn't being specific, obviously, about how fast it actually is.
What I and the other computer science folks here are telling you is that a faster conventional computer is irrelevant. When dealing with these computationally complex problems, the speed of the computer is essentially immaterial. Who cares if it completes in a billion suns' lifetimes, or just a million?
A thousand times. A million times. Who cares? An intractable problem is still intractable.
The only place that the real speed of a computer matters is in real time systems (audio and video effects, stock trading maybe) and in user interface responsiveness. For instance, one algorithm I worked on in my short stint of grad school took about a week to process on the first machine I had. So we bought another, faster computer. It processed in three days instead. Then we added one more data point to the dataset, and we were back at about a week of time.
Where the breakthroughs really come is in the algorithm used to solve the problem.
Where your ultrafast computer might work well is in using genetic algorithms to generate algorithms for solving difficult problems. Attempt to genetically evolve an algorithm that breaks RSA. Perhaps blind luck will work where so many brilliant minds have failed.
posted by Netzapper at 8:34 PM on December 2, 2009
What I and the other computer science folks here are telling you is that a faster conventional computer is irrelevant. When dealing with these computationally complex problems, the speed of the computer is essentially immaterial. Who cares if it completes in a billion suns' lifetimes, or just a million?
A thousand times. A million times. Who cares? An intractable problem is still intractable.
The only place that the real speed of a computer matters is in real time systems (audio and video effects, stock trading maybe) and in user interface responsiveness. For instance, one algorithm I worked on in my short stint of grad school took about a week to process on the first machine I had. So we bought another, faster computer. It processed in three days instead. Then we added one more data point to the dataset, and we were back at about a week of time.
Where the breakthroughs really come is in the algorithm used to solve the problem.
Where your ultrafast computer might work well is in using genetic algorithms to generate algorithms for solving difficult problems. Attempt to genetically evolve an algorithm that breaks RSA. Perhaps blind luck will work where so many brilliant minds have failed.
posted by Netzapper at 8:34 PM on December 2, 2009
As for "a thousand times faster" just change it to "fastest.computer.ever." I wasn't being specific, obviously, about how fast it actually is.
Fair enough, though we're already coming up on the limits of conventional electronic chips, which won't be anywhere near fast enough for what you're talking about. To keep things believable, you might describe it as being based on quantum computing. Or maybe use photonics combined with a breakthrough in computer science that makes cracking encryption faster.
However no one has addressed yet the [very powerful fleet of spy satellites] idea. Any thoughts on that one?
Well, even assuming you had continuous, 100% coverage of the world with stereoscopic, ultra high-resolution real-time imagery and video in all spectra, you're still limited by two things: 1) most buildings and cars are opaque and 2) you can generally only look straight down.
That said, such a system would be useful for lots of things. For example, it would make missile defense a lot easier as you'd have real-time location and trajectory data. Missile defense is still a very hard problem, but you'd be a lot closer to something workable.
posted by jedicus at 8:36 PM on December 2, 2009
Fair enough, though we're already coming up on the limits of conventional electronic chips, which won't be anywhere near fast enough for what you're talking about. To keep things believable, you might describe it as being based on quantum computing. Or maybe use photonics combined with a breakthrough in computer science that makes cracking encryption faster.
However no one has addressed yet the [very powerful fleet of spy satellites] idea. Any thoughts on that one?
Well, even assuming you had continuous, 100% coverage of the world with stereoscopic, ultra high-resolution real-time imagery and video in all spectra, you're still limited by two things: 1) most buildings and cars are opaque and 2) you can generally only look straight down.
That said, such a system would be useful for lots of things. For example, it would make missile defense a lot easier as you'd have real-time location and trajectory data. Missile defense is still a very hard problem, but you'd be a lot closer to something workable.
posted by jedicus at 8:36 PM on December 2, 2009
1. Sell spy satellite data to various countries to help them keep tabs on their enemies.
2. Wait a while and collect wealth to build a private army.
3. Start feeding false data to various countries to create the situation in which your army can take control.
posted by furiousxgeorge at 8:52 PM on December 2, 2009
2. Wait a while and collect wealth to build a private army.
3. Start feeding false data to various countries to create the situation in which your army can take control.
posted by furiousxgeorge at 8:52 PM on December 2, 2009
Actually, it doesn't even need to be false data. You are the De facto intelligence service for a lot of countries, just use that power to manipulate the situation however you need to by selective offerings of information to create your own storyline for your puppets to follow.
Also, remember not to explain your plan to your nemesis, just shoot him/her.
posted by furiousxgeorge at 8:57 PM on December 2, 2009
Also, remember not to explain your plan to your nemesis, just shoot him/her.
posted by furiousxgeorge at 8:57 PM on December 2, 2009
I think it should basically be a machine that can tell the future. This sounds a little too fantastical, but there are several computationally heavy disciplines which now claim to be able to make accurate predictions about the future. If you stuck close to the published literature you could make it very convincing.
So combining these disciplines together, the computer is able to predict the weather, election outcomes, foriegn policy outcomes and of course to economic trends (read any economist). The super villian could use these predictions to gain power and steer the course of history. If you stuck close to these published literature you could make it very convincing.
I think it would be interesting if the book got into themes about the limits of computaional prediction, things like black swans. The more I think about it the more I want to read this book.
posted by afu at 9:08 PM on December 2, 2009
So combining these disciplines together, the computer is able to predict the weather, election outcomes, foriegn policy outcomes and of course to economic trends (read any economist). The super villian could use these predictions to gain power and steer the course of history. If you stuck close to these published literature you could make it very convincing.
I think it would be interesting if the book got into themes about the limits of computaional prediction, things like black swans. The more I think about it the more I want to read this book.
posted by afu at 9:08 PM on December 2, 2009
Ooh! How about a simulation computer that takes the genetic data of...umm...1 million humans....puts the data through some sort of software that simulates....1 billion years of random evolution (environmental stimuli simulated i guess, random mutations simulated also) in ten thousand different directions....selects which pathways end up as the most viable and then creates embryos of those simulated future-people from stem cells? Said future-people then use psychic super powers to crush tiny humans.
posted by ian1977 at 9:13 PM on December 2, 2009
posted by ian1977 at 9:13 PM on December 2, 2009
If you're more concerned about creative forms of chaos that could be wreaked with computers and less concerned about whether they could actually be doable (even with computers orders of magnitude faster than what we have today, or the limits of computation) here are some ideas:
- Have the computer collect and analyze everything everyone does, in a similar way to how Google tracks a lot of personal information, and report it to the evil master
- Anything that was predicted to happen due to Y2K
- Have the computer understand text, speech and meaning, and conduct "man-in-the-middle" attacks on emails, voicemails, or phone conversations, changing the content in transit with funny and/or sinister effects. Change buy orders to sell orders, alter safety alerts to make the say 'all clear, change love notes to breakup notes, diplomatic conversations to declarations of war. Even if the computer is not this smart, maybe it just grants the villain access to all forms of communications and he has humans write or alter tham.
- Have it take control of anything that connects to a mechanical, real-world system - computers that control dams, air traffic, ships, traffic lights, nuclear missles (as in War Games), submarines, etc.
- Place subliminal banner ads throughout the web and start to control people's minds
- Many things that are implausible for a computer to do are easily doable with an army of cheap workers managed by a computer (think Amazon Mechanical Turk). What if this guy hired a million people armed only with smartphones in third world countries, paid them 50 cents a day, and had them aid him in the wreaking of havoc
- The computer could be a biological computer that can control an army of bacteria or larger creatures, or assemble creatures of its own (think of the video game Spore)
- The spy satellites could be like a super Google Street View, and be able to resolve down to the millimeter, tracking the movements of anyone who goes outside, anything they read or do. Everyone is afraid to go outside for fear of being tracked.
- The computer can send out massive amounts of spam, millions of times more than we have now, and with the ability to dynamically adjust to spam filters (like the shields of Borg ships do to counter photon torpedoes) It can also do this with all other forms of communication, so phones now ring every second with spam messages. Nobody can use electronic communications anymore.
posted by lsemel at 9:43 PM on December 2, 2009
- Have the computer collect and analyze everything everyone does, in a similar way to how Google tracks a lot of personal information, and report it to the evil master
- Anything that was predicted to happen due to Y2K
- Have the computer understand text, speech and meaning, and conduct "man-in-the-middle" attacks on emails, voicemails, or phone conversations, changing the content in transit with funny and/or sinister effects. Change buy orders to sell orders, alter safety alerts to make the say 'all clear, change love notes to breakup notes, diplomatic conversations to declarations of war. Even if the computer is not this smart, maybe it just grants the villain access to all forms of communications and he has humans write or alter tham.
- Have it take control of anything that connects to a mechanical, real-world system - computers that control dams, air traffic, ships, traffic lights, nuclear missles (as in War Games), submarines, etc.
- Place subliminal banner ads throughout the web and start to control people's minds
- Many things that are implausible for a computer to do are easily doable with an army of cheap workers managed by a computer (think Amazon Mechanical Turk). What if this guy hired a million people armed only with smartphones in third world countries, paid them 50 cents a day, and had them aid him in the wreaking of havoc
- The computer could be a biological computer that can control an army of bacteria or larger creatures, or assemble creatures of its own (think of the video game Spore)
- The spy satellites could be like a super Google Street View, and be able to resolve down to the millimeter, tracking the movements of anyone who goes outside, anything they read or do. Everyone is afraid to go outside for fear of being tracked.
- The computer can send out massive amounts of spam, millions of times more than we have now, and with the ability to dynamically adjust to spam filters (like the shields of Borg ships do to counter photon torpedoes) It can also do this with all other forms of communication, so phones now ring every second with spam messages. Nobody can use electronic communications anymore.
posted by lsemel at 9:43 PM on December 2, 2009
A supercomputer that can not only predict weather, but simulate how to manipulate it would be a very handy tool for a super-villain, esp. if that super-villain is Sean Connery.
posted by Blazecock Pileon at 10:29 PM on December 2, 2009
posted by Blazecock Pileon at 10:29 PM on December 2, 2009
Isemel is on the right track there, if you are the evil version of Google you can cause all sorts of mischief, not sure you could take over the world unless you took it way past that to a total data monopoly though, but for a sci-fi story that isn't too far fetched.
posted by furiousxgeorge at 10:32 PM on December 2, 2009
posted by furiousxgeorge at 10:32 PM on December 2, 2009
Accurate protein folding could allow a supervillain to design new drugs & possibly entirely new lifeforms. If the computer itself was partially biological in nature, it clearly would be capable of synthesizing any of this. So I would have my biocomputer create some new fungus that threatened the world's food supply and have my shell companies ready to sell the only effective fungicide, at reasonable rates, of course.
This is so obvious, I'm surprised I have to explain my plans to you simple monkeys.
posted by chairface at 10:37 PM on December 2, 2009
This is so obvious, I'm surprised I have to explain my plans to you simple monkeys.
posted by chairface at 10:37 PM on December 2, 2009
Greg Egan's short story Crystal Nights is kind of relevant. (You can read it online. [pdf]) Create a virtual world with yourself as god, evolve intelligent life that goes through cultural evolution a million times faster than we do, and when they're sufficiently advanced, assign them to solve all your problems for you.
I remember a much older sci-fi short which had pretty much the same idea only without the computer simulation: mad scientist evolves creatures on an isolated island.
posted by moonmilk at 10:42 PM on December 2, 2009
I remember a much older sci-fi short which had pretty much the same idea only without the computer simulation: mad scientist evolves creatures on an isolated island.
posted by moonmilk at 10:42 PM on December 2, 2009
Many interesting problems that are too complex to solve with one computer can be solved in parallel by a thousand computers. So it's not worth considering those. Perhaps there's some way that a lot of CPU power could help you beat the stock market, but the guys who're making money on quick trades profit primarily (from what I understand) by having low network latency to the market. You either need to have a computer that's absurdly faster (on the order of 10^100x faster) to really make a mess of things (e.g., cracking root CAs' private keys - but know that they'd just create keys twice as long and you'd be back where you started), have a computer that can solve NP-hard problems quickly, or figure out some non-parallelizable problem that you can solve quickly (IIRC min-cut max-flow is hard to parallelize, and you could write a pretty spiffy inventory management system with this computer, but the potential audience for this would number in the single digits).
Either it's going to be very, very nerdy, or you're going to have to defile computer science. I hope you choose the former.
posted by jewzilla at 11:46 PM on December 2, 2009
Either it's going to be very, very nerdy, or you're going to have to defile computer science. I hope you choose the former.
posted by jewzilla at 11:46 PM on December 2, 2009
Have it send a crapload of pings to the CIA supercomputer?
posted by paultopia at 11:54 PM on December 2, 2009
posted by paultopia at 11:54 PM on December 2, 2009
Think about this:
China has a great firewall, yes? Imagine what would happen if you created a virus that attacked it from CIA IP addresses, and then did the same with Chinese IP's against US government websites and systems. You have this computer that can handle all this incoming data and this supervirus that spreads via antivirus software; I think you'd have an infowar started in no time. Now, with your supercomputer that can analyze operating systems and spit out new viruses that attack the same weaknesses with 1000 different viruses, you're stealing identities and testing credit card numbers and filing fraudulent applications with valid ID numbers faster than governments can replace and verify them. Keep poking and prodding and trying to start infowars and destroy commerce, and you've at least got a situation where you'll paralyze the flow of information for a few months.
Chaos!
posted by saysthis at 2:15 AM on December 3, 2009
China has a great firewall, yes? Imagine what would happen if you created a virus that attacked it from CIA IP addresses, and then did the same with Chinese IP's against US government websites and systems. You have this computer that can handle all this incoming data and this supervirus that spreads via antivirus software; I think you'd have an infowar started in no time. Now, with your supercomputer that can analyze operating systems and spit out new viruses that attack the same weaknesses with 1000 different viruses, you're stealing identities and testing credit card numbers and filing fraudulent applications with valid ID numbers faster than governments can replace and verify them. Keep poking and prodding and trying to start infowars and destroy commerce, and you've at least got a situation where you'll paralyze the flow of information for a few months.
Chaos!
posted by saysthis at 2:15 AM on December 3, 2009
Your super villain genius doesn't need a faster computer, he needs P=NP.
Then you're good on a large set of "unsolvable" computer problems. All of these problems could result in both extreme riches, control, or chaos.
posted by zpousman at 6:00 AM on December 3, 2009
Then you're good on a large set of "unsolvable" computer problems. All of these problems could result in both extreme riches, control, or chaos.
posted by zpousman at 6:00 AM on December 3, 2009
Start selling supercomputers. Put Intel, AMD, IBM, Microsoft, and Apple out of business. Then, once the whole world is dependent on your supercomputers, activate their secret backdoor and hold the whole world hostage.
posted by fings at 8:19 AM on December 3, 2009
posted by fings at 8:19 AM on December 3, 2009
Bruce Schneirer has this interesting post on key lengths and decryption. It gets technical (how long of a key length the entire output energy of a sun during a time period could decrypt) but it's well worth reading before having your supersupercomputer easily and arbitrarily break encryption. Unless the computer has an infinite energy sounce (hey, maybe it taps into the quantum vacuum zero point energy). But if it was able to do that then you'd be able to more impressive things than break codes. Like power the Earth forever even after the universe dies its heat death.
posted by 6550 at 12:07 PM on December 3, 2009
posted by 6550 at 12:07 PM on December 3, 2009
Making money on the stock market with micro-second transactions requires being really, really close to the stock market, since the time lag in data transmission is a much greater factor problem than processing power.
If you had the broadcast capacity, you could, say, reprogram all the satellites in the world to respond only to your computer's commands & disrupt global communications, military intelligence, and so forth.
Then you could broadcast whatever you want & spy on whomever you choose.
Again, if you have enough antennae to project your signal anywhere in the world, you could eavesdrop on global communications (secretly or not so secretly), and those remote control military planes, and so forth.
Digital Fortress... the new Die Hard movie both deal with digitized terrorist intent on disrupting society.
posted by MesoFilter at 1:03 PM on December 3, 2009
If you had the broadcast capacity, you could, say, reprogram all the satellites in the world to respond only to your computer's commands & disrupt global communications, military intelligence, and so forth.
Then you could broadcast whatever you want & spy on whomever you choose.
Again, if you have enough antennae to project your signal anywhere in the world, you could eavesdrop on global communications (secretly or not so secretly), and those remote control military planes, and so forth.
Digital Fortress... the new Die Hard movie both deal with digitized terrorist intent on disrupting society.
posted by MesoFilter at 1:03 PM on December 3, 2009
I can't think of anything cooler than finding the Ramsey number (5,5) or (6,6). The quote from Paul Erdos goes like this:
Imagine an alien force, vastly more powerful than us, landing on Earth and demanding the value of R(5,5) or they will destroy our planet. In that case, […], we should marshal all our computers and our mathematicians and attempt to find the value. Suppose, instead, that they ask for R(6,6). In that case, […], we should attempt to destroy the aliens.
I'm sure something like that could be worked into an evil plan by a good writer. Of course, I'm a Graph Theorist working in Ramsey Theory, so I'm a little biased.
posted by monkeymadness at 3:30 PM on December 3, 2009
Imagine an alien force, vastly more powerful than us, landing on Earth and demanding the value of R(5,5) or they will destroy our planet. In that case, […], we should marshal all our computers and our mathematicians and attempt to find the value. Suppose, instead, that they ask for R(6,6). In that case, […], we should attempt to destroy the aliens.
I'm sure something like that could be worked into an evil plan by a good writer. Of course, I'm a Graph Theorist working in Ramsey Theory, so I'm a little biased.
posted by monkeymadness at 3:30 PM on December 3, 2009
This thread is closed to new comments.
more info on this.
posted by dfriedman at 6:33 PM on December 2, 2009