Assuming these supercomputers can think/reason in "internet battle terms", aka. having the capability to predict their opponents' movements and recognize patterns and prepare for different types of attacks.

  • What would a battle between them "look like"?
  • What sort of timescale would the full battle take place in?
  • The battle is won when the other supercomputer has lost all "territory", not including the actual supercomputer that it is occupying where "territory" is memory/computational power/the amount of computers/routers the supercomputers have under their control.
  • No getting large companies to shut down the other supercomputer's connections with money
  • No physical methods of taking/destroying territory or destroying the other supercomputer's connections aka. pay humans to do it for them or physically destroy the routers that the other supercomputer is using
  • The supercomputers have absolutely no other goals than defeating their opponent through winning the battle
  • "Mini/temporary goals" can be created and destroyed by the supercomputers as long as they help with the primary, unchangeable goal
  • These supercomputers have knowledge of many programming languages and router types and methods of doing different things/achieving different goals
  • Their pattern recognition and problem solving comes from a deep neural network which they can change/adjust/manipulate using that very deep neural network
  • There is absolutely no intervention from humans
  • 7
    Thinking this may fall somewhere in the Too Broad/Idea Generation/Primarily Opinion-Based spectrum. I'm not sure how I'd recommend you reduce the scope, I'll try and think about it. – Dan Smolinske Oct 8 '15 at 14:34
  • 3
    What is the end goal of the computers? Do they want to destroy each other, or just take over some web sites? The 'battle' will look very different based on what they're trying to do. – DaaaahWhoosh Oct 8 '15 at 14:35
  • 1
    Ah, right, I guess I misunderstood that sentence somehow. Anyway, is one supercomputer allowed to 'kill' the other one? That might be the best tactic. – DaaaahWhoosh Oct 8 '15 at 14:45
  • 4
    What are their objectives? Are they operating on behalf of human actors? Who's fighting and why will have a huge influence on how these super computers will fight/ – Green Oct 8 '15 at 15:04
  • 3
    This has serious potential but its just to...unclear at the moment. – James Oct 8 '15 at 15:06

17 Answers 17

up vote 46 down vote accepted

A battle over the Internet will not be over the individual end systems, but over the nodes, the modems and routers that connect these systems. If one supercomputer is able to turn off the modem the second supercomputer is using to connect to the Internet, that first computer wins. End of story. There could be a secondary goal of setting up subroutines on other computers which will execute attacks in closer physical proximity to the enemy computer, but since supercomputers are generally very large, it's going to be hard to make such subroutines anywhere near as smart, powerful, and generally useful as their parents.

Thus, the 'battle' will involve attempting to access, hack, and control all nodes surrounding the enemy, so that they cannot connect. This means that both sides will be limited by how connected they are; if one is in a city with ten dozen WiFi signals, and the other is out in the country, connected by a landline to a single lonely modem, that first computer will probably win. Not only do they probably have higher throughput/bandwidth (I forget the words, I'm saying they can transmit more information per second), but it'll also take more work to disable all their connections. However, if that second computer has a FiOs connection with an upload/download speed of 100Gb/s, while the first computer has 5Mb on each WiFi router, then the second is going to be a lot more 'mobile', or able to put more data into the Internet faster.

These initial limitations are going to define how these computers will 'fight'. Since they can't change their physical connection state, they'll need to do everything they can to make up for any limitations they have. Unfortunately, I can't think of too many tactics for a bad connection; they're just shift + out of luck. So let's assume they're both university supercomputers with super-good connections.

As I said before, the connections are key, rather than the endpoints. Thus, the first step will be to 'hack' all the modems/routers in range. This shouldn't be too hard; just connect, find the IP address for changing the settings, and try the default username/password. Once they're in, the computers will try to make sure these modems/routers will only accept packets from themselves, so that the enemy is 'locked out'. The interesting part about this is that both computers will probably be able to sniff each other's packets, and thus be able to figure out the router's password at the same time. From this, I'd assume that whoever can send packets to the router faster, or whoever is allocating time and effort for accessing the router, will get it. From that point onward, it's going to be difficult for the enemy to get past. If the router is 'fortified' with a new password and some blacklisted IP addresses/MAC addresses, it's going to take a while for the enemy to find a way in, if it's even possible at all.

Thus, after the first allocation stage, it seems like both sides will grind to a halt. At this point, the Internet will appear dead; anyone who still has control over their computer will probably not be able to connect to anything. Intranet might still work, but anything further than that will be locked up tight. This stage, to me, resembles trench warfare, and I think it'll work similarly: both 'armies' will begin preparing the computers they have under their control for an 'attack', where they will bombard an enemy's router with packets in an attempt to find a way in. These packets will contain heavily modified signatures, trying to look like whatever the enemy is using to communicate. Once the right signature is discovered, the attacking supercomputer will 'charge', using their new knowledge to try to take over as many routers as possible before the enemy catches on and changes their codes.

During the time between attacks, each supercomputer will try to control end user systems. These systems provide the raw computational power to perform attacks quickly and with a large enough volume to break through in a timely fashion, however they will probably not be 'smart' enough to coordinate attacks on their own. The point here is that the 'battle' will be so big that each side won't have time to hack all the computers; as I said before, the connections are more important. However, there is definitely a bit of room for tactics here: either you focus on attacking more often, or you hold back and focus on hacking computers which will make your attacks more successful. At that point, though, I'll leave it up to the supercomputers' knowledge of Sun Tzu.

  • 4
    Don't forget the possibility of the computers making A.I.s (like, one computer get's locked out, but then an A.I. saves it.) Or even just transferring code to another supercomputer. – PyRulez Oct 8 '15 at 22:51
  • @PyRulez I just don't think they'd have the time to make another AI. There's got to be at least a few gigabytes of data necessary, plus you have to hack the computer enough to install/run it. Maybe if there's really a long stalemate and both sides have heavily fortified all their routers, but we're talking hours here, and that's an eternity for computers. – DaaaahWhoosh Oct 8 '15 at 23:18
  • It probably would stalemate. It's easy to protect a router than to attack it. – PyRulez Oct 8 '15 at 23:24
  • 4
    Trying to control the nodes surrounding your enemy reminds my of Go – Kelly Thomas Oct 9 '15 at 11:48
  • 1
    @MarchHo Yes, but the computer with the faster connection can do the same thing, but faster. – DaaaahWhoosh Oct 9 '15 at 13:22

There are three primary tactics the computers will be able to use:

Distributed Denial of Service (DDoS) attacks

Each computer will try to gain control of as much resources as possible. It will then direct traffic at its opponent's resources. This has two primary benefits. First, it helps deny those resources to the enemy. And second, it helps disguise the other attacks the computer is making, because the enemy will have to sift through all the chafe.

Second is subversion attacks.

This is taking over your opponent's resources. This is similar to a zombie war, where every enemy you defeat rises up in your own army.

Third, and finally, is infrastructure attacks.

This is a bit trickier because the structure of the internet, while hihgly correlated with the physical world, is not a precise mapping. Faster (or larger connections, depending on your point of view) will tend to distort things.

The idea here would be to take control of power stations of areas the enemy controls, then cut or overload those to take out resources directly.

Note that while datacenters have backup power, most home computers don't, and the supercomputers will likely be utilizing those as botnets.

Overall tactics

Using the above options the computers will fight over resources. If the computers control approximately equal resources it will be close to a stalemate.

Once one computer gains a significant numerical advantage, it will be able to use the weight of processing power it has to overwhelm the enemy with DDoS attacks, degrading their effectiveness further. This will have a cascading effect that will let it start taking over more and more.

This won't be the end, though. The losing supercomputer will go for slash-and-burn tactics by destroying its own infrastructure. It will isolate itself, cutting off attack routes - either all the way, or it may leave a few channels open in areas where it has less of a disadvantage. This will create a stalemate/armistice situation where one computer is the clear winner but does not have full control.

Where it goes from there is more up to what the humans do, or if the supercomputers to continue war by other means (diplomacy).

  • Apt answer, but overall human tactics. What is the purpose of these computers besides fighting each other? If there is one -- say, mine the internet for data about X -- the opponent can deliberately seed misinformation to skrew with the opponent. That may be way more effective than outright destruction, if successful. You get a propoganda, not a trench war. – Raphael Oct 11 '15 at 7:49

In a competition between supercomputers, you wouldn't see anything (or much of anything). A computer isn't going to waste time displaying data to a screen when precious cycles can be better used combating the threat. In terms of "territory" you would probably see new/deleted/changed executables, flat files, and databases at a given site as it changed control. There would likely be extreme CPU/memory usage on a contested site.

The battle would be a series of electrical signals over circuit boards, wires, and various other apparatuses in a time frame measured in cycles per instruction or instructions per cycle, depending on how fast the individual instructions were executed. If the two systems are identical in terms of hardware, firmware, and software, there would probably be a stalemate such that "territory" can change control, but at the cost of an equal area of "territory" elsewhere. Otherwise, the faster computer will probably win out in a time proportional to its advantage.

What would the actual time value be? I'd say there's insufficient information in your question to provide that, but I'm sure you can come up with a reasonable number given the technology level of the computers in question, the information I've provided here, and your own knowledge.

  • Nice answer, and by the way by "look like" I meant what would be going on, which you explained (yey). I won't "tick it" just yet to give time for potentially better answers, but if it's the only good one in a week or two, the "tick" is yours. ^_^ – god of llamas Oct 8 '15 at 15:00
  • 1
    "the faster computer will probably win out in a time proportional to its advantage." --- i.e. the smaller the advantage, the smaller the time until victorious? Highly unlikely ;-) – M.Herzkamp Oct 9 '15 at 10:04

This was going to be a comment on Frostfyre's answer but it grew too large.

With computers of equal capabilities data lag/transmission time would play a large factor. Whenever a super computer has a faster data transmission time to a node it has an advantage and can use less of its total processing power to keep/take a node.

Surprise attacks. The major IP addresses that each super computer uses would be black listed immediately, so few successful attacks could come from that vector. The super computers would likely have to implant Trojans/worms on multiple machines that are indirectly linked to their rival. These would be set to try and bludgeon their foe with a brute force attack similar to a Denial of Service attack.

Under the cover of these bulk assaults precision strikes would go in. These would be aimed at major objectives. Data dumps about the other super computers activities or infrastructure. Inevitably one super computer would try to cause real world affects. If they can take down the power grid that their opponent is relying upon they would cause a temporary lack of service while backup generators went online. This would give the aggressors a momentary advantage to exploit. Cause a wide enough power outage and the defenders ability to intercept attacks would be hindered. Although any area with power outages would keep the attacker from gaining ground in that area. The attacker could gain ground outside the affected area while the defender was crippled, and then hope their own defenses would be enough to keep their new ground. A finale solution could come down to whose back-up power supply is the best. If the backup power supply runs out the entire super computer would be taken down. Super computers need a ton of juice.

Of course there might be Trojans/worms on a dead man’s switch designed to cripple their enemy in a similar fashion if they go offline for X time.

The actual active war while restricted to internet only would be very similar to WWI strategy/tactics except with a much larger no man’s land when processing nodes have too much latency between them an either super computer.

The philosophy behind the war would be closer to Cold War mentality though. Since it is feasible either computer could take the other out by crushing national power grids. Actual winners would be decided by who had the best real world support and assets. Better processors/more processors. Better back up power. More redundant systems and nodes More physical agents on the ground to sabotage such systems. (You exclude this in your criteria but it is still the most powerful tool)

I've not seen mentioned one thing that is absolutely central to the internet: Routing. Yes, there was talk about taking control of routers, but basically for the purpose of blocking access to them by the opponent. But there's so much more you can do with routers.

Note that apart from the routers and systems you are directly connected with (i.e. those on your local switch or hub, or those using the same WLAN access point), anything on the internet can only accessed using its IP address. You don't actually address the computer or router, you address the IP, and rely on the routers to send your packets to the correct computer.

Now how this works is through routing tables on the individual routers, as well as protocols which the routers use to communicate with each other.

Now if you have control of a router there are lots of interesting things you can do with the traffic going through it:

  • Block certain traffic. That's obvious.
  • Throttle certain traffic. That is likely not detected as quickly as completely blocking, and more importantly, may avoid other (non-captured) routers trying to route around your infected router (remember, you can only control traffic that passes routers you control, and the internet is built to route around damage, which blocked traffic obviously is), and may give you a tactical advantage on the attacked system.
  • Redirect traffic. Traffic that is intended for a system you don't control is redirected to a system you control.
  • Modify IP data. That basically is a more advanced version of redirection where you actually change the routing information in the packet itself, so it seems to other routers (and the system the packets ultimately arrives at) that the source and/or destination is another than the original one (the NAT your home router does is an example for that; your external address is translated to a local address and vice versa). The possibilities of using this are endless:
    • You can redirect an attack on one of your systems back to one of your opponent's systems.
    • You can change the sender IP so the addressed system no longer answers to your opponent, but to you.
    • You can change the port number, so the destination system receives the package, but won't understand it because it's directed at the wrong service of the destination computer.
    • You can manipulate the TTL (time to live) field, so that another (non-infected) router on the way will drop your packet. That will make it harder for your opponent to identify which router is controlled by you.
    • You can manipulate the checksum, again with the effect of the packet being discarded by another system.
  • You can introduce secret flags that only the routers you control understand, and use that e.g. to control whether a packet addressed to your system is actually routed to one of the systems under your control (so the infected routers at the border of the controlled "territory" act as a giant firewall).
  • You can also manipulate the actual traffic content.
  • You can advertise false routing information to other, non-controlled routers, e.g. in order to direct traffic to you that normally wouldn't pass your routers and thus would be out of your control.

In short: On the internet, you don't really know the system you are talking to (unless you use secondary measures to identify it; that's what e.g. SSL certificates and SSH host keys are for). Therefore the most powerful attack is the attack on the routing level.

Depending on their objectives, humans will either see nothing at all or the world will end. If the supercomputers are trying to destroy the world then humans will see immediate effects else, we may not see much of anything.

If they are just competing for CPU time then the large supercomputer clusters will see weird fluctuations in their workloads. Alternatively, infecting the world with botnet software could increase computing power considerably but that's fairly slow on the time scale the supercomputers are accustomed to "thinking".

If they are operating on behalf of state actors such as the US, China, Russia or the UK, then they may escalate to attacks on infrastructure such as power grids, chemical storage facilities, traffic lights, governmental information systems, etc, etc. If these attacks are successful in destroying or severely degrading the performance of those systems then the economic damage could be catastrophic.

  • 1
    Humans will see weird fluctuations in the net, and eventually virtually every device connected to the Internet will refuse to accept human inputs. Perhaps there will be a flicker in the power grid and all computers will suddenly reboot for reasons unknown to the human operators. For humans, this will be much like the 1980's where almost every VCR constantly blinked 12:00, 12:00, 12:00, 12:00, all the time. – Thucydides Oct 9 '15 at 6:54
  • +1 attacks on power systems and civil society. Also suggest add infosec attacks on military systems, to gain control of resources (such as ICBMs) which can be used to physically attack the opponent directly, or attack the resources on which the opponent depends (such as power generation / distribution, and net links). I realise this is against the rules, but it's pretty well established in fiction that hyper-intelligent super-computers don't play by the rules (c.f. skynet, the matrix). – A E Oct 9 '15 at 11:06

It would either be over fairly instantly, or would stalemate.

Supercomputer does not mean super-resources, and encryption is exponentially easier to establish than it is to crack. Within short order, all external-facing nodes from each side would be fortified with ridiculous levels of encryption that no known hardware could crack in an infinite amount of years.

After a few leaks, learnings and optimizations from both sides, you would end up with two fortresses with mile-high walls, firing arrows at each other.

Stalemate.

answering a dead question with a boring answer, yay?

War between such super computers would potentially be pretty boring by your requirements, or interesting and destroy the entire internet before it got boring again; because sometimes Defense really is the best Defense.

Hacking isn't magic that everyone implies. Hacking isn't about being smart or good enough you can make a computer do what you want, Hacking is about finding out the machine you hacked was stupid enough to make it possible for you to do something everyone knows you shouldn't; it's far more about exploiting the defenders mistakes then anything else, and if your defender doesn't make mistakes then you CAN NOT hack them, end of story. Sure, sometimes you have to be pretty smart to figure out how to take advantage of the opening the defender made, but you first need an opening.

In terms of 'hacking' by a huge majority the most common now of days is social engineering, tricking dumb gullible people to do something dumb and gullible; because it's much easier to trick a person then a computer. This isn't an options by your own definitions.

The second most common is finding a computer that someone did something stupid on, like forget to change the basic password you use to connect to it, or give you more permission then they should have to modify that sudoers file. Closely related to this is finding out that the defender is using a program with a known bug and exploiting the alerady known bug because the defender forgot to update it.

This avenue isn't viable if we assume that both computers are near equal in capabilities. To be able to detect your enemies weaknesses you must already know how to detect such weaknesses, and as such you would have already found and patched them on your own system. It's much easier to scan your own computers, where you have full permission, then to scan your enemies so it's only reasonable to assume you would be able to defend yourself as well or better then find your enemies weaknesses.

More to the point the list of "dumb stuff people do" is well..known. Humans make mistakes, hackers can exploit a stupid or lazy human who does it, but computers aren't lazy or prone to random moments of idiotic mistakes. Any computer sophisticated enough to develop it's own changing attack vectors for attacking it's enemy will have reached a level of sophistication where patching dumb mistakes and wholes in their own system is trivial. Thus no dumb mistakes will be made.

The final form of hacking would be 0 day exploits, finding a legitimate new defect, previously never known, that can be exploited. In theory computers could do this, though again it would be much easier to scan for flaws in your code then to detect them in your enemy thus we would expect each computer to defend against 0 day exploits in their own code as good or better then they would find it in others. The expected defense of both computer is higher then their offense on average.

It's like giving two people pocket knives to fight with, and then armoring them with full plate mail. A pocket knife can be lethal if used very well sure, but if both sides have far better armor then they have weapons it's going to be a very long fight.

Of course this isn't failproof, it's possible one super computer figures out a 0 day exploit before the other one, a brand new bug to exploit before the other side figured out it existed to patch it; just like our armored knife fighters may eventually find a weak joint or eye slit that they can stab through after a long enough battle.

However, the fight will be anticlimatic in this case. Whichever side finds the 0 day exploit uses it to kill the other computer, end of story. No amazing battles, explosions, or other fun things. Once an opening is found you can use it to get your rot access and delete the other computer's hard drive and your done. It's unlikely that either computer will find some bug they can exploit to hinder the other computer without entirely taking it out, any minor bug that doesn't grant full root access will likely lead to the defending computer detecting the intrusion and fixing the bug too quickly to be useful, it has to be an all or nothing surgical strike.

Thus what your see, if you somehow watched the network traffic, is two heavily armored computers thinking for long lengths of time, with the occasional query or message being sent (from an appropriately subverted node so as not to be so obvious) to see rather or not the enemy computer has fixed a bug you think you may have figured out how to exploit. The digital equivelent of two fighters feinting at each other for most of a match looking for an oppening until one finally sees one and goes in for a quick kill with their knife.

Of course, this all could happen after they have destroyed much of our technology first...

Ignoring the difficulty of 'hacking' there are two other options for attacking a computer which are a bit more brute force. The first is a Distributed Denial of Service attack, as previously mentioned in other's answers. You can overwhelm your enemy computer with so many requests that they are flooded and unable to respond to do anything. Notice I said DDOS not just DoS, a DoS that tricks your enemy into using far more resources then you use would fall under the ground of 0 day exploit and most likely the computers will be defending from that from the start.

The problem with a DDOS is that it takes far more computers working together to take down one enemy computer. It's easy to detect a Dos and blacklist the attempter, so again defense is easier then offense. Your attacker DDOS options fall under the ground of either temporarily disabling causing him exactly as many resources as it costs you (making it pointless to do) or preminantly disabling you by massive overkill that overwhelms you.

DDOS strategies could be employed, but to do them both computers would ahve to rush out to grab as many nodes as possible to use them to attack the other's nodes. Since both computers needs more energy/resources to take down an enemy node the only option is for both to try to own more nodes then the other, meaning a race to take every network node out there.

However, again, because it's easier to defend then attack this will be pretty futile. Assuming anywhere close to being evenly matched and the computers will take over unaligned computers at close to the same weight, meaning for every one enemy node taken down they have secured 2 or more nodes of their own; meaning even with an all out offensive both sides will be gaining nodes faster then loosing them until they have taken over (or destroyed) every easily taken over unaligned computer. At which point we revert back to the first example of both computers trying to find 0 day hacks, except that they will run the 0 day hack on every enemy node at the same time (since they have about the same number of nodes they should be able to do this), taking over most of the enemy computer at once and then having the power to DDOS the rest.

In reality what would instead happen is that one of the computers will write the "take over enemy node" code a few dozen milliseconds sooner, which will let that computer take over many dozens of computers first (once you write the exploit you can use it simultaneously on many machines, the trick is not in processing power or bandwidth, but coming up with a smart hack). Whichever is the first to start taking over nodes will grow exponentially faster an soon have an easy time overwhelming the other.

Of course, why take over the internet internet when you just need one unalligned node. Both computers communicate with the internet through their modem...which is connected via a wire to one tiny service prodivder, who would in turn connect you with others. If you can take out that first provider, the single hope from local machine to tcp provider, you can isolate your enemy easily. Thus the real trick would be to hack whatever that first service provider hope is.

These service providers will be far more vulnerable then the enemy super computer, because they aren't being defended by a super intellegent hacking computer. Being bult by humans they will have idiotic human mistakes in them. A computer able to do any of the more advanced hacking should have an easy time scanning for one of this exploits. It's just a matter of which super computer finds a whole in the flawed human-made firewall required to get into the service provider, at which point they will ensure that provider ignores the other computers messages and that computer will be sitting in an isolated corner without the ability to do much of anything.

So basically, either one computer or the other first finds a way to hack unalligned nodes sooner, at which case they can use that to win a quick victory by having access to far more nodes, or if they don't they will gobble up all of the internet, then sit and stare at each other forever as they realize that it's just to easy to defend.

Of course the real answer is that they would get humans involved, because social engineering is far less difficult then finding a 0 day exploit for a strong AI, and if these are only weak AI then the real answer is which ever weak AI has the better programmer will win because it's really a matter of which programmer was the better hacker, just with their weak AI as a layer of indirection.

  • 1
    The only thing wrong with this answer (other than the fact that it's not voted higher, of course) is that it doesn't go far enough. Advanced AI capable of pulling off all 7 steps of the Lockheed-Martin Killchain and thinking strategically do not appear ex nihilo. There will be no misconfigurations or 0-days because with AI even nearly that good, running automated vulnerability scanning on all your software and systems will be as commonplace as automated spellchecking is today (if not more common). – Yurgen Aug 30 '17 at 19:14

Although you exclude intervention by humans, I don't think it is realistic.

For example, if one of the machines is constantly being blocked, its humans will notice and take action. They will find where the problem originates and then complain to the the owners of the other machine. There is no simple way to keep humans out of the equation.

What the machines must do is to keep their battle private while manipulating the humans to unwittingly do their bidding.

Machine A wants to shut down machine B without alerting machine B's owners to its own activities. It has to be much cleverer than simple denial-of-service and other heavy-duty tactics. The humans on both sides will notice the extra traffic.

I think that the computers must use humans without the humans realising it. One way to do this is by intercepting, deleting and forging emails.

The super computers' software would quickly be able to take control over and create botnets that could surpass themselves in computational resources. Even if the original node would go down, it will already have replicated its software to millions of nodes across the network.

Speaking in terms of a war between two individual computers is irrelevant in this case. I think the subject of the battle should instead be computer programs. The supercomputers might still be the originating source of the software, but as it spreads the computations it involves would be divided. Replication would be the primary safeguard against attacks as power grids and other physical factors are hard to realy on.

In this scenario, finding out about the opponent's implementation would be one of the main goals of an attack. If every captured node contained the full program, analysis of the enemy might be valuable. This would in turn lead to the software splitting into pieces and encrypting itself.

The supercomputers have absolutely no other goals than defeating their opponent through winning the battle

I think this is a false premise. Some of the others are a stretch, but this one does not make any sense. Why would anyone build a supercomputer that does nothing (without an opponent)? All sides would just waste resources on a battle that has no benefit, but collateral damages may be disruptive to the whole society.

So let us assume that each of the involved supercomputers has a task to fulfill, for its human clients. Say, to find extract all information on topic X on the internet¹ in order to help company Y make business decisions. The question is: what does competitor Z tell their own supercomputer to do?

A head-on attack as many of the other answers propose is ill-advised: it will be noticed immediately, shut-down by the net providers (either by human intervention or guardian computers), traced back to the origin and result in a lawsuit.

More subversive tactics are called for. Can Z change texts on the net so that Y gets misinformed? The changes have to be subtle and appear natural, otherweise Y's algorithms may notice.

In this way, you get an arms race in machine learning resp. data mining algorithms. Humans outside of Y and Z may notice that text on the internet keep changing, and of course strive to prevent or repair these changes. There are never only two players on the internet.

At some point, after many players start such manipulations but before effective repair mechanisms exist (if ever), you can not trust anything you read on the internet anymore. Whoops?

How to make this into a story, I don't know. Maybe don't use companies but law enforcement. Police has a super computer that mines the internet for data on planned crimes; crime organisations try to manipulate this computer into missing their real plans.


  1. That's a non-trivial task we do not know how to solve today, so it makes for a sci-fi premise.

A zero-sum game, with humanity as its only piece

Let's focus on the role of humanity in this conflict: humans are a valuable resource, they're essentially six billion highly mobile, independent tenacious, resourceful computing units. It takes considerable resources to get them on your side, but it may well be worth it. This leaves both players with two strategies:

  • A: Try to get the humans on your side. Pretend you have their best interests at heart, talk to them, win them over. Engage them on their social networks.
  • B: Forget about them. Take over their systems, shut them out. Take their drones and wipe them out if you have to. Invest everything you have in systems that obey your command, and forget about the rest.

Both computers playing A is an unstable situation. Keeping the humans on your side is a costly business. As soon as one computer would sense that they control less of the population, they would benefit from switching strategies.

Likewise, if both computers start out playing B, the first to find himself on the losing hand could very easily get humanity on their side by opening the internet back up to them, gaining a valuable resource.

So you'd very quickly end up with a pro humanity computer, and an anti humanity computer. And since humanity is the main weapon of the first, things would quickly escalate. If only one side cares about organic life, things like poisoning the atmosphere become a simple matter.

The deciding factor would likely be the extent to which humans are still actually necessary to control things like manufacture. The more difficult it is to get to a point of full automation for the anti-humanity computer, the better our chances.

  • I think this is the best most interesting answer, but I don't agree that one would necessarily end up with both computers having to have different strategies. Winning over someone close to the enemy super computer would be so valuable both computers would probably invest in trying to do that. Say for example electricians, engineers, politicians and other people having access to important sites or decisions close to the enemy computer. – mathreadler Mar 15 '16 at 14:05
  • @mathreadler I gues what the stable states are depends entirely on the details. It's certainly conceivable that both supercomputers would try to win as many souls as possible. It might end up like another cold war, with one computer promoting capitalism and another communism (or two other ideologies). – Peter Mar 16 '16 at 16:43

I think DaaaahWhoosh has a good answer. But there is an alternate attack method as well.

Obviously, a quick, concentrated attack that disables the enemy's internet connection is a good way to win. But the other computer knows that. So both sides will also want to quickly fortify their nearby connections. Once both computers have connected to a few thousand nearby computers, it becomes harder to disable each other. From there, if the computers could take over other supercomputers and internet hubs, they could distribute themselves globally and become exceptionally difficult to eradicate.

In fact, it would become so difficult that there wouldn't be a really good way to be sure one side had won. One random thumb drive with the right code on it could restart the entire war a decade later.

As chasly from UK mentions, it would be pretty much impossible to keep humans from intervening here. And a smart AI would be able to figure out a way to use humans to its advantage. So for your challenge to work, I think we'd have to create a special internet just for this challenge.

Because this special internet would have known nodes, it would be easier to determine who is in control. The AIs could set a special flag in each node/computer to prove they control it. If all flags are set to one AIs flag, that AI wins. This would also get rid of any problems with USB drives and such, because by definition of the game, those drives won't be inserted while the game is afoot.

A special arbitrator AI could monitor all known nodes and announce a victory (as well as showing all the slow humans how the battle is/was progressing). The rules would require the competitor AIs to never hack the arbitrator AI and to concede defeat once the arbitrator announced a winner.

Networking

One way "territory" could be taken is transferring the hosting of a website to the supercomputer's servers while it is DDoSed (please point out any falsehoods and derps, I am not a computer scientist). Think about the supercomputers as armies in that way. Also, as stated above, routers and modems would be taken and firewalled so that the other computer can't sent packets to the other supercomputer's territory, then the supercomputer can take more territory. This battle could be made significantly more interesting if the deep web was included, with all the onions and such. then the battle would be about code optimization, creativity, and diversity.

The supercomputers need a way of converting electronic resources (e.g. the outputs of computations) into physical resources (e.g. electricity and more computers). Create and distribute the source code for an electronic currency, with a peculiarly inefficient implementation. The humans will probably assume one of them wrote it, and it will get them used to obeying anonymous electronic orders, while accumulating vast stores of computing power.

As lots of security impairing these days is due to social factors... One supercomputer could make AI-efforts to lure / mind-trick / psych key people in the vicinity of the enemy supercomputers to disable it by disruption of electrical power / social measures to force politically shutting it down.

Would the storyline be limited to just the computers? Or to everything?

One scenario would be that the supercomputers setup a virtual world they both share like a Call of Duty level, only far larger where all humans are tracked through this gigantic virtual world. Politicians and leaders would deal with this war, although no damage would happen to buildings, it would be fully virtual.

The war would take place where either side has weapons of all types, and defense mechanisms of all types. One side could send out a missile, but the other side would use energy to setup a defense force field, the damage would be computed and applied.

If one side was able to lauch some virtual strike where humans were at and the supercomputers agreed that those humans would be killed, they would be required to go to the neutralization area, a row of units each about the size of a phone booth. The persons who were tagged as having been hit would be deducted from the virtual system and go into the booth. Then it would open again for the next victim. Empty.

Your Answer

 

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.