patternMinor
Are all computer viruses man-made?
Viewed 0 times
madeallarevirusescomputerman
Problem
I'm curious if the origin of any successful computer virus is known to NOT have an intentional, human originator. To put it another way, do computer viruses ever "spontaneously" originate through the process of natural selection in the same way biological viruses originate; without an intentional "designer?"
I imagine (though I'm not sure) that it's possible for sufficiently knowledgable software engineer to make use of software engineering tools like neural networks, genetic algorithms, adversarial networks, etc to write a virus according to his/her specifications, but that's not what I want to know. I want to know if computer viruses ever begin on their own, and if they don't, why not?
The three core tenets of evolution by natural selection are descent with modification, variable fitness, and resource limitation. Some scientists say that wherever these conditions are in place, evolution occurs necessarily. Importantly, being alive is not a prerequisite, which is why biological viruses continue to evolve even though they are not technically alive. Richard Dawkins says that thoughts and ideas survive and proliferate through the same mechanism. So, does computer code ever evolve organically without human intervention to form a successfully self-replicating entity? If not, why not?
I imagine (though I'm not sure) that it's possible for sufficiently knowledgable software engineer to make use of software engineering tools like neural networks, genetic algorithms, adversarial networks, etc to write a virus according to his/her specifications, but that's not what I want to know. I want to know if computer viruses ever begin on their own, and if they don't, why not?
The three core tenets of evolution by natural selection are descent with modification, variable fitness, and resource limitation. Some scientists say that wherever these conditions are in place, evolution occurs necessarily. Importantly, being alive is not a prerequisite, which is why biological viruses continue to evolve even though they are not technically alive. Richard Dawkins says that thoughts and ideas survive and proliferate through the same mechanism. So, does computer code ever evolve organically without human intervention to form a successfully self-replicating entity? If not, why not?
Solution
Interesting question. I think the biggest barrier preventing something like that from happening is that computers are engineered to be extremely robust to random, non-human errors. Computers will do what programmers instruct them to do, and unless the hardware is being pushed beyond its limits or subjected to extreme environments like cosmic radiation in space, they do not deviate from programmer instruction.
It is necessary for computers to have these near-zero error rates due to the sheer quantity of operations they perform. An error rate of one in a million seems pretty good...until you consider that a modern desktop computer is capable of performing about 100 billion operations per second. A one in a million error rate would result in 100,000 errors per second, rendering your computer completely unusable due to constant crashes!
The same reasoning applies to storage mediums as well. If you assume that one in a million bits stored on a 1TB drive gets corrupted every day, then that would result in 360 megabytes of corrupted data per year!
The necessary low error rates are achieved through a combination of robust engineering and manufacturing, as well as software tricks to detect and repair errors such as checksums and error-correcting codes. In situations where these measures are insufficient, such as on spacecraft computers where cosmic radiation can wreak havoc, fully redundant computers check each others' results though voting.
Therefore I'd argue that "descent with modification" doesn't really exist here. Computers operate more or less deterministically and without error by design, so it would be very difficult for an unexpected program to spontaneously emerge.
If we designed computers that were much more error-prone, then it would be more likely that such an event could happen. But such a computer is also likely to be pretty useless, so we don't build computers like that.
It is necessary for computers to have these near-zero error rates due to the sheer quantity of operations they perform. An error rate of one in a million seems pretty good...until you consider that a modern desktop computer is capable of performing about 100 billion operations per second. A one in a million error rate would result in 100,000 errors per second, rendering your computer completely unusable due to constant crashes!
The same reasoning applies to storage mediums as well. If you assume that one in a million bits stored on a 1TB drive gets corrupted every day, then that would result in 360 megabytes of corrupted data per year!
The necessary low error rates are achieved through a combination of robust engineering and manufacturing, as well as software tricks to detect and repair errors such as checksums and error-correcting codes. In situations where these measures are insufficient, such as on spacecraft computers where cosmic radiation can wreak havoc, fully redundant computers check each others' results though voting.
Therefore I'd argue that "descent with modification" doesn't really exist here. Computers operate more or less deterministically and without error by design, so it would be very difficult for an unexpected program to spontaneously emerge.
If we designed computers that were much more error-prone, then it would be more likely that such an event could happen. But such a computer is also likely to be pretty useless, so we don't build computers like that.
Context
StackExchange Computer Science Q#123639, answer score: 6
Revisions (0)
No revisions yet.