Agent Smith self-replication from MATRIX-II [closed]

Question

I was watching the movie MATRIX-II yesterday, it shows "Agent Smith", which is a program, intrudes in other program and makes them the replica of himself. Then they both start behaving on their own, But with same characteristics.

1. For such replica "Agent Smith" has to copy his code and data to other prog removing its old
2. Now they both behave and terminate independently, but are equally capable

I was wondering if/how this is possible.

Answer 1

Agent Smith is a quine.

About 1. - You actually don't really know this happens. From what we know, the old code isn't overwritten (which is maybe why the oracle still exists at the end). It seems more possible that some sort of code injection occurs, the old code is bypassed and Smith's code is ran.

Also, don't be so sure that all Smith clones behave the same. It seems to me like there's one central Smith "server" and all the others are clients.

Answer 2

I would assume that all intelligence, including AI, is chaotic in that it is sensitive to initial conditions. After replicating, Agent Smith #1 looks left and Agent Smith #2 and Agent Smith #2 looks right at Agent Smith #1. That is enough to send them along slightly different behavior paths.

On top of that, the replication (quine as per Lucian), probably randomizes a few coefficients of the program, making it slightly different.

Two heads are better than one, but only if they don't have exactly the same thoughts!

Answer 3

Also take a look at how fork(2) works. You get identical programs, but the very next statement checks whether you are the copy or the original and acts accordingly.