A horizontal object serves as the basis for an autonomous generative system which uses an evolutionary algorithm to find a method for constructing programs that strive to survive in their environment and to combat – that is, crash – other programs. The work generates programs which initially consist of random instruction sequences, starts and tests them, discards the inoperative ones, and introduces mutations to produce new versions of the optimally working code.
Evolution is by nature unpredictable even though, if it unfolds in a hermetic environment, it should be at least “ontically isolated.” However, sundry metaphysical-anxiety-suffused theoretical constructs and even practical experiments have recently sought to falsify the thesis that the world is a simulation. All such theories and experimentation pursue the same goal of making a hole in reality. But is an ontic leak actually possible? If the computer simulation analogy were directly transposable onto reality, the answer would be “yes, it is.” Simulations limit resolution, effectiveness, and errors (Robert Tappan Morris caused such a leak by writing his famous “Worm” in 1988).
The piece is based on the programing language – a kind of a machinic pseudocode and an imaginative assembler – used in Core Wars, a programmer’s game developed by D. G. Jones and A. K. Dewdney in 1984 and, thus, passing for ancient in programming categories. The game consisted in writing programs which endeavored to “crash” each other, and was enveloped in a playful, “martial” terminology, with programs as “warriors” fighting on an “arena” (i.e., a separated block of computer memory). Alongside Darwin (its conceptual predecessor developed at the Bell Labs by Victor A. Vyssotsky, Robert Morris, and Douglas McIlroy in 1961), Core Wars inspired early computer bugs and a plethora of speculations on the possible evolutionary trajectories of autonomous software.