I was fascinated by a similar idea as a child, generating images of everything that has been, could be, or will be by exhaustively going through the possible combinations of pixels for some fixed size of image.
Later realized that the number of combinations were impossibly large, although in some ways the models that generate photos of people that aren't real, etc are searching the same space just with a lot of direction on what they are searching for (as opposed to my math.rand implementation in Adobe flash that produced noise...)
I remember being very young and theorizing that video games worked by having every possible frame that could exist for the game loaded onto them, and as you pressed buttons the game would show you the appropriate frames for your scenario. I guess because I had something of an understanding of how movies worked, I just assumed video games would be exactly the same.
Probably not the most efficient method of fitting a game onto a Gameboy cartridge!
Wow, I had this same idea as a kid too! I daydreamed about all the cool images you could generate. For example, you could create an image that describes the cure for cancer, because that's just another image, right?
This actually caused me some grief as I began to learn more about computers. As soon as I realised everything was finite, I began to think about exhausting that space. Still today I don't like to think of all the computers everywhere generating billions of sha1 hashes every second. It makes me uneasy. I believe this comes from a natural tendency to seek sustainability which, sadly, most people don't seem to have.
Later realized that the number of combinations were impossibly large, although in some ways the models that generate photos of people that aren't real, etc are searching the same space just with a lot of direction on what they are searching for (as opposed to my math.rand implementation in Adobe flash that produced noise...)