To run out of memory, in the language of computing, is to have too much of it and also not enough. Such is our current situation: we once again find ourselves in a crisis of memory, this time marked not by dearth but by surplus. Simply put, we are running out of space. There is no longer enough room to store all of our data, our terabytes of history, our ever-accumulating archival detritus. As I type, my computer labors to log and compress my words, to convert each letter into a byte, each byte into a hexadecimal “memory address.” This procedure is called “memory allocation,” a process of sifting, sorting, and erasing without which our devices would cease to function. For new bytes to be remembered, older ones must be “freed” — which is to say, emptied but not destroyed — so as to prevent what are called “memory leaks.” Leaks are to be avoided because, wherever they occur, blocks of precious computing memory are forever fated to remember the same stubborn information, and therefore rendered useless. For memory allocation to function smoothly, the start and finish of each memory block must be definitively marked. “In order to free memory, we need to keep better track of memory,” one developer advises. Operating systems, unlike the humans for whom they were designed, are built to tolerate little ambiguity about where memory begins and where it ought to end.
There's nothing Artificial about our Intelligence