I'm not going to argue over this one: x86 (and most other architectures) use two's complement for representing negative numbers. Under that system, you invert all the bits and add 1. It's just the way it is. Don't believe me? Go ahead and read
this.