Also, you only decompile to level of basic instructions that the processor understands. When you compile code to add two numbers, well, the processor only adds bytes. There are a quite a few steps that the compiler has to fill in.

Ok, all that is not a big deal. But then you deal with compiler optimization. Optimizing basically tells the compiler to take its time and find some clever ways to save machine steps. So now the “standard way” for a compiler to implement adding numbers may have other stuff rolled into it because the compiler may see an opportunity to save steps in a seemly unrelated calculation by inserting steps into the addition it is implementing. Now it’s basically unrecognizable. A human didn’t write, and wouldn’t have written that mess that the decompiler gives.