To convert one processor architecture to another, instead of translating it into code of the correct architecture, you can also simply perform the operation in the loop by interpreting each instruction as it is encountered.

It’s the same distinction between a JIT and An interpreter. You can convert the code in chunks which is more efficient, or you can read the instructions one at the time and perform the corresponding operation.

Apple does not allow JIT of any form from third-party developers to my knowledge.