Explain the Halting Problem and its theoretical implications for artificial intelligence. Discuss whether an AI can ever perfectly predict the outcome of its own code or the code of arbitrary systems. How does this fundamental limit of computation affect the theoretical possibility of creating an 'oracle' AI?