How does AI programming relate to computability and halting? I can't determine whether a program will halt either. But I can kill it if it doesn't halt and just

How does AI programming relate to computability and halting? I can't determine whether a program will halt either. But I can kill it if it doesn't halt and just try another program, because I can operate a stopwatch. Where does this recursive argument form? We can already see LLMs solving simple programmer tasks: creating software from an informal description.