It's no different than what the internet has been doing for us for decades. People tell us commands to run, we use our best judgement, maybe check a couple things, and then run the commands. If the internet suggests a command or a LLM suggests a command, what's the difference?