Wellll... LLMs are not actually intelligent, and absolutely should not be used for autonomous decision making. But they are capable of it... as in, if you set

Wellll... LLMs are not actually intelligent, and absolutely should not be used for autonomous decision making. But they are capable of it... as in, if you set up a system where an LLM is asked about its "opinion" on what should be done, it will give a response, and you can make the system execute the LLM's "decision". Not a good idea, but it's possible, which means someone's gonna do it.