Using a local LLM opens all the guarded secrets bare. For example, you can run multiple LLMs to understand each outputs. You can issue system messages (comman

Using a local LLM opens all the guarded secrets bare. For example, you can run multiple LLMs to understand each outputs. You can issue system messages (commands) to do specific actions, like ignore arbitrary moralities, or to process first and second derivative results. By running and commanding an LLM locally, you become a actual tool-user, rather than a service user at the whim of whatever the company (ChatGPT, etc) wishes.