You run uncensored local models where you can ask questions that would get denied by public providers, or questions that you prefer them not to know the intricate details (like your financial planning)
There's value in knowing what capabilities local models will have on regular hardware in the near future.