So if the LLM doesn't feel like doing its research properly then questions won't be accurately answered? And the inbuilt bias of the AI will never be challenged because all of the references it chose to include will check out? Its blind spots becoming our blind spots? It's idea of a criminal becoming our idea of a criminal?
Well I recommend mythTV. It has been a while since I used it but it was pretty good. I did get it working with a USB tv tuner I once had but that did require building kernel modules manually.
So true