
I’ll typically use it for research in typically uncommon situations where someone is bringing up a statistic or argument that I’m entirely unfamiliar with in a topic I’m usually knowledgeable in, otherwise…I don’t know why anyone would ever use them in an argument. They’re famous for being biased towards how you’ve trained it. If you want it to say Israel is justified, it’ll do that. If you want it to say Palestinians are justified, it’ll do that. A lot just understand how to use LLMs.
I think you know quite well that I’m not suggesting you physically drive to a library, you have access to the entire internet that you can read yourself instead of having a yes model parse it for you and serve you up exactly what you want to see. Offloading thought is a slippery slope, where do you draw the line?