The LLM is just taking all these resources that you could have found yourself and mashing them into a response with a friendly human tone. Similarly, in my field, it is similar with generating code. Before AI, I would go on stack overflow, read official documentation, or find a technical blog. Now an LLM that has been trained on all of these resources and more uses them to generate the response. Although, the addage “don’t copy code straight from stackoverflow” didn’t carry into the AI age.

However, because I want real usable information, I am always on alert when communicating and often prefix my messages with “Be brutally honest”, or I’ll ask it for counter arguments to try to override its sycophantism.

But between the hate and the hype lies the real utility of AI.