Absolutely! Very well-articulated response. Thank you for that. Yes, essentially that is the 'goal' of LLMs -- to provide insights with as little need for 'prompting/coding' as necessary. Right now LLMs (even the best ones) make a lot of assumptions and respond with confidence with generic responses (OK, some are emerging that attempt to elicit context, it's a bit creepy in my opinion, but I understand why).

When we are presented with a question from a human, we rarely respond the same way. Much depends on who is asking the question and why. Even a question as simple as 'What's new?' requires context to be truly useful (you need to know when to answer 'not much' or 'the client is trying to change the scope of the project again' or when you just smile and say nothing walking away as quickly as you can). All answers are contextually dictated -- LLMs don't have that context at the beginning of the conversation (hence the 'Custom Instructions" in ChatGPT). A conversation with LLMs is like questions at a speed dating meetup (it's a real skill to elicit clues about whom to stay away from, ha!).

Geek alert: Notice how Star Trek conversations with 'the computer' differ in the original series and the Next Generation. Would you rather converse with Comdr. Data or Kirk's Enterprise computer?

Overall, that's where the 'holy grail' of LLMs resides, the seamless conversational experience. I'm already noticing how much easier some chatbots are to reason with than most humans. Perhaps the future of humanity can be saved by the New Algorithmic Age of Enlightenment...