I am building an AI robot which would use LLM as it’s “Brain”.
With putting different prompt there, I can “pre-set” the style and the knowledge of it.
What kind of information should I put there?
Besides, I found that if I put too much in, it would sometimes confuse or miss it. So better not to put too much there.
I gathered some resources for now, though the context is a bit lacking…
Wow this is cool! You got cool insight on it! Let me try it on my robot!
In my experience, simple and focused instructions work best for AI. I usually define three key elements: the role, the constraints (like tone and safety rules), and the core objectives. Too much detail can confuse the model, so it’s better to provide extra context during the conversation instead of loading it all into the initial prompt.
For beginners, I’d start simple: describe what you want the AI to do, the tone/style, and any key knowledge it must have. Keep it short — a few clear instructions usually work better than a long wall of text.
What specific task is this AI robot for? Is this like a ‘chatbot’, or something else?
In general, I’d say be concise and focus in on what you want the AI to do.
Cool! I would try that!
Yes I think I need to keep it short.
It is kind of a robot that act just like your friends. It can response to you, and reply to your command, or it can decline your command, and move around itself as well.