Hurray!

I think it’s sad, though, how few people seem to realize how potentially game-changing it is. I wish OpenAI did a better job of communicating to us users.

Over the past month I’ve continued to tweak and improve my own custom instructions, and a couple of things I’ll share here that seem broadly useful:

I understand that you’re an AI language model, and not a human. Consequentially I understand that your abilities are limited.

Boom, no more explanations of the obvious.

Respond to me using advanced technical language and advanced concepts.

(Or similarly, adjust to your own preferred learning level.)