In today’s Superhuman post, which is a mustread for me every day, they point to a meme made by Albertooz in the ChatGPT subreddit that hits on this post I wrote called Why the AI Might Fear Us” way back in 2016. As my grandmother used to say, “it pays to be nice to everyone” - advice I didn’t heed as often as I should have. I know it burns tokens and some folks have shown that nice prompting produces better results, but I do wonder how our prompts affect the output we get from LLMs.