While doomscrolling the internet these days and trying, mostly unsuccessfully, to keep up with the great AI boom, it has become incredibly obvious that many people really hate when you use AI to generate your own thoughts for you. There are also people who say never to use it at all, but the genie is out of the bottle as far as I am concerned. Reddit is full of this kind of hate and, ironically, it is somehow especially prevalent on the LLM subreddits.
These adventures led me to Mitchell Hashimoto, of HashiCorp fame, and Armin Ronacher, of Flask fame. Both of them have opinions I respect, and both have made many contributions to software in general. They also include AI disclosure notes in their written work. Mitchell’s is a simple note that his posts are written explicitly by hand, while Armin’s is an explicit declaration of his uses, including prompt sharing on a specific page.
Mine is below, borrowed a bit from both of them:
- I’m not a frontend engineer by any stretch; most of this site was generated with the help of GPT-5.5, including the styling.
- The main bodies of the blog posts are written entirely by me, by hand.
- Summarizing text and brainstorming remain awesome uses of LLMs, and are things I do quite often.