Hi HN! I've been working on a series of small experiments that use LLMs under the hood.

For the first one, I made a typical chatbot UI but with a twist: you can enable a mode that writes the user interactions on your behalf. Peak laziness!