Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I will often write a bunch of stuff and then use an LLM to pre-process it a little bit and suggest some improvements. I will then work through the differences and consider them individually and either accept them, or use them to write my own improvements. This is kind of like having an okay editor working for you. No substitute for a real editor, but it means that a: what I intended to say is preserved, b: there's no additional waffle (the prompt includes instructions not to expand on any topic, but only ever to summarise where possible), c: everything still goes by me in the end, and if it doesn't feel like something I would actually write then it doesn't get used.

I believe that it has improved my writing productivity somewhat, especially when I'm tired and not completely on the ball. Although I don't usually reach for this most of the time (e.g. not for this comment).



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: