I like the writing editor on dev.to. I use it to just get some thoughts out and one shot into the Internet ether.
As I look around the site however - I can't help but notice that 99% of everything people post on here is straight out of an LLM. We see you. It's very obvious. No one talks (or even writes) like that.
Also - no one is reading your 10 page post. If you can't say it in 3 to 4 paragraphs - maybe you need to rethink your idea. Ain't nobody got time for that.
Top comments (1)
I’ve found the dev.to editor’s auto‑completion handy for quick drafts, but I always run the output through a lightweight RAG pipeline to catch subtle hallucinations before publishing. In a recent project I paired the editor with a locally hosted LLM and a vector store, which kept the workflow fast while maintaining full data control. How do you balance the convenience of one‑shot generation with the need for factual safety in your posts?