alt Hacker NewsI dislike one-shot LLM prose as much as the next person, but there's no chance this use case ever goes away. Writing is difficult and takes a long time. Using an LLM to write for you gets your message across with the least effort. It's here to stay. Our only hope is that the models simply start writing better.
Replies
randomblock1 • today at 4:36 PM
Sure, it's here to stay. But the least effort solution is often not the best one. I would also argue it gets the point across with the least effort on the writer's part, but the most on the reader's. LLM prose is not easy to read and often not interesting because we're all used to slop. And I like a little human touch. I don't know if LLMs will ever be able to replace the individual human style differences.
> Writing is difficult and takes a long time.
It gets a lot less difficult when you don't outsource it to a chat bot at every opportunity. Communication matters. If you do it badly because you let a bot that can't write do it all for you people should complain. The fact that lots of people will do it doesn't mean that readers should lower their standards.
I'm never going to complain about people using AI to write their private documents, but if someone publishes something, explicitly for others to read, and what they published is annoying to read and that wasn't the goal of the "writer" they should probably know about it. Even if they never check the comments, those of us who were made to read that slop can commiserate with each other.