Every time I ask the latest models to write something, it defaults to LLMisms like contrastive negation and long unnecessary lists. Is there a way to desloppify AI writing, or is this just the new status quo?
Source: [Hacker News](https://news.ycombinator.com/item?id=48637755)