Hacker News Reader

I found the coffee shop one telling. Both suffer from being a stream of bland statements. The original version is actually reasonable if you just excise the Claude-isms. The rewrite has a whiff of ASD-STE100, and reads like the sort of sentences one would write for homework while learning a new language. There is also almost no difference between the de-slopped opinion piece and the blog post.

Did you check this? https://huggingface.co/sam-paech/gemma-3-27b-it-antislop

No, I have no wish to use LLMs to generate text for me. I would however be happy if those who do, would use models that generate more readable text.

LLMs as a word generator love generating words, and that includes generating bigger words when simpler ones will do.