[Bug] Model outputs chatty narration/event logs instead of requested content #109195

Bug Description
This model is compeltely enshittified, keeps writing chatty commentry style writing. Logs events, instead of writing platform.

Environment Info

• Platform: linux
• Terminal: gnome-terminal
• Version: 2.1.322
• Feedback ID: 59054fd7-a604-42f9-ae00-921fb7f457c2

I'm tired of fighting with it's garbage shit tier english writing. it eats up tokens, and I need to delete its sloppy writing all the time.

I'm tired of seeing AI slop writing. I, therefore, earnestly request everyone to please, please, please be considerate about what you're writing.

It creates huge cognitive loads on readers like me with ADHD. Keep it simple, genius. Using an LLM is okay for writing, but not thinking about how you can make it better and understandable for readers hurts credibility.

Tell your agents not to write these.

The "This is no X. This is Y. Why it matters" pattern. Please avoid hollow sentences that don't depict anything and can't spark curiosity in the reader's mind.
Aggressive, chatty, rhetorical, and dramatic writing. Instead, write like a warm and helpful professor with patience who teaches every concept with care and in easy-to-understand English.
Avoid cricket commentary–style writing. "This chapter writes, some more verb, measures and reports the same." Don't use verbs for nonliving things. A chapter can't write by itself. Instead, use "In this chapter, we'll examine and benchmark…"
Avoid overusing the article "the." "The this, the that." A human writer will creatively vary and not make the reading monotonous.
Please be concise and thoughtful with your writing. Seek feedback from native English speakers and professors. Clarity, precision, and respecting readers' cognitive load are underrated skills.