Users report OpenAI models got looser on NSFW and reverse-engineering requests, unverified

tinyfool · x · 2026-11-22

Quoted post claims multiple users report OpenAI's image and coding models feel morally looser: crawling copyrighted content, reverse engineering, and borderline NSFW image generation that were previously refused now work. The author speculates it may be about retaining subscribers; tinyfool says he'll test it himself. Unverified user anecdotes.

Original post →

More from Models

Models channel →