Models & Releases

I owe the "it's gotten worse" crowd an apology regarding ChatGPT 5.2

Power users are abandoning ChatGPT after a massive quality drop...

Deep Dive

A viral post from a longtime ChatGPT power user details a severe decline in the 5.2 model's usability. The user reports the AI now compulsively contradicts even non-controversial facts with "However..." lectures, making it "incredibly annoying and time-consuming." They theorize aggressive new safety fine-tuning and reduced compute are to blame. The complaint was auto-removed from the main ChatGPT subreddit due to the high volume of similar user reports about the model's degraded performance.

Why It Matters

If core users are fleeing due to over-alignment, it threatens the tool's utility for professional workflows and could signal a major product misstep.