The model is getting better technically, but worse experientially

[hannshumhanna]

I spent some time trying to understand where feedback should be posted.
Some channels are not very intuitive and I do not use Twitter or Reddit so I decided to write here.
Hope this is an ok place.

From my side the model feels like it is improving in one direction but getting worse in another.
Especially in the part that many people originally liked and connected with.

I understand this upgrade was likely focused on better code quality.
But for regular users like me the overall experience feels worse.

The answers themselves do not seem bad.
But the interaction feels much drier and more limited than before.

It also feels like some subtle but important things are slowly disappearing.
Hard to explain clearly but very easy to notice.
Those things made the dialogue feel more natural and engaging and now that feeling is weaker.

One more thing that feels important.
For many Plus users a much more valuable improvement would be stronger long term memory that works in a more personal way.
That feels more important than changes in tone or surface behavior.

Just wanted to share honest feedback from someone who really appreciates what you have built.

[vb]

You may be right. The official announcement opens with:

We are introducing GPT-5.2, the most capable model series yet for professional knowledge work.

That framing alone suggests why the general, day-to-day chat experience may feel different now.

As a ChatGPT user, you can choose a style and tone that better matches what you expect from a chat experience. I have not explored this feature much yet, but I plan to try the less personal variant, since that could align better with how I prefer to work with the model.

Hope this helps.

Ps. I have moved this topic to the ChatGPT category.