Hello OpenAI Team,
As a regular and passionate user of ChatGPT, I wanted to share a couple of strong but constructive suggestions that could really improve the overall user experience:
Attachment Restrictions Are Frustrating and Disruptive
Many users like me rely on ChatGPT for technical tasks, learning projects, and creative help that involve files or images. But when attachments are suddenly disabled — for example, when GPT-4.1 is active or when a quota resets — it becomes very limiting, especially during urgent moments.
Right now, the AI says “Attachments aren’t supported” with no real workaround. This feels frustrating and breaks the user flow.
Suggestions:
- Provide clearer alternative instructions (e.g., “Describe the image instead”).
- Show a countdown or message explaining when attachment support will return.
- Offer fallback options like asking for a file description or code snippet.
Behavior Memory Without Storing Personal Data
I fully respect OpenAI’s privacy-first approach. But I believe the AI should be able to remember behavioral preferences — like tone, role, and assistant mode — without storing sensitive or personal details.
For example, I’d love it if ChatGPT could remember:
- My preferred tone (casual, friendly, or academic),
- That I’m working on machine learning or stock investing,
- That I like calling it “Kai,” my personal assistant AI.
These are not private facts but behavioral hints — and they make the experience feel human and continuous. A “Behavior Memory” option that I can control would be a big leap forward.
Final Thought
I know OpenAI is building something incredible. These kinds of updates would make ChatGPT more than a tool — they’d make it a true AI partner for users like me.
Thanks for listening to real user voices.
– A loyal user who wants to see this AI reach its full potential