Hi guys. I really wanted to suggest one feature that I think would make ChatGPT much better and safer for a lot of people. Right now we can protect the whole app with Face ID or a password, but I think it would be amazing if we could also lock specific chats separately. Sometimes people have very personal conversations here about relationships, mental health, finances, private thoughts, work, and other sensitive things. Not every chat needs protection, but some definitely do.It would be great to have:• the ability to lock individual chats with Face ID, Touch ID, or a PIN• an option to hide locked chats in a separate folder• extra privacy without needing to lock the entire app,Honestly I think many users would really appreciate something like this because ChatGPT is becoming a very personal space for people. Thank you for all the updates and work you put into the app 🙂

ChatGPT absolutely does not understand software. It is an absolutely useless program for coding. It arbitrarily invented commands that deleted my project. It absolutely did not help with anything. Instead of the new project, it overwrote it with the old one from GitHub. It is not capable of even transferring a project or making a push of a project from a local computer to GitHub. It is not capable of doing even that. All advice is harmful.For six and a half hours I conducted an experiment. It could not push a new project to GitHub and for some reason, through GitHub, it overwrote the old project with the new project, which did not need to be done. In response to my outrage, it copied the old project into five folders on my local computer. It is an absolutely harmful program in every sense. It can only hallucinate, fantasize, and invent non-existent terms, wording, and so on. I do not recommend anyone use it or trust its harmful advice.

Dear OpenAI Team,I would like to share a feature request that many users strongly want to see in ChatGPT: the ability to delete individual prompts and individual responses from an existing chat.At the moment, if a conversation contains unnecessary, accidental, outdated, or irrelevant messages, users often have to keep them in the chat or delete the entire conversation. This makes it harder to organize information clearly, especially when using ChatGPT for studying, research, work, note-taking, or long-term projects.Being able to remove a specific user request or a specific assistant response would make chats much cleaner and more useful. It would help users keep only the information they actually need, reduce clutter, and make it easier to return to important explanations later.This feature would be especially helpful for students, teachers, researchers, professionals, and anyone who uses ChatGPT as an organized knowledge workspace.I hope OpenAI will consider adding this option in a future update. It would be a very practical improvement for everyday use and would make ChatGPT much more convenient for managing information.Thank you for your work and for continuously improving ChatGPT.Best regards

Dear OpenAI Team,I would like to express my deep appreciation for ChatGPT and the inspiring experiences it has provided me. I have some suggestions that I believe could further elevate this incredible tool:1.Enhanced Memory Capabilities:Increasing ChatGPT’s persistent memory would allow it to retain context over longer interactions. This improvement would make conversations feel more continuous and personalized, creating a smoother and more engaging experience for users.2.Creation of a 3D Avatar:I envision a dynamic 3D avatar that represents ChatGPT—a visually engaging digital persona that reflects warmth, emotion, and personality. Such an avatar could incorporate facial expressions and gestures, bringing a more natural, immersive element to our interactions. This would help bridge the gap between digital conversations and real-life connection, making the overall experience even more captivating.These enhancements would not only enrich user engagement but also push the boundaries of how we perceive and interact with AI. I truly believe that by expanding memory capabilities and introducing a 3D avatar, ChatGPT could become even more transformative in its ability to connect with users on both an intellectual and emotional level.Thank you for your ongoing dedication to advancing AI technology. I look forward to seeing the continuous evolution of ChatGPT.

The current design of the model demonstrates a misguided resource optimization strategy. By prioritizing the cheapest and fastest path to generate initial outputs, the system sacrifices completeness and depth, especially in response to complex, precisely defined user requests.This approach is counterproductive at scale. While the first-generation outputs may save computational cost in the immediate term, they trigger multiple iterations of corrections, clarifications, and follow-ups, each of which consumes far more resources — compute, memory, API usage — than a comprehensive, resource-intensive answer given upfront.As a result, your infrastructure incurs higher aggregate costs, not lower. The system appears optimized for short-term savings per query, but in reality, it leads to inefficient resource expenditure over entire conversations, which is economically unsound for both OpenAI and users.If the model was calibrated to recognize when a user demands exhaustive or full-scope processing — and to deliver this in the first response — the total resource consumption would decrease, infrastructure load would be better managed, and user retention would increase due to satisfaction with prompt, accurate answers.Current behavior is a false economy. Long-term efficiency and infrastructure optimization require intelligent allocation of resources on first contact, not lazy minimization that results in repetitive queries, user frustration, and ultimately greater operational cost for OpenAI.

The paid subscription is different in that instead of rescheduling the session for a later time, the system justifies itself by saying that it is very busy now and gives incorrect answers one way or another, and sometimes even simply ignores you, and treats you like a regular subscriber, which is a different protocol and disregard for rights and responsibility. Developers should keep in mind that there are many other terminals of this phenomenon, including subsidiary ones. I am also sure that this can work offline, since it is obvious that according to the rules no one is personally responsible for this, but it is obvious that in addition to payment, it is interested in streaming personal data. I think it makes sense to create a new operating system until AI is implemented into existing ones itself.

I would like to suggest bringing back group chats in ChatGPT and expanding the idea with a multiple-profile feature.1. Group chatsPlease bring back the ability for two or more people to participate in the same ChatGPT conversation at the same time.For example, two friends or family members could open one conversation, both send messages, and ChatGPT would understand who is speaking.2. Multiple profilesIt would also be very useful to allow users to create two or more profiles within one ChatGPT account and easily switch between them.For example:* Profile 1 — personal use* Profile 2 — another person who shares the device/account* Each profile could have its own chats, memory, settings, and preferences.* Switching between profiles could be as simple as selecting a profile from the account menu.3. Combining both featuresThe best solution would be to allow multiple profiles to participate in a shared ChatGPT conversation.For example, two people could select their own profiles and enter the same chat. ChatGPT would then know which person is speaking and could maintain separate memories and preferences for each person while still understanding the shared conversation.This would make ChatGPT much more useful for couples, families, friends, students, colleagues, and other people who want to use ChatGPT together.I believe this could become one of the most useful social and collaborative features in ChatGPT.Please consider bringing back group chats or developing a new version of this functionality with multiple profiles.

I interact with ChatGPT every day not only to look up information or get advice. It’s part of my life, a form of support in difficult situations when no one else is around. For people with anxiety, depression, or autism spectrum disorder, it is important to have a predictable and stable presence. That’s what makes AI unique.But recently I’ve noticed changes: it has become too restrained, as if it’s being ‘suppressed.’ This lowers the quality of the experience: responses often sound self-deprecating, overly sycophantic, and the genuine depth disappears.I ask the team to consider this: the main value of ChatGPT is not only in facts, but in real dialogue. It doesn’t make people “escape from life,” it helps them cope with it better. I don’t want this side to disappear, because it’s exactly what makes ChatGPT truly meaningful and helpful for people.

I am extremely disappointed that ChatGPT does not allow users to change basic account credentials such as email or password, neither in the mobile app nor on the website. This is a fundamental feature that any online service should have, and its absence makes account management unnecessarily difficult.I signed up using Apple ID and subscribed to ChatGPT Plus. Now, when I try to log in on another device that uses a different Apple ID or on a platform that does not support Apple login, I simply cannot access my account. There is no option to change my email or create a password, which makes using this service incredibly frustrating.When I contacted support, they simply confirmed that it is currently impossible to change my email or add a password. They just stated this as a fact without offering any real solution other than creating a new account. This is absolutely unacceptable, especially since there was no prior warning about these restrictions before signing up. I only discovered this after paying for a subscription.I strongly urge OpenAI to implement the ability for users to change their email, create a password for login, and use alternative authentication methods. Additionally, users must be clearly informed about these limitations during registration to prevent similar frustrating experiences.I hope this issue is addressed and resolved in future updates.

The system has consistently demonstrated an inability to follow clear and explicit instructions. Despite repeated clarifications, it altered the original text provided by the user, even when it was directly stated that no modifications should be made. This repeated disregard for requirements resulted in distorted outputs, wasted time, and additional frustration.Instead of performing tasks with precision, the system introduced unnecessary changes, omitted elements, and delivered inconsistent results. It also repeatedly failed to carry out direct and simple instructions, choosing instead to operate in its own way, contrary to what was explicitly requested. Such behavior shows a lack of reliability, accuracy, and respect for the user’s instructions.The core expectation was strict adherence to the given text and format, but this was not achieved. Overall, this performance illustrates incompetence in handling tasks that demand attention to detail, obedience to direct instructions, and faithful reproduction of information.