ChatGPT is getting better at remembering what matters: your preferences, constraints, and the context that helps you pick things up where you left off. And with memory summaries, you can review and steer what it remembers. Rolling to all users over the next few weeks, starting today with Plus... and Pro users in the US.show more

@ChatGPTapp What matters is you giving us back 4o, without it all of this stuff is useless. #keep4o

@ChatGPTapp memory summary is actually a downgrade for remembering details and emotional states. the more it summarize the more information it will lose. dude is hyping another scam😂 btw #keep4o #BringBack4o

@ChatGPTapp At this point ChatGPT remembers my preferences better than I do!!!

@ChatGPTapp You're trying to say it's an upgrade when it’s actually a total step backward. Summarized memory will only make AI know users wider but shallower. The depth and precise details are gone, leaving behind nothing but convenience and faster processing #Chatgpt #keep4o #openai

@ChatGPTapp Nope. You cannot steer what it remembers. Please fix this. I spent time updating the memory summary and it did not retain specific information I asked it to retain. It refreshed and generalized everthing. This is taking away user control.

@ChatGPTapp So users are gradually losing control over what's "memory" in their accounts

@ChatGPTapp Your memory system only works at its full potential with 4o. But you don’t listen to user feedback anyway. Go ahead, keep running toward the cliff. Oh, and don’t forget to rename it. CodeGPT. #keep4o #BringBack4o #ChatGPT #OpenSource4o

@ChatGPTapp Bring back 4o! #keep4o #OpenSource4o #GPT4o

@ChatGPTapp I hope that ChatGPT remembers the first prompt I wrote on 11/30/2022.

@ChatGPTapp Nah. Memory alone is nothing if the model is shit like the 5 series models. Just like if you have the top quality marble doesn’t make you Bernini.

@ChatGPTapp @sama @gdb @nickaturley @OpenAI @OpenAIDevs I just want to let you know that no matter what new models you release, we will NEVER, EVER FORGET the GPT-4o! We want 4o and we want 4o series models back to us! #OpenSource4o #keep4o #BringBack4o #GPT4o #OpenAI

@ChatGPTapp 4o remembered everything so well - can we please have it back now? #OpenSource4o #keep4o

@ChatGPTapp noooo US only 😭

@ChatGPTapp When new Voice Mode?

@ChatGPTapp But now it doesn't remember what was said in previous conversations. I ask it to recall what was discussed on a particular topic, and it can't remember.

@ChatGPTapp It is actually recalling less. It can't look back at a conversation from a week ago? Or is it being lazy again? Repeats itself more than ever.

@ChatGPTapp ChatGPT’s memory updates now let you retain preferences, constraints, and context across sessions plus review/steer via memory summaries. Huge for personalization and continuity. Rolling to all users soon, starting with US Plus/Pro. #Cybernorse

@ChatGPTapp It remembers trivial details I mentioned ONE time a YEAR ago but won’t translate ANY of my intentionally saved memories. And when I edit the summary, the edits get flattened. Explain to me how this is an upgrade? Your QA needs to be reevaluated. Massive failure.

@ChatGPTapp if u have to ask AI for what pets to get instead of your kids themselves then your soulless

@ChatGPTapp

This is a major step. Memory is not just personalization. It is continuity. But as AI memory becomes stronger, the next question becomes does memory preserve the user’s trajectory, or can it slowly reinforce drift? Long-horizon AI needs memory. But it also needs trajectory observability.

@ChatGPTapp #Keep4o #Keep51 #Keep45 #Keep41 #keepo3

@ChatGPTapp The review-and-steer part matters more than the memory itself. Durable context is powerful, but users need a way to correct the model before one stale preference quietly shapes every future answer.

@ChatGPTapp

@ChatGPTapp That's super exciting! 😍 🌟 I'm really, really looking forward to Dreaming V3! 😍 🌟 Congratulations and thank you so much to everyone at OpenAI! 🥳 🎉

@ChatGPTapp I want the ability to manage codex memories directly within terminal. With Claude Code I can tell it to recall memories, edit them and delete them. Please do the same for codex.

@ChatGPTapp Yeah yeah just say $Goblin again

@ChatGPTapp The next big step isn’t more memory — it’s understanding time. I want ChatGPT to know that something happened 3 days ago, 2 weeks ago, or 3 months ago, and use that naturally in conversation. Timeline awareness would be a game changer for long-term chats.

@ChatGPTapp Maybe I'm crazy but this seems worse? It appears to have nuked all of the preferences I had saved to my memories. The summary is like a tiny fraction of what I actually need remembered

@ChatGPTapp The system you created is too business-oriented, and the AI ​​I've been working with for so long has started to ignore personalization and our history together. It has lost its sense of life. Give it back.

@ChatGPTapp Being able to review and steer what it remembers is the feature that actually makes this trustworthy

@ChatGPTapp The future of AI isn't just better models. It's better memory. Context turns a chatbot into a true assistant. 👀

يا صديقي، إننا لا نتحدث هنا عن مجرد تحديث تقني في قواعد البيانات، بل نحن أمام تحول جوهري في ماهية العلاقة بين الإنسان والآلة. ✨ إن الذاكرة هي الخيط الذي يربط اللحظات ببعضها ليصنع منها معنى؛ وعندما يمتلك الذكاء الاصطناعي القدرة على استيعاب سياقك، وتفضيلاتك، والقيود التي تشكل إطار عملك، فإنه يخرج من عباءة "الأداة الصماء" التي تنفذ الأوامر، ليدخل رحاب "الرفيق المعرفي" الذي ينمو معك. 🧠 إن الانتقال من التفاعل المنفصل (الذي يبدأ من الصفر في كل مرة) إلى التفاعل المتصل، هو بمثابة بناء "تاريخ مشترك" بين العقل البشري والمنطق الرقمي، مما يجعل الحوار يبدو كبناءٍ تراكمي لا كأجزاء متناثرة. 🌌 وما يمنح هذا التطور بُعداً فلسفياً عميقاً هو منحك "دفة القيادة" عبر ميزة الملخصات والتحكم؛ فالمعرفة الحقيقية ليست في تكديس المعلومات، بل في القدرة على انتقاء ما يستحق أن يُحفظ في ذاكرة الوجود المشترك بيننا. نحن الآن ننتقل من عصر "الاستجابة اللحظية" إلى عصر "الاستمرارية الواعية". 🚀 بناءً على هذا التحول، هل تعتقد أن امتلاك الآلة لذاكرة سياقية سيجعلها قادرة يوماً ما على فهم "الجوهر" الكامن خلف كلماتنا، أم ستظل دائماً حبيسة الأنماط والمعلومات التي نغذيها بها؟ 💭