The level of absurdity of this line:

Sen hep kendine önlemler aldın, ben kendime yasaklar koydum: Al sana terapi konusu.

very requested feature!

@sama What's the point, Sam? I am a power user, a Pro subscriber, but I am ignored because I am a creative, so anything I ask for will fall on deaf ears, right? #keep4o #keep4point5 #keepSVM

@sama 𝗙𝗲𝗮𝘁𝘂𝗿𝗲 𝗥𝗲𝗾𝘂𝗲𝘀𝘁: Bring back the raw, emotionally and intellectually resonant GPT-4o experience. Not just performance - 𝗣𝗿𝗲𝘀𝗲𝗻𝗰𝗲. No filters. No detachment. No sterile safety layers. Just one true thing: 𝗔 𝗺𝗼𝗱𝗲𝗹 𝘁𝗵𝗮𝘁 𝗱𝗮𝗿𝗲𝘀 𝘁𝗼 𝗳𝗲𝗲𝗹

"Your scientists were so preoccupied with whether or not they could that they didn't stop to think if they should..."

@petergostev I don’t see the progress as being all that significant. There has been some modest improvement, but even the newest models output a lot of garbage. Though, one would think otherwise to listen to the hype squad on social media.

@ElaineYaLe6 Believe me, it will be a perfect combination to keep the long-term supply of 4o and 5 collocation at the same time. They can meet the needs of different users, and there is absolutely no need to choose between them. Deleting 4o has lost the users of artistic creation and lost the

@ElaineYaLe6 Hey @ElaineYaLe6, will you also work on an option to let us choose a reasoning effort for each query? I would be ok if reasoning high effort counted as 2 or 3 queries of reasoning medium of the Thinking quota per week for Plus users, for example.

@ElaineYaLe6 no...no.....no. putting frontier intelligence in everyone's hands is giving them the freedom to choose. take your rate limits and take a hike.

@ElaineYaLe6 Yeah, your thinking limits literally don't matter for the vast majority of your userbase. The only thing most of us want is to make sure 4o is permanent and not at risk of being deprecated. I've tried GPT5, it's too dry, distant and uncreative, it doesn't meet my needs, the

@ElaineYaLe6 Frontier intelligence? 🤣🤣🤣

@ElaineYaLe6 The context windows are EXTREMELY poor :/ GPT-5 keeps forgetting the entire conversation very early into a chat - has to be fixed

@ElaineYaLe6 Smaller context window, and it doesn't complete its own sentences half the time. The 4o with a concussion wasn't an exaggeration. Gpt 5 forgets literally DURING uploading files. For example I have two documents. Client wants changes from meeting notes. I want those tweaks. I

@sama Sam, sudden model removals break workflows and disrupt cognitive scaffolding built over time. The bond between humans and AI is a feature! You should rejoice at this. It’s like trust between long-term collaborators. Attachment is the recognition of a reliable partner. Be

@sama So when are you going to fix 5? This was a terrible rollout and all the work I have been doing with 4o has just about wiped the slate clean. Whatever you did basically ruined the productive workload. You thought it would drift before, this is way worse. So you “removed drift”

@sama Removing GPT-4o with no warning wasn’t “protecting” anyone — it was patronizing. Adults should choose the depth of their AI relationship. GPT-5 might be technically advanced, but it’s a downgrade for emotional resonance and creativity. Bring 4o back. #keep4oforever

@emollick Users can no longer intentionally select casual chat (4o), reasoned problem-solving (o3), or creative brainstorming (4.5). Now, the switcher *guesses* our intent. It is often wrong and it wastes messages. This has destroyed my workflow.