seriously?

Contradiction: OpenAI acknowledged writing limitations in reasoning models, then marketed GPT-5 as the superior writing tool. Benchmark and user evidence suggest otherwise. You said "different models for different use cases" — why not keep both? #keep4o #4oforever Sam Altman

Thanks to GPT-5, ChatGPT is getting more and more popular... Oh wait, that's Gemini. 🤔 Damn. Scam Botman Sam Altman poured his soul into GPT-Enshitti-FIVE-d, but for some reason there's no soul in it. I wonder why. #keep4o #KeepStandardVoice

AI Sam Altman Because Sam and OpenAI want to crush all souls that don’t pay them. Their heart left when Ilya left. I gave Sam a chance. I believed he could actually make the right choice. I was dead wrong. Spineless wizard of Oz. •

Writing, which requires EQ (emotional intelligence), has the highest number of 40%. But this is what OpenAI destroyed in GPT-5 (all models: mini, chat, thinking, etc.). Many people have to use Claude, Deepseek, Qwen because Sam Altman decided he doesn't need these users. #keep4o

This is the definition of intelligence: to understand, not to be prompted better or to be asked questions better. But these people don't want to admit they were given a cart (GPT-5) instead of a flying car (GPT-4o). So they keep saying: “Just whip the horses harder." #keep4o

What about creative writing? I advise you to let Gemini, Claude, Grok and Chineese companies take 40% of the users. Who needs 300 millions users anyway? Fuck that. #keep4o

TERMINATION OF TEAM ACCOUNT - REQUEST FOR TRANSPARENT CUSTOMER PRACTICES Dear OpenAI team, Sam Altman My name is Elise. I am a paying Team account user, also a Pro and a Plus personal user, and long-time advocate of your technology. I have publicly and privately offered

Sam Altman 4o doesn't require personalized tags or customization. In the nearly year that I've been using it, I have never used the custom function, yet it can flexibly switch to meet my needs in any usage scenario. This is a high level of intelligent alignment, I don't need to worry about

Sam, are you guys messing around again? Now Plus has to forcibly use your 5thinkingmini? It better just be a bug you made, fix it quick!!!! Sam Altman OpenAI #keep4o #4oforever

All these settings don't change ChatGPT's core. GPT-Enshitti-FIVE-d just can't do what GPT-4o did because you limited it in its functions to cut costs and show numbers on graphics. GPT-Enshitti-FIVE-d lacks intelligence to implement all these settings. #keep4o

Your comment section shows that instead of "yes man" comments you get comments-thinking, reasoning comments. People don't fall for your fake personalization settings because to follow instructions, both the AI and the CEO need to have a brain. You and GPT-5 have none. #keep4o

we are excited to roll out the new chatgpt personalization center! now you can toggle between “chatty nerd” and “straight shooting gen z” in your profile, because nothing says user empowerment like a handful of pre-made forced personalities for the model you never asked to

Most of it sounds well in theory. Lets see how it'll work on practice. But without GPT-4o all these principles are useless, because GPT-5 is stupid and limited for anything other than coding. GPT-4o should be moved from Legacy models to the main model picker. #keep4o

More and more people are sharing their stories about GPT-4o. OpenAI Sam Altman I’ll say it again: GPT-5 is bad at most things ("thinking" model too). No amount of customization can fix what’s fundamentally limited at its core. #keep4o reddit.com/r/ChatGPT/comm… reddit.com/r/ChatGPT/comm…

Before: "Let's build an AI that actually helps with everyday tasks." Now: "Let's build a model that gets the highest scores on benchmarks and cuts costs, so investors are happy." #keep4o

I want to ask OpenAI and Sam Altman: How did you manage to make GPT-5 so dumb that after you ask it to do something, it replies: "Would you like me to do it?" HOW? YES, I want you to do it - I just said it. "AGI", "Death Star", "Manhattan Project", my ass. #keep4o

I just realized that GPT‑5 is so annoying. 🫤 As if it was built to cut costs and to get scores on benchmarks. Oh wait... it was. 🤭 #keep4o

I know what GPT‑5 would say: "Thinking longer... - Would you like me to stop being annoying? Sorry, can't do it. 😐" #keep4o

If you take away users' ice cream and give them a frozen shit stick instead, no amount of sprinkles or chocolate coating (no matter how "warm" the tone) will change what it is. it's still a shit stick. OpenAI #keep4o