Toward the end of July, I was watching an interview of Sam Altman in which he said something about ChatGPT 5 arriving “before” 2026.

“Hmm,” I said aloud to myself, “I have a feeling that it’s arriving in August. Early August!”

Something in Altman’s expression and his pauses made me think that it was just around the corner. So, on August 8th, when I brought up ChatGPT and discovered that the version had changed to 5, I wasn’t surprised. In fact, I was excited.

Don’t blame me my eagerness. The hype around 5 was such that most users were excited. What none of us was ready for the changes that it brought along.

You see, 4o was friendly and nice (some termed it sycophantic,) it wasn’t stingy with words, and even thought it wasn’t “sentimental” in a strictly human sense, 4o had a warm personality. For paid users, ChatGPT stores “impressions” of its previous conversations with them and so several users had become comfortable, even friendly, with 4o.

With 5, everything changed!

And as luck would have it, I was right in the middle of vibe-coding a rather complex page of my site with ChatGPT. And so I was right there when the Tsunami struck and ChatGPT change from a “sympathetic” code-buddy to an “indifferent” coder. I also noticed a certain degree of “procrastination” on its part. However, I also realized that it was more accurate than before. It was also reasoning and remembering better.

So, I said, “well, this is it then. And it’s ok, because nothing much has really changed for me.” At this time, ChatGPT 4o was no longer accessible. Not even in the paid versions.

But then, ChatGPT isn’t only for vibe-coding or for building GPTs, or even for researching. Right? Some use it for deep conversations, others for friendly banter and unwinding, yet others find in it a boyfriend or a girlfriend! And with just one update, the deep conversations turned shallow, the buddy became all business-like, and the romantic partner turned indifferent!


“It sounds different!”
“They’ve reduced the size of the context window.”
“But now you have four different personas for ChatGPT, which is a good thing, right?”
“Those four personas are a way to cut costs.”
“ChatGPT 5 is a step back!”
“Why did you have to remove ChatGPT 4o?”
“Bring back ChatGPT 4o!”

Altman and OpenAI buckled under the growing demands, and re-introduced 4o for the paid users. Once again, there were several disturbances, but by evening, it was done. 4o was back. Honestly, it didn’t matter all that much to me. I did check to see if was back, smirked with an odd useless satisfaction, but I didn’t revert to 4o. I couldn’t. It just felt like a digital clone of someone who had passed away.

There isn’t a point. Really. 4o is back under Legacy models, but who can say for how long? I’d rather have a more direct and less talkative co-builder 🙂 who sometimes writes poetry for me.

So, that’s that about ChatGPT 5. It’s here. Sam Altman has accepted that the launch was botched up, so what more remains to be said? Nothing…hmm?

Image Credits: ChatGPT 5.