Openai CEO Sam Altman and several other scientists and engineers came to Reddit the day after debuting the powerful new GPT-5 AI model for the honorary tradition of an Ash Me Nothing thread.
Although the discussion varied over all kinds of technical and product elements, there were a few topics that stood out as particularly important for posters based on the frequency and passion with which they were discussed. Here are a few of the most notable things we learned from Openai Ama.
Pining for GPT-4O
The biggest recurring theme in AMA was a sorrowful crying from users who loved the GPT-4o and felt personally attacked by its removal. It’s not an exaggeration that a user posted, “Bring 4o GPT-5 carries the skin of my dead friend.” To which altman replied, “What a … atmospheric picture. Ok, we hear you at 4o, work on something now.”
This was also not just an isolated request. Another post asked to keep both GPT-4O and GPT-4.1 along with GPT-5 and argued that the older models had different personalities and creative rhythms. Altman admitted that they “looked at this now.”
Most requests were a little more muted, with a poster that asked, “Why do we get rid of the variants and 4o when we all have unique communication styles? Bring them back!”
Altman’s answer was short, but directly admitting the point. He wrote, “Ok, we hear you all at 4o; Thanks for the time for giving us feedback (and the passion!). We will bring it back to plus users and will see need to decide how long to support it.”
It is interesting that so many heavy users seem to prefer the style of the older model and prefer it rather than the objectively better newer ones.
Filtration history
Another big topic was Chatgpt’s security filter, both at the moment and before the GPT-5, which many issued complaints about being exaggerated. A user described a scenario where they had been marked to discuss historical topics, with an answer that Gauguin was marked and deleted because the artist was a “sex pest” and the user’s own clarification question is marked himself.
Altman’s answer was a mix of deal and reality check. “Yes, we will continue to improve this,” he said. “It’s a legitimate tough thing; the lines are often really quite blurred sometimes.” He emphasized that Openai wants to allow “very wide latitude,” but admitted that the boundary between uncertain and safe content is far from perfect, but that “people of course should not be banned for learning.”
New level
Another asker reset on a hole in Openai’s subscription model: “Planning in guys to add another plan for solo power users who are not professional? $ 20 plan offers too little for some and the $ 200 level is excessive.”
Altman’s answer was brief and simply said, “Yes, we will do something here.” No details, just a confirmation that the idea is on the table. This brevity leaves open options from ‘next week’ to just say ‘the discussion starts now.’ But the price difference is a big thing for power users who find themselves limited by the plus level but cannot justify a company prices. If Openai creates an intermediate level, it can reshape how dedicated individual users engage in the platform.
The future
At the end of AMA, Altman shared some new information about the current and future state of chatgpt and GPT-5. He started by admitting some problems with release and wrote that “We expected some unevenness as we roll out so many things at once. But it was a little more uneven than we hoped for!”
This unevenness ended up making GPT-5 seem not as impressive as it should have so far.
“GPT-5 will seem smarter from today,” Altman wrote. “Yesterday we had a seven [severity, meaning system issue] And Autoswitcher was out of service for part of the day and the result was GPT-5 seemed dumber. “
He also promised more access to chatgpt plus users with twice as large interest rates as well as the upcoming return of GPT-4o, at least for the same subscribers. Ama painted a clearer picture of what Openai is willing to change in response to public pressure.
The return of GPT-4O for Plus users at least recognizes that raw capacity is not the only metric that matters. If users are this vocals of keeping an older model alive, future releases of GPT-5 and above may be designed with more conscious taste built in beyond just the personality types promised to GPT-5.
