OpenAI CEO Sam Altman says it needs to fix these 3 weaknesses to make ChatGPT as good as possible
Published on Mar 11, 2026 at 7:52 AM (UTC+4)
by Claire Reid
Last updated on Mar 11, 2026 at 7:52 AM (UTC+4)
Edited by
Claire Reid
OpenAI CEO Sam Altman has admitted that ChatGPT has ‘three weaknesses’ but has promised that the company ‘will be able to fix’ them
OpenAI released the AI chatbot in November 2022, and in just two months, it had amassed a staggering 100 million users.
However, ChatGPT has had a much bumpier time more recently, with a reported 1.5 million people leaving the app, and many of them opting for Claude instead.
Now, CEO Altman has agreed that GPT-5.4 has some weaknesses it needs to address.
Click the star icon next to supercarblondie.com in Google Search to stay ahead of the curve on the latest and greatest supercars, hypercars, and ground-breaking technology
ChatGPT CEO Sam Altman has said GPT-5.4 is ‘favorite’ model to talk to
Back in February, OpenAI caused a stir when it removed the 4o model, and replaced it with 5o.
At the time, many ChatGPT users were upset by the move, as they felt that 4o had a better ‘personality’ than the newer one.

There have even been some wide spread social media campaigns urging the CEO to rethink the decision the older model, but so far the company has resisted.
In a recent post on X, Altman admitted that previous 5o models ‘missed the mark’ when it came to the chatbot’s personality.
But that had all changed with with the newly released GPT5-4.
In the post, he explained that while the chatbot was ‘great’ at tasks such as coding and ‘knowledge work’, he had also found that it was his ‘favorite model to talk to’ and said it was ‘extra good to be moving in the right way’.
However, he has admitted that three areas will be getting fixes
OthersideAI CEO Matt Shumer is also a fan of the updated model, even going so far as hailing it ‘the best model in the world, by far’ in a recent post on X.
But, he also shared what he believes to be three weaknesses of ChatGPT – and it looks as though Altman might agree.
In the post, Shumer said that the ‘frontend taste’, which presumably means how it looks and is styled, of ChatGPT is ‘far behind’ rivals Gemini and Opus.
He went on to say that once that issue was reminded he would have ‘literally no reason’ to use any other chatbot.
But that wasn’t the only issue he pointed out.
“It can still miss obvious real-world context,” he continued.

“For example, I had it plan an itinerary for a trip. At first glance, it looked perfect, but it failed to take into account that it chose locations that would be mobbed by spring breakers, so I had to re-run the prompt from scratch with more context.”
Finally, he said when testing it inside OpenClaw, GPT5-4 ‘kept stopping short before finishing tasks’.
The post, which racked up more than 3,000 likes, came to the attention of Sam Altman, himself, who appeared to agree but also looked to have a plan in mind.
“We will be able to fix these three things,” he wrote in response.
He didn’t give a timeline for when these changes might be implemented, but we all know things can move very fast when it comes to ChatGPT.
DISCOVER SBX CARS: The global premium car auction platform powered by Supercar Blondie
With a background in both local and national press in the UK, Claire has covered a range of topics, including technology, gaming, and cryptocurrency, since joining the editorial team at Supercar Blondie in May 2024. Her ability to be first to a story has been integral to making SB’s coverage of scientific discovery, AI, and global tech news a slick 24/7 operation.