- OpenAI’s Sam Altman says the firm ‘screwed up’ with ChatGPT 5.2
- The focus was on technical ability, not user interactions
- Altman said OpenAI focuses on some features and must ‘neglect’ others
ChatGPT users are not shy about voicing their concerns when they feel an update has broken OpenAI’s chatbot, as was ably demonstrated when ChatGPT 5.2 made its debut. This update was met with howls of displeasure, and now OpenAI has admitted that it “screwed up” when implementing the change.
The comments came from OpenAI chief Sam Altman, who was speaking to a developer town hall meeting earlier this week. At the event, Altman was asked about negative user feedback surrounding ChatGPT 5.2 and claims that this version of the artificial intelligence (AI) tool had produced content that was “unwieldy” and “hard to read.”
In response, Altman was forthright: “I think we just screwed that up,” he said, before continuing: “We will make future versions of GPT 5.x hopefully much better at writing than 4.5 was.”
Interestingly, Altman lay the blame on OpenAI’s decision to focus on ChatGPT 5.2’s technical aspects rather than its writing ability: “We did decide, and I think for good reason, to put most of our effort in 5.2 into making it super good at intelligence, reasoning, coding, engineering, that kind of thing,” he stated. “And we have limited bandwidth here, and sometimes we focus on one thing and neglect another.”
Making sacrifices
Altman’s explanation is illuminating because it sheds light on OpenAI’s decision-making practices. It suggests that there will always be elements of ChatGPT that have to be sacrificed when trying to improve others. In the case of ChatGPT 5.2, that was perhaps more noticeable than in previous updates because it concerned the way the chatbot spoke to you and worked with your prompts.
We can see this differing emphasis in the ChatGPT 4.5 update, for example. At the time, OpenAI said it had improved the way the chatbot interacted with users, claiming that the result was a bot that “feels more natural” compared to past iterations.
In the case of ChatGPT 5.2, OpenAI noted its improvements to tool use, coding and document creation, but it was personal interactions that felt off for many people. It highlights how OpenAI’s releases can at times focus on one area, and at other times on another.
When we pitted ChatGPT 5.2 up against Google’s Gemini 3 chatbot, we found the two options to be neck-and-neck across a range of tests. But it’s obvious that for some people, ChatGPT 5.2 just wasn’t up to scratch.
With OpenAI acknowledging the discontent, there’s a good chance we’ll see amendments to the chatbot that point it back in the right direction for disgruntled users.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.
https://cdn.mos.cms.futurecdn.net/w4FM89EMTyEckeVYVYGBYM-1920-80.jpg
Source link
alexblake.techradar@gmail.com (Alex Blake)
