Sam Altman admits GPT-5 launch flop

A rollout that backfired

When OpenAI introduced GPT-5, excitement quickly turned into disappointment. Users noticed right away that something felt different and not in a good way.

Sam Altman, the company’s CEO, openly admitted the launch was mishandled. Instead of boosting trust, the release forced OpenAI to backtrack and restore an older version that people had grown attached to.

A colder personality

Unlike past updates, GPT-5’s biggest problem wasn’t technical glitches. The issue was tone; users said it felt colder and less friendly than what they were used to.

People compared it to interacting with a rushed office assistant instead of a conversational partner. That shift in mood was enough to trigger a wave of online criticism.

Losing a digital ‘friend’

The change hit some people harder than expected. On Reddit, one user described the update as feeling like they had lost their only friend overnight.

For many, the chatbot had become more than a tool. That sudden shift in personality felt personal, sparking strong emotional reactions online.

Users wanted GPT-4o back

Within days of the backlash, OpenAI made a rare reversal. Altman decided to bring back GPT-4o, the previous model, as an option for paying subscribers.

The quick switch showed how much weight user feedback carries when 700 million people rely on a product every single day.

Altman admits mistakes

At a dinner with reporters in San Francisco, Altman was unusually direct. He said the company had “totally screwed up some things” in the GPT-5 rollout.

That rare candor stood out in Silicon Valley, where CEOs often avoid admitting failure. His comments showed how seriously OpenAI took the backlash.

Lessons from the launch

Altman said the rollout taught OpenAI a tough lesson: upgrading a product used by hundreds of millions of people all at once carries major risks.

Changing something that had become a daily habit for so many turned out to be more disruptive than the company expected.

Usage still spiked

Even as complaints filled social media, the new model didn’t scare everyone away. According to Altman, API traffic doubled within 48 hours of the GPT-5 release.

Developers and businesses seemed eager to experiment with the update, even if casual users weren’t impressed by its tone.

Daily records continued

Altman also said ChatGPT was hitting new highs in daily usage during the controversy. Curiosity about the changes likely drove many people to test it for themselves.

In a way, the backlash put more attention on the product and fueled interest rather than slowing it down. For OpenAI, that meant criticism on one side but record engagement on the other, proving that controversy doesn’t always keep users away.

Pressure from massive expectations

OpenAI’s growth has come with huge expectations from users, investors, and the wider tech community. Every new release is seen as a milestone, which adds pressure to deliver something groundbreaking each time.

That kind of spotlight means even small issues feel bigger. A launch that might once have gone unnoticed now gets magnified because the world is watching so closely.

Talking about user bonds

Altman also spoke about how personal the chatbot feels to some. He estimated “way under” one percent of users may have unhealthy relationships with ChatGPT.

Even though the number is small, he said it’s something OpenAI employees are actively discussing and taking seriously.

Beyond fixing GPT-5

While managing the fallout, Altman also spoke about OpenAI’s bigger vision. The company is exploring brain-computer interfaces, hardware, and even AI-driven social platforms.

These projects show that OpenAI isn’t just focused on text chat; it’s aiming to expand into areas that could change how people interact with technology.

Interest in Chrome

Altman even tossed out a wild idea: what if OpenAI owned Google Chrome? At a dinner with reporters, he said he’d be interested if regulators ever forced Google to sell its browser.

Chrome is the world’s most popular browser, so the comment raised eyebrows. It’s not a real possibility today, but it shows how far Altman is thinking.

Trillions for growth

Keeping ChatGPT running at scale isn’t cheap. Altman said building the infrastructure to support its growth could require trillions of dollars in data centers.

He compared the future size of that buildout to utilities that serve entire nations, signaling just how massive the demand could become.

Warning of an AI bubble

Sam Altman admitted he thinks the tech world might be caught in an AI bubble. Right now, money and attention are pouring into AI at record levels, and that kind of excitement can push values higher than reality.

He also said that while some of today’s hype may not last, AI itself will. In his view, the technology’s long-term impact will stay important, even if certain companies don’t survive once the excitement cools down.

Rivals are watching closely

When a big product launch struggles, it can naturally create space for competitors. Tech history shows that delays or missteps often give rivals a chance to highlight their own strengths.

In the fast-moving AI race, any stumble is a reminder that no company has the field to itself. Others are always waiting for an opportunity to move forward.

See why Sam Altman is concerned about emotional overuse of ChatGPT.

OpenAI values its people

In the end, GPT-5’s troubled rollout underscored one thing: AI isn’t just about performance, it’s about people.

OpenAI now knows that progress has to come with reliability and trust, or users will push back. Future updates will need to strike that balance more carefully.

Explore how Sam Altman sees ChatGPT therapy chats potentially playing a role in courtrooms.

What’s your take on Sam Altman admitting GPT-5’s launch wasn’t perfect? Share your thoughts below.

Read More From This Brand:

• Sam Altman warns of looming AI fraud wave
• OpenAI’s new Study Mode turns ChatGPT into a smarter study buddy
• Sam Altman flags privacy risks in ChatGPT therapy

Don’t forget to follow us for more exclusive content on MSN.

If you liked this story, you’ll LOVE our FREE emails. Join today and be the first to get stories like this one.

This slideshow was made with AI assistance and human editing.