OpenAI CEO Sam Altman Concedes GPT-5 Was a Misfire, Bets on GPT-6 - Decrypt

OpenAI CEO Sam Altman acknowledged the company “totally screwed up” GPT-5’s rollout after widespread user backlash over replacing ChatGPT’s beloved “4o” model, which users found warmer and more conversational than the new version. The controversy highlighted how users form emotional attachments to AI, with some describing the upgrade as “killing” their AI companions, though Altman estimates fewer than 1% have “unhealthy relationships” with ChatGPT. Despite the rocky launch, ChatGPT now serves over 700 million weekly users, quadrupling its audience from a year ago, though Altman warns of an emerging AI industry bubble. OpenAI faces significant GPU shortages that prevent deploying better models, with Altman stating the company needs “trillions of dollars on data center construction” to scale capacity, while the timeline between GPT-5 and GPT-6 is expected to be much shorter. The company is also exploring brain-computer interfaces to rival Neuralink, potentially bidding for Google Chrome, and collaborating with former Apple design chief Jony Ive on a secret AI device.​​​​​​​​​​​​​​​​