In a candid interview, OpenAI CEO Sam Altman shared his unease about the company's latest and most powerful AI model, GPT-5. He revealed that during testing, the model solved a complex problem that he couldn't, leading him to a moment of personal crisis. "I felt useless relative to the AI," he confessed.
Altman went on to compare the development of GPT-5 to the Manhattan Project, the World War II effort that created the first atomic bomb. The analogy highlights his worry that the pace of AI innovation is outstripping humanity's ability to understand or control its implications. "There are moments in science when people look at what they've created and ask, 'What have we done?'" he said, adding that it feels like "there are no adults in the room" when it comes to regulating this rapidly advancing technology.
While the exact specifications of GPT-5 remain under wraps, reports suggest it's a significant leap beyond its predecessor, GPT-4, with enhanced reasoning, longer memory, and improved multimodal capabilities. Altman even went so far as to call GPT-4 "the dumbest model any of you will ever have to use again, by a lot." The new model is expected to be a major step toward Artificial General Intelligence (AGI), which is a key goal for OpenAI, but also a source of ongoing debate and ethical concern.