The race for Artificial General Intelligence, or AGI, has been a driving force behind the AI boom. For years, OpenAI‘s mission has been to safely develop AGI, a form of AI capable of performing any intellectual task a human can. But now, CEO of the company Sam Altman, at the center of that mission, is changing his tune regarding AGI pursuit.
In a recent interview, OpenAI CEO Sam Altman admitted that “AGI is not a super useful term.” His new stance comes as the company continues to make rapid advances in the AI space. This apparent new vision makes the concept of AGI difficult to define.
“AGI” has become a vague and hype-driven term for OpenAI CEO Sam Altman
The problem, according to Altman, is that different people and companies use different definitions of AGI. For some, it’s an AI that can perform “a significant amount of the work in the world.” But even that is an ever-changing target as the nature of work evolves. Ultimately, Altman believes it’s a distraction from the real, exponential progress being made in AI model capabilities.
He’s not alone in this thinking. Industry experts agree that while AGI serves as a “North Star” for inspiration and can help secure funding, its vague, sci-fi definition often creates a “fog of hype.” It can obscure the tangible, specialized progress that is making a real difference in the world right now.
From binary to breakthroughs
This new perspective seems to be part of a broader shift at OpenAI. Altman now prefers to talk about levels of progress toward general intelligence rather than a binary “is it AGI or not?” state. He believes this is a more useful way to measure progress as AI gets closer to that goal.
Still, the company still maintains AGI as its ultimate goal. However, Altman still expects major breakthroughs in specific fields, such as new math theorems and scientific discoveries, within the next couple of years. This focus on specific achievements over a nebulous concept of “general” intelligence may be the new direction for the industry’s leaders. It’s a reminder that while the grand vision of AGI is still alive, the real and measurable progress in specialized AI will truly shape the future.
Meanwhile, OpenAI has been busy releasing its GPT-5 AI model. The company’s promises of a unified model that would always automatically fall back to “ideal mode” for every request fell apart after initial tests. The initial bumpy rollout even forced the company to return access to older versions of GPT while they fix GPT-5.
RELATED POSTS
View all
