OpenAI CEO Sam Altman states the company is confident it knows how to build AGI as traditionally understood, referring to the tech industry's long-sought benchmark of artificial general intelligence. He predicts AI agents capable of autonomously performing certain tasks may start to 'materially change the output of companies' this year. Altman made this announcement in a blog post published on Monday, where he discussed the past and future of OpenAI.
The company's next goal is 'superintelligence in the true sense of the word,' he says. 'Superintelligent tools could massively accelerate scientific discovery and innovation well beyond what we are capable of doing on our own, and in turn massively increase abundance and prosperity.' Altman's description of superintelligence, broadly defined as AI agents that surpass human intelligence, sounds much like how OpenAI once described AGI: 'AI systems that are generally smarter than humans.' OpenAI has repeatedly stated that its foundational goal is to develop an artificial general intelligence (AGI) system that 'benefits all of humanity.' But Altman tried to lower expectations for AGI last month, saying it would 'matter much less' than people thought. Downplaying its previous definition of AGI may work in OpenAI's favor, since the company is tied to Microsoft via exclusivity deals it made in 2023 until OpenAI officially declares that AGI has been achieved. But reports that Microsoft itself classifies AGI as a system capable of generating $100 billion in profits may prove that loophole ineffective. And while The Verge has heard that OpenAI plans to blend its large language models together as an 'AGI' offering, potentially to hit this goalpost faster, the company isn't turning a profit. Altman also says it's currently losing money on its $200-per-month ChatGPT Pro subscription