OpenAI enters third phase, aims for personal AGI for everyone

OpenAI's new plan draws a direct parallel between the rollout of electricity in the 1920s and the current state of AI, arguing that the technology's deeper impact will come from the new possibilities it opens for people. The company frames its goals within a historical context where widespread access to a transformative technology led to a tripling of median inflation-adjusted income and a lifespan increase of over 20 years by the end of the 20th century.

The announcement outlines a transition where AI does AI research, potentially handling a significant fraction of OpenAI's work by March 2028. This push for speed is coupled with a stated need for human judgment to become more important, not less. The company emphasizes that faster technical progress makes public coordination more important, and it reiterates support for an international organization to coordinate leading AI efforts and reduce catastrophic risk.

The ultimate test of this vision lies in the company's ability to make the technology both powerful and resilient. As the post notes about the automobile, the goal is not to stop progress but to build the societal systems—like seatbelts and traffic laws—that make a powerful technology safe for widespread use. Whether OpenAI and other developers can foster that kind of resilience ecosystem will likely determine if the benefits of AI are broadly distributed or become concentrated.