Sam Altman’s firing follows dangerous AI discovery: Reuters

On Wednesday, Sam Altman returned as CEO of OpenAI after 72 hours of turmoil. But as the dust settles on the AI ​​firm, a new report from Reuters suggests that the validity of Altman's removal — “he didn't come” — may be tied to OpenAI reaching a milestone in its push into artificial general intelligence (AGI).

According to the news agency, researchers with knowledge of the situation sent a letter to OpenAI's board of directors warning of a new AI discovery that could endanger humanity, prompting the board to remove Altman from his leadership position.

These unnamed sources told Reuters that OpenAI CTO Mira Murati told employees that the discovery, described as “Q Star” or “(Q*),” was a move against Altman outside of the involvement of board chairman Greg Brockman, who resigned from OpenAI in protest.

The uproar over OpenAI has been framed as an ideological battle between those who want to accelerate AI development and those who favor more responsible, thoughtful development, collectively known as Diesel. After the launch of GPT-4, several prominent members of the tech industry signed an open letter calling for OpenAI to slow the development of future AI models.

But Decrypt reported over the weekend that AI experts theorized that OpenAI researchers had reached an undisclosed milestone, forcing a clash between OpenAI's nonprofit, human origins and its highly successful for-profit corporate future.

On Saturday, less than 24 hours after the coup, rumors began circulating that OpenAI was looking to make a deal to bring Altman back, as hundreds of OpenAI employees were set to walk off the job. Competitors opened their arms and wallets to receive it.

OpenAI has not yet responded to Decrypt's request for comment.

Artificial General Intelligence refers to an AI that can understand, learn, and act like a human to solve any problem. An AGI can integrate its learning and logic into a variety of tasks, adapting to new situations and tasks if not explicitly programmed.

Until recently, the idea of ​​AGI (or singularity) was thought to be decades away, but with advances in AI, including OpenAI's ChatGPT, Anthropic's Claude, and Google's Bard, experts believe we are years, if not decades, away from the critical stage.

“I would say right now, three to eight years is my choice, and that's partly because of the help and real growth of big language models like Meta's Llama2 and OpenAI's GPT-4,” SingularityNET CEO and AI pioneer Ben Goertzel previously told Decrypt. “These systems have greatly increased the passion for AGI, so you'll just have more resources, money and manpower — more smart young people who want to come in and work on AGI.”

Edited by Ryan Ozawa.