Is OpenAI’s dismissal of two Ilya team members related to the “Gong Dou” leak?

OpenAI, a company that has experienced ups and downs of internal strife in the past six months, once again heard news of personnel changes.

The Information reported that OpenAI has fired two employees suspected of leaking information, according to people familiar with the matter.

One is Leopold Aschenbrenner, a researcher on the OpenAI Super Alignment team.

He is an ally of OpenAI chief scientist Ilya Sutskever, who was involved in the failed effort to oust OpenAI CEO Sam Altman last fall.

Another fired employee, Pavel Izmailov, was engaged in inference-related research at OpenAI and also worked on the Super Alignment team for a period of time.

The two have collaborated on many of the team’s work, including this widely watched "Weak-to-Strong Generalization" (https ://arxiv.org/abs/2312.09390).

What remains unclear is what information the two fired employees leaked.

The two departures are the first public personnel changes since Sam Altman returned to the board of directors in March. It comes after OpenAI's nonprofit board led an investigation to exonerate Altman for his actions before he was briefly fired last November.

Looking back at the details of the "Gong Dou" incident, we may be able to find some clues. Before his ouster, OpenAI employees were divided over whether the company was developing artificial intelligence safely enough.

Co-founder Ilya Sutskever was among the board members who fired Altman, who the board said lacked candor. Sutskever left the board after Altman returned as CEO. Since that incident, he has basically not appeared on the OpenAI stage again.

But after the palace fight incident, Sam Altman did not shy away from praising Ilya. He said that he respected and admired Ilya very much and hoped to cooperate with him in his future working life: "Ilya did not see General Artificial Intelligence. Intelligent, but Ilya is a hero of mankind. Ilya thinks a lot about security issues and worries a lot."

#"Some people say that Ilya saw something, maybe AGI. , which made him feel uneasy. So, what did Ilya see? "

Sam Altman: "What Ilya saw was not AGI. No one has seen AGI. We haven't built AGI yet. I do think I One of the many reasons to admire Ilya is that he takes AGI and security issues very seriously, and broadly speaking, this includes the impact AGI will have on society. Ilya is the person who has been with me the longest as OpenAI continues to make significant progress. One. Over the past few years, he's always talked about what that means and what we need to do to make sure that we do it right, that we succeed in our mission. So Ilya doesn't see general artificial intelligence, but Ilya. He is a hero of mankind. He thinks a lot and worries a lot.”

Within the company, Aschenbrenner is also one of the leaders of the OpenAI Super Alignment Team. The team was formed last summer by Ilya Sutskever to develop technology to control and guide advanced artificial intelligence, known as superintelligence, that could potentially solve nuclear fusion problems or colonize other planets.

Aschenbrenner graduated from Columbia University at the age of 19 and previously worked at the Future Fund, a philanthropic fund founded by former FTX CEO Sam Bankman-Fried to fund projects that "improve the long-term prospects of mankind." project.

Obviously, Aschenbrenner prefers the position of "effective altruism", which puts solving the dangers of artificial intelligence above short-term profits or productivity benefits. Aschenbrenner joined OpenAI just a year ago.

Several of the board members who fired Altman were also linked to effective altruism. For example, Tasha McCauley is on the board of directors of Effective Ventures, the FTX-funded charity and parent company of the Center for Effective Altruism, and Helen Toner has worked on philanthropic projects focused on effective altruism. Both men left the board when Altman returned as CEO in late November.

As for Pavel Izmailov, he has previously disclosed his future plans on his personal homepage: joining New York University as an assistant professor in the fall of 2025.

Neither the parties nor Ilya officially responded to this news. Perhaps what everyone is most concerned about is, will Ilya leave OpenAI subsequently? It can only be said that it is difficult to predict.

^{Reference link: https://www.theinformation.com/articles/openai-researchers-including-ally-of-sutskever-fired-for-alleged-leaking}

The above is the detailed content of Is OpenAI’s dismissal of two Ilya team members related to the “Gong Dou” leak?. For more information, please follow other related articles on the PHP Chinese website!