news

Power reshuffle: New forces emerge amid OpenAI turmoil

2024-08-14

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

The successive core senior personnel changes at OpenAI, a leader in artificial intelligence technology, have attracted widespread attention in the industry. In the past few months, OpenAI has experienced a series of high-profile departures, including two co-founders, a product director, a security director, and a former board member. Some of them left permanently, while others left temporarily. Only two of the 11 senior executives remain. In addition, the company has also undergone multiple reorganizations.

Media reports said that after a wave of senior executives leaving the company, a new group of leaders has entered the company to fill the vacancies.

Among them, Jakub Pachocki has been promoted to chief scientist. Pachocki is a senior member of OpenAI and has been with the company since 2017. In May of this year, Pachocki was promoted to chief scientist after the departure of co-founder and researcher Ilya Sutskever. Sutskever is one of OpenAI's highly respected co-founders, and he supported the brief ouster of CEO Sam Altman last November. Holding this position gives Pachocki the power to set the company's research direction and vision.

However, current and former employees told the media that Pachocki's influence was evident before his promotion. It is reported that Altman had given Sutskever and Pachocki similar responsibilities in the past, although Pachocki reported to Sutskever, which led to tension between the two. Since Sutskever's departure, Pachocki has become a key decision maker within the company with the support of Altman.

Barret Zoph serves as the head of the post-training department. Analysts believe that the recent departure of co-founder John Schulman will make his only direct subordinate Zoph more prominent in the company. According to an employee who revealed to the media, under Schulman's leadership, Zoph is responsible for the day-to-day management of the post-training team, while Schulman is responsible for setting high-level plans for the team. The task of the post-training team is to ensure that OpenAI's basic models can be successfully applied to ChatGPT and its API for developers. As a former Google Brain researcher, Zoph coordinated the team's research work and ensured that the project was completed on time, which earned him a high profile in the company.

Mark Chen becomes head of frontier research. Chen joined OpenAI in 2018 after working in quantitative trading for many years. He now leads a large and well-known frontier research group under the department of Vice President of Research Bob McGrew.

Media reports said Chen's influence at OpenAI was highlighted during the board's removal of Altman last November. At the time, Chen, Zoph and post-training researcher Liam Fedus were the main liaisons between leadership and employees. According to a former employee who spoke to the media, the three conveyed the strategy behind the letter in support of Altman, which was signed by most employees and stated that they would join Microsoft if Altman failed to be reinstated.

Lilian Weng becomes head of the safety systems department. In July, OpenAI transferred Aleksander Madry, the team leader who was previously responsible for ensuring that artificial intelligence does not cause harm to society, to a basic working group focused on the field of reasoning. At the same time, the company integrated its safety team under Weng's leadership. Now, Weng will manage teams focused on long-term and short-term safety, an organizational decision that some researchers worry could lead to conflicts of interest.

Wall Street Journal previously stated that the tension between its non-profit mission and commercial transformation has become increasingly apparent, causing core team members to leave. Some analysts also pointed out that the reason for the staff loss may also be related to Musk's lawsuit this week. In addition, funding and competitor pressure have also made the situation facing OpenAI more complicated.