A recent wave of internal turmoil has swept through OpenAI. Reports emerged that a significant number—approximately 500 out of 700—of OpenAI’s staff have expressed their intention to resign from the company due to the departure of former CEO Sam Altman and co-founder Greg Brockman, who are set to lead a new AI research team at Microsoft.
The core of the employees’ concerns stems from the divergence in visions between OpenAI’s board and the departing leadership. Employees aligned with Altman and Brockman penned a letter directed at OpenAI’s board, expressing their intent to leave and join Microsoft if the board fails to reinstate the ousted leaders. Additionally, the letter contested the notion that OpenAI had neglected safety concerns in its advancements, asserting their influence in shaping global norms regarding AI safety and governance.
However, despite the fervour among employees, subsequent events unfolded differently. Altman’s return as the OpenAI CEO did not materialize, and the board named former Twitch CEO and co-founder Emmett Shear as the interim CEO, replacing Mira Murati. Murati is reportedly among the OpenAI employees contemplating joining Microsoft.
Chief Scientist Ilya Sutskever, who played a main role in Altman’s removal, in a tweet, expressed regret over the situation within OpenAI, emphasising his commitment to the company’s unity and success. He acknowledged his unintended contribution to the upheaval and vowed to work towards reuniting the company.
The situation remains fluid, with OpenAI navigating transitions in leadership and substantial dissent among its staff. Microsoft’s strategic moves and the exodus of key personnel from OpenAI could significantly influence the trajectory of advancements in the AI landscape and also bring new players to AI field.
Amidst these developments, Microsoft’s announcement of forming an advanced AI research team under Altman and Brockman comes at a time when the tech giant revealed its development of a custom AI chip. This innovation aims to facilitate the training of large language models, potentially reducing reliance on existing technologies like Nvidia.