OpenAI Director John Schulman Joins Amazon-Backed Anthropic Amid Leadership Shake-Up

John Schulman, OpenAI's director of alignment, has joined Anthropic, an AI startup backed by Amazon. His departure follows the recent exit of OpenAI's president, Greg Brockman, and product manager, Peter Deng.

John Schulman Leaves OpenAI for Anthropic, Joining Amazon-Backed AI Startup Amid Leadership Changes

John Schulman, the director of alignment at OpenAI, was responsible for customizing the model responses to align with the company's values. He has since departed the company to collaborate with Anthropic, which Amazon backs. Anthropic and OpenAI share a common history, and the former comprises former executives from the latter who have chosen to establish their own company. Upon announcing his decision regarding X, Schulman was greeted by responses from OpenAI chief Sam Altman and others who expressed gratitude for their time working together at the company.

Schulman's X post was published shortly after The Information reported that OpenAI's president, Greg Brockman, had taken an extended leave of absence and that a product manager, Peter Deng, was also departing the company. Schulman, previously accountable for the model alignment process at OpenAI, will now be employed by Anthropic, which Amazon supports. Former members of OpenAI established Anthropic, and Schulman's departure has reduced the number of OpenAI's founding members. Brockman and Altman are two of the remaining members.

As OpenAI's director of alignment, Schulman was accountable for guaranteeing that the firm's products, including ChatGPT, generated outputs corresponding to a predetermined set of values. Examples include restricting harmful outputs and tailoring responses against widely held beliefs to ensure that all responses generated by AI models are secure for human use.

Schulman discloses in his statement that he is departing OpenAI to "further" his concentration on AI alignment and "return to hands-on technical work." He further clarifies that the decision is not the result of a lack of emphasis on alignment at OpenAI but rather a "personal one."

OpenAI Faces Leadership Changes as John Schulman Joins Anthropic, Focusing on AI Alignment and AGI Goals

Following his departure, Wojciech Zaremba, the head of language and code generation at OpenAI, is the third founder currently employed by the company, after Altman and Brockman.

In response to Schulman's announcement, Altman, the CEO of OpenAI, expressed gratitude to Schulman for his contributions to the organization. Altman also disclosed the specifics of his initial encounter with Schulman, stating that they had encountered each other at a café at the University of California, Berkeley, where they discussed artificial general intelligence.

While ChatGPT represents a form of restricted intelligence, the ultimate goal of all AI firms, including OpenAI and Anthropic, is to develop artificial general intelligence (AGI) software. According to Wccftech, this software aims to generate multifaceted outputs that resemble the cognitive capabilities of a human. The future of AI at both OpenAI and Anthropic is promising, with the pursuit of AGI at the forefront of their objectives.

Following the departure of AI safety researcher Jan Leike in May, Schulman has also departed the company. Leike left OpenAI after the company disbanded an alignment team and joined the Amazon-backed Anthropic to oversee its alignment initiatives. Schulman assumed responsibility for what OpenAI calls "post-training"after Leiki's departure. His post emphasized that post-training is "in good hands" and has a "deep bench of amazing talent." Additionally, it is "coming together with some promising projects."