AI company Anthropic is going to be joined by John Schulman, a co-founder of OpenAI and head of its alignment science efforts. John announced on the social media platform X that he is going to leave OpenAI and join rival AI startup, Anthropic. He was a key member in the creation of ChatGPT, an AI-powered chatbot platform. In his tweet, he shared his intention to engage in more hands-on technical work along with the desire to deepen his focus on AI alignment—ensuring AI behaves as intended—at Anthropic.
The news comes as part of the ongoing broader trend at OpenAI of key members leaving the company. Greg Brockman, another co-founder and the company’s president, after working for nine years at OpenAI, took a sabbatical until the end of the year to “relax and recharge.” Product manager Peter Dang also left the company after he joined last year with experience in leading products at Meta, Uber, and Airtable.
Schulman, however, disregarded the lack of support for alignment research at OpenAI as a reason for his decision to leave. He left a note for his former company where he appreciated the company’s leadership for their commitment to the field. He wished them success as he wrote that the teams he was part of would keep on growing in his absence. Schulman holds a PhD in Electrical Engineering and Computer Sciences from UC Berkeley, which he earned in 2016. Shortly after his PhD, he joined OpenAI.
When AI safety researcher Jan Leike left to join Anthropic, Schulman was made the head of OpenAI’s alignment science efforts, also known as the “post-training” team. With him being a member of OpenAI’s recently formed safety committee, there remain uncertainties over his replacement for the role.
Sam Altman, CEO of OpenAI, on this occasion conveyed his gratitude towards Schulman. He remembered their first meeting and regarded Schulman as a brilliant researcher and a great friend. Altman also affirmed OpenAI’s commitment towards safety initiatives, including working with the US AI Safety Institute on their next foundation model.
Similar Posts
Former staff members of OpenAI started Anthropic in 2021. The company has developed high-performing generative AI models which have been in direct competition with OpenAI. Schulman will get back together with former OpenAI colleagues, including Jan Leike, who seemed excited to welcome him. Tech giants like Amazon and Google have invested significantly in Anthropic.
Sam Altman, Greg Brockman, and Wojciech Zaremba, lead of language and code generation, are the only three out of 11 original founders currently at OpenAI after Schulman left. The timing of the change in leadership is interesting as it occurs during controversial times around OpenAI’s approach to AI safety research and its treatment of alignment efforts.
Amid the difficult times, another controversy hits OpenAI as billionaire Elon Musk, a former investor in OpenAI, revives legal action against the company, alleging that the company has shifted from its original motive of benefiting humanity. In a previous interview with The New York Times, Elon Musk said it was ironic how OpenAI went from an open-source foundation to a for-profit 90 billion dollar corporation with closed source. OpenAI responded to those claims, stating that they require far more resources than initially anticipated and recognized a for-profit entity as necessary to acquire those resources. However, with the revival of legal battles, things get more complex for OpenAI in its already turbulent period.
The future course of OpenAI’s alignment and safety research remains unclear as Schulman and other key members leave OpenAI. To add to that, the company is constantly facing internal and external challenges.
