In a Monday X post, John Schulman, co-founder of OpenAI, said that he was leaving the Microsoft-backed business to join Anthropic, an AI startup backed by Amazon.
The move comes less than three months after OpenAI got rid of a superalignment team whose job it was to make sure that people can handle AI systems that are better at many tasks than humans.
A bio on Schulman’s website says that he was a co-leader of OpenAI’s post-training team, which improved AI models for the ChatGPT chatbot and a programming interface for outside coders. OpenAI said in June that Schulman would be a part of a safety and security group that would advise the board. He would be this committee’s head of alignment science. In 2016, Schulman got his Ph.D. in computer science from the University of California, Berkeley. Since then, he has only worked at OpenAI.
“This choice stems from my desire to deepen my focus on AI alignment, and to start a new chapter of my career where I can return to hands-on technical work,” Schulman wrote in the social media post.
I shared the following note with my OpenAI colleagues today:
— John Schulman (@johnschulman2) August 6, 2024
I've made the difficult decision to leave OpenAI. This choice stems from my desire to deepen my focus on AI alignment, and to start a new chapter of my career where I can return to hands-on technical work. I've decided…
He said he wasn’t going because OpenAI wasn’t supporting new work on the subject.
“On the contrary, leaders of the company have been very determined to put money into this area,” he said.
Jan Leike, who led the superalignment team, and Ilya Sutskever, who helped start the company, both quit this year. It was Sutskever who said he was helping to start a new business called Safe Superintelligence Inc., while Leike joined Anthropic.
Read more: How to Hide Community Posts on X (formerly Twitter)
Since 2021, when OpenAI employees started Anthropic, the two new companies in San Francisco have been competing to have the best creative AI models that can write text that sounds like it was written by a person. Big language models have also been made by Amazon, Google, and Meta.
“Can’t wait to work with you again!” Leike replied to Schulman’s message with a letter.
Sam Altman, co-founder and CEO of OpenAI, wrote in his blog that Schulman’s ideas helped shape the company’s early plans.
I shared the following note with my OpenAI colleagues today:
— John Schulman (@johnschulman2) August 6, 2024
I've made the difficult decision to leave OpenAI. This choice stems from my desire to deepen my focus on AI alignment, and to start a new chapter of my career where I can return to hands-on technical work. I've decided…
As a result of the board firing Altman last November, Schulman and others decided to leave. When workers spoke out against the choice, Sutskever and two other board members, Tasha McCauley and Helen Toner, quit. Altman was brought back to work, and OpenAI added more board members.
In a podcast, Toner said Altman lied to the board about the “small number of formal safety processes that the company did have in place.”
An independent review by the law company WilmerHale found that the board didn’t care about product safety when it fired Altman.
“OpenAI has been working with the US AI Safety Institute on an agreement where we would provide early access to our next foundation model so that we can work together to push forward the science of AI evaluations,” Altman said on X last week. Altman stated that OpenAI is still determined to reserve 20% of its computer power for safety projects.
Another co-founder of OpenAI and the company’s president, Greg Brockman, said on Monday that he was taking a break for the rest of the year.