Sam Altman and Greg Brockman, the CEO and president of ChatGPT maker OpenAI, respectively, recently took to X.com to publicly discuss the recent departure of Jan Leike, arguably the company’s former top safety officer.
As Cointelegraph recently reported, Leike was the head of alignment. He resigned his post on May 17 citing irreconcilable differences with the company’s leadership.
Among Leike’s allegations was the assertion that “safety culture and processes have taken a backseat to shiny products” at OpenAI.
It didn’t take long for Brockman and Altman to respond, both taking to X.com within 24 hours of Leike’s post to respond.
As for Brockman’s part, he laid out a long and winding post replete with a three-pronged strategy for safety alignment at the company.
He began by expressing gratitude for Leike’s work at the company before disputing claims that OpenAI wasn’t focused on safety.
“First, we have raised awareness of the risks and opportunities of AGI,” Brockman wrote, pointing out that the company has “called for international governance of AGI before such calls were popular.”
“Second, we have been putting in place the foundations needed for safe deployment of increasingly capable systems. Figuring out how to make a new technology safe for the first time isn't easy.”
The final prong of Brockman’s lecture was to mention that “the future is going to be harder than the past. We need to keep elevating our safety work to match the stakes of each new model.”
He further intimated that the company wasn’t following the big tech prescription of moving fast and breaking things. “As we build in this direction, we're not sure yet when we’ll reach our safety bar for releases, and it’s ok if that pushes out release timelines.”
Sam Altman, co founder and CEO, kept his message short but did mention that he’d have more to say on the subject in the coming days:
“He's right.” Altman posted, referring to Leike’s comments, adding “we have a lot more to do; we are committed to doing it.”
Related: Reddit shares jump after-hours on OpenAI data-sharing deal