Hieu Pham, a former researcher who worked with both OpenAI and xAI, has announced his decision to leave frontier artificial intelligence (AI) labs, citing burnout as the primary reason. Pham shared his plans to return to his native Vietnam to focus on recovery and pursue new opportunities.
Pham’s departure was made public via a post on the social media platform X, where he described his tenure at OpenAI and xAI as a “once-in-a-lifetime experience.” During his time at these organizations, Pham contributed to the development of advanced AI systems intended to significantly improve daily life. He expressed pride in this work but also acknowledged the personal toll the demanding environment had taken.
“I cannot believe I would say this one day, but I am burnt out,” Pham stated. He emphasized the mental health challenges that emerged during his intense work schedule, noting that conditions he once underestimated proved to be “real, miserable, scary, and dangerous.”
Pham began his role at xAI in August 2024 before joining OpenAI a year later. According to his LinkedIn profile, he spent approximately seven months at OpenAI before the announcement of his resignation. He plans to return to Vietnam with his family, intending to explore new pursuits and focus on improving his health.
Pham’s exit occurs amid a broader trend of AI researchers stepping away from fast-paced AI labs. Earlier in the same month, Mrinank Sharma, who led the Safeguards Research Team at Anthropic, also left his position. Sharma highlighted concerns about the rapid advancement of AI and the need for wisdom to match its increasing global impact.
Other notable departures in the AI research community include Dylan Scandinaro, who transitioned from Anthropic to OpenAI as head of preparedness but has publicly acknowledged the risks associated with rapidly developing AI technologies. Discussions about workplace intensity and long hours have drawn comparisons to China’s “996” work culture, which involves working from nine in the morning until nine at night, six days a week.
Senior research scientist Nathan Lambert of the Allen Institute for AI commented that the high-pressure, often self-imposed workload is common in organizations like OpenAI and Anthropic. Many employees embrace these demands motivated by their dedication to AI research.
Pham’s decision to step back underscores growing concerns within the AI community about the personal costs of working in leading-edge labs amid an environment characterized by rapid development and high expectations.
Pham concluded his announcement expressing hope for healing and the possibility of returning renewed: “I hope I will heal. Until then.”
