- Added details about Meta hiring two leading AI researchers from OpenAI
- Added details about Meta hiring four leading AI researchers from OpenAI
- Added Bosworth's statement
- Added Beyer's statement
Update July 16, 2025:
Jason Wei and Hyung Won Chung, two prominent AI researchers, are leaving OpenAI to join Meta’s superalignment team, according to multiple sources cited by Wired. Wei previously worked on OpenAI’s o3 model and its Deep Research group, with a focus on reinforcement learning. Before that, he contributed to chain-of-thought research at Google, which helped lay the foundation for improved reasoning in large language models.
Chung also specialized in reasoning and agent-based systems and was involved in OpenAI’s o1 model. According to his online profile, his work focused on structuring decision-making processes in AI systems.
The two researchers had already worked closely together at Google and made the move to OpenAI as a pair in 2023. Their internal Slack accounts at OpenAI have since been deactivated.
Meta’s superalignment team is focused on building advanced AI systems that go beyond the capabilities of today’s models. The company has previously recruited coordinated research teams from both OpenAI and Deepmind to accelerate progress.
Update July 2, 2025:
Wired reports that Meta is offering top AI researchers compensation packages worth up to $300 million over four years, with more than $100 million paid out in the first year alone, according to multiple sources with direct knowledge of the deals.
At least ten of these offers have reportedly gone to OpenAI employees. The packages include stock grants that vest immediately and are primarily targeted at leadership roles within Meta's newly formed superintelligence team.
One OpenAI staffer employee told Wired that "that's about how much it would take for me to go work at Meta." Others questioned whether they would have as much influence at Meta as they do at OpenAI, despite the massive pay packages.
Update June 29, 2025:
OpenAI loses four more top researchers to Meta as even its own engineers call it a "huge loss"
Meta has hired four more top AI researchers from OpenAI. Jiahui Yu, Hongyu Ren, Shuchao Bi, and Shengjia Zhao previously worked on core models like GPT-4.1, o3, o1, and handled multimodal post-processing at OpenAI.
With these hires, Meta is bringing in key expertise for the development of Llama 5, if it moves forward, as well as for its new superintelligence team. The four are considered some of the most experienced researchers in OpenAI's ranks. There are currently no details about the circumstances of their move or their future roles at Meta.
In a now-deleted tweet, OpenAI engineer Cheng Lu called the departures a "huge loss."
"Not too many people outside the company know how talented and hardcore they are. Such a huge loss to OpenAI and I feel really disappointed that the leadership didn't keep them," Lu wrote.
Update June 27, 2025:
Meta CTO confirms massive offers for top AI executives
Meta CTO Andrew Bosworth pushed back on recent statements by OpenAI CEO Sam Altman about 100 million dollar bonuses for AI researchers leaving for new jobs, calling them "dishonest" and "exaggerated" in a leaked internal meeting, according to The Verge.
Bosworth said that while some high offers do exist, they are limited to a small group of top leaders on OpenAI's superintelligence team, and aren't simple cash bonuses but part of more complex compensation deals. He also noted that Altman himself is countering offers to retain talent.
Meta CPO Chris Cox acknowledged that, although Meta AI reaches a billion monthly users, engagement falls short of ChatGPT's numbers. The separate Meta AI app gets just 450,000 daily users, many of whom access it through Ray-Ban smart glasses. Looking ahead, Meta plans to shift its focus toward entertainment and social connection, while rivals like OpenAI, Google, and Anthropic concentrate on productivity. Meta declined to comment on the meeting.
Update:
Meta poaches three top AI researchers from OpenAI, who had poached them from Deepmind
Lucas Beyer has confirmed that he, Alexander Kolesnikov, and Xiaohua Zhai are joining Meta. On X, Beyer addressed rumors about his contract, saying he did not receive a $100 million signing bonus - a figure previously floated by OpenAI CEO Sam Altman, though not specifically in connection with this hiring. Beyer says he will remain in Zürich.
It's unclear whether Altman actually heard those numbers internally or was just trying to disrupt Zuckerberg's recruiting spree by pushing Meta to offer bigger salaries.
Original article:
Meta has hired three top AI researchers from OpenAI: Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai.
All three are known for their work in machine learning and computer vision, including developing scalable image models and the Vision Transformer (ViT) architecture.
They had only recently opened OpenAI's Zurich office in late 2024 and previously worked at Google Deepmind.
According to reports, Meta CEO Mark Zuckerberg personally led recruitment efforts after Meta's latest AI model failed to meet expectations. He reportedly offered compensation packages worth up to $100 million.
Meta also recently invested about $14 billion in AI data labeling company Scale AI, reportedly to attract skilled employees.
