ChatGPT co-creator Shenjia Zhao resigns from Meta within days of joining the company, hinting at a return to OpenAI

Shengjia Zhao , co-developer of OpenAI's ChatGPT, has reportedly resigned from Meta after just a few days at the company and offered to return to his former employer, OpenAI, a move that has been criticized as showing signs of confusion in the largest reorganization of Meta's AI division in its history, led by CEO Mark Zuckerberg.

Zuckerberg's AI hires disrupt Meta with swift exits and threats to leave
https://www.ft.com/content/110786e7-6443-4dff-adee-0ac02c55aaa6

What's really happening with the hires at Meta Superintelligence Labs | The Verge
https://www.theverge.com/ai-artificial-intelligence/767746/meta-ai-superintelligence-lab-departures-scale-zuckerberg-memo

Zuckerberg has been on a massive talent hunt for months to gain an advantage in the AI field, aiming to build a 'personal superintelligence for everyone.' He's poached prominent researchers from competitors like OpenAI and Apple, promising them huge contracts and access to vast computing resources. As part of a $14 billion investment, Zuckerberg brought on Alexander Wang, the former CEO of Scale AI, to lead the company's new AI research division, Meta Superintelligence Labs (MSL).

Meta plans to build an AI lab led by Scale AI CEO Alexander Wang, specializing in the pursuit of 'superintelligence,' and is offering hundreds of millions to tens of billions of yen in compensation to lure AI researchers from OpenAI and Google.

Zhao, who has been involved in the development of a number of groundbreaking products, including ChatGPT, the large-scale language model (LLM) GPT-4, and OpenAI's first AI inference model, o1, has also been appointed as a Principal Investigator at MSL.

Meta appoints former OpenAI employee Shenjia Zhao as chief scientist for its new AI research division, Meta Superintelligence Labs (MSL) - GIGAZINE

According to the Financial Times, Zhao reportedly quit Meta shortly after joining the company and signed employment documents to return to OpenAI, but shortly after that, Meta appointed him as its new Chief AI Scientist.

It's unclear why Zhao wanted to leave Meta so soon, but the Financial Times speculates that it was due to conflicts between the new recruits and the longtime Meta staff.

According to sources, these radical changes have created internal friction, with new executives like Zhao trying to prove themselves while longtime executives are being sidelined. As a result, Meta has seen a string of short-term departures, including researchers quitting just a few weeks after joining and a former OpenAI researcher who completed the onboarding process but didn't show up on his first day. Among the recent departures appear to be veterans with nine or ten years of experience at Meta.

Meanwhile, some new hires have expressed frustration with Meta's bureaucratic culture and internal competition for access to promised computing resources. Wang's leadership style, lack of management experience at a major tech company, has also caused friction, the Financial Times reported.

The turmoil isn't limited to personnel. Meta's flagship model, the Llama Behemoth, didn't perform as expected, and active development toward a public release was halted. Wang's new division, the To Be Determined (TBD) Lab, is now focused on building newer, more cutting-edge models.

Meta plans to release the massive AI model 'Llama 4 Behemoth' with a total of 2 trillion parameters, but development has been delayed until fall 2025 - GIGAZINE

In response to this situation, Meta shared a memo with management announcing that it would temporarily suspend all MSL hiring except for business-critical roles. Meta explained that this was a measure to carefully plan its workforce for 2026, and is typical for organizations that have experienced rapid growth.