Meta's advanced AI ambitions leads to heavy reliance on Chinese researchers

Source Cryptopolitan

Meta Platforms has quietly assembled a powerhouse of Chinese AI researchers to staff its newly announced Meta Superintelligence Labs (MSL).

According to reviewed memo, CEO Mark Zuckerberg tasked this unit with crafting the next wave of large-scale models, promising “personal superintelligence” that anyone could tap into. The move has sparked pride back in China, as many tech watchers note the country’s outsized role in global AI research.

Meta’s AI unit is spearheaded by Chinese talents

Seven of the 11 publicly listed hires at the lab, not including former Scale AI CEO Alexandr Wang and former GitHub CEO Nat Friedman, who were tapped to lead the unit, are from China: Bi Shuchao, Chang Huiwen, Lin Ji, Ren Hongyu, Sun Pei, Yu Jiahui and Zhao Shengjia.

According to Chinese local media, these Zuckerberg recruits graduated from some prestigious Chinese universities, including the University of Science and Technology of China, before furthering their studies and careers in the US.

Of the new hires at Meta, four of them are reportedly from Beijing’s Tsinghua University, which some call the equivalent of the Massachusetts Institute of Technology (MIT).

The influx of Chinese talent into Meta’s new AI unit has sparked widespread discussion in the country’s tech industry, which highlighted the high concentration of mainland talent working in AI globally.

That development was pointed out in May by Nvidia CEO Jensen Huang, who said at the Hill & Valley Forum in Washington that “50 per cent of the world’s AI researchers are Chinese, which should play into how we think about the game.”

Meta ignites tensions with rivals in the sector

According to the SCMP, Meta’s strategy has also heightened tensions within the sector among peers. In a recent podcast, OpenAI CEO Sam Altman revealed that Meta was offering signing bonuses as much $100 million to attract potential recruits from the startup.

Altman also reportedly criticized the social media network firm in a memo to OpenAI staff.

“There will always be some mercenaries,” he reportedly wrote, adding that “missionaries will beat mercenaries.” According to reports, Altman also hinted at re-evaluating the company’s compensation.

Among the team that Meta hired, Chang Huiwen is a graduate of Tsinghua University’s elite Yao Class. Chang earned her PhD at Princeton, focusing on image processing. After internships at Adobe and Facebook, she snagged a Microsoft fellowship in 2016. Chang later joined Google in 2019 before moving to OpenAI in mid‑2023, where she co-developed the advanced image generation features in GPT‑4o.

Another recruit is Lin Ji, who completed his bachelor’s at Tsinghua in 2018 and wrapped up a PhD at MIT by 2023. His résumé boasts internships at Google, Adobe and Nvidia. He joined OpenAI late last year, diving into multimodal reasoning and synthetic data generation.

There is also Sun Pei, who earned a Master’s from Carnegie Mellon after graduating Tsinghua, then started at Google in 2011. He briefly returned to China with Alluxio before joining Waymo in 2017. At DeepMind, he became a principal researcher, helping shape the Gemini AI suite’s reasoning and post‑training pipelines.

According to SCMP, another recruit is Zhao Shengjia, who after a Tsinghua bachelor’s in 2016 and a Stanford PhD in 2022, went straight to OpenAI. He led the synthetic data team and was instrumental in building ChatGPT, the core GPT‑4 model and related mini‑models.

Adding to the list is Bi Shuchao, who studied mathematics at Zhejiang University and earned advanced degrees at UC Berkeley. Launching his career at Google in 2013, he optimized Ads with deep learning and co‑founded YouTube Shorts. In May 2024, he joined OpenAI to head multimodal post‑training, contributing to GPT‑4o’s voice mode and the o4‑mini models.

Also confirming Altman’s assertion is Ren Hongyu – a 2018 Peking University grad, Ren took his PhD at Stanford, interning at Microsoft, Nvidia, Google and Apple along the way. He moved to OpenAI after graduation, leading a team on post‑training efforts for the company’s flagship models.

Then Yu Jiahui, who was from the Special Class for Gifted Young at the University of Science and Technology of China. He completed his PhD at UIUC and his eclectic career includes roles at Microsoft, Megvii, Adobe, Snap, Baidu, Nvidia and Google’s DeepMind, before joining OpenAI in October 2023 to oversee perception development.

KEY Difference Wire helps crypto brands break through and dominate headlines fast

Disclaimer: For information purposes only. Past performance is not indicative of future results.
placeholder
US October PPI Preview: Fed Rate Hike Expectations Rise, PPI May Be Key Variable; How Will US Stocks, Dollar, and Gold Move?On Friday (Nov. 13) Eastern Time, the U.S. will release the October Producer Price Index (PPI). Following the October CPI once again exceeding expectations, the market will look to the PP
Author  TradingKey
12 hours ago
On Friday (Nov. 13) Eastern Time, the U.S. will release the October Producer Price Index (PPI). Following the October CPI once again exceeding expectations, the market will look to the PP
placeholder
Gold closed below $4,000 again — the ETF crowd and central banks stayed, the momentum money leftSpot gold is trading near $3,999, its third consecutive lower daily low, after a 5.00% 10-year yield and a 102.85 dollar index raised the odds of a December hike to 87%. But the selling is not coming from the strategic holders: managed-money net longs fell to 324 tonnes, a 21-week low, and Goldman Sachs attributes roughly 95% of five weeks of selling to non-reportable positions. ETFs still hold a record 4,256 tonnes and central banks bought 321 tonnes in Q3. The table below sets the triggers, directions and levels.
Author  Irene Q.
17 hours ago
Spot gold is trading near $3,999, its third consecutive lower daily low, after a 5.00% 10-year yield and a 102.85 dollar index raised the odds of a December hike to 87%. But the selling is not coming from the strategic holders: managed-money net longs fell to 324 tonnes, a 21-week low, and Goldman Sachs attributes roughly 95% of five weeks of selling to non-reportable positions. ETFs still hold a record 4,256 tonnes and central banks bought 321 tonnes in Q3. The table below sets the triggers, directions and levels.
placeholder
Nvidia slid 7.7% after hours while AMD and Broadcom rose — the AI trade changed hands, it didn't breakNvidia closed Wednesday at 173.73, up 2.87%, then fell as much as 8.1% to 159.44 after reports that its board was preparing for a possible leadership change. The same session, the Philadelphia Semiconductor Index rose 5.7% with all 30 components higher, CoreWeave jumped 24.7% on 168% revenue growth, and Micron added 11.9%. If AI demand were being repriced, AMD and Broadcom would not be bid. This is a governance discount on one name, and 204.54 — the 200-day average — is the level that settles it.
Author  Irene Q.
17 hours ago
Nvidia closed Wednesday at 173.73, up 2.87%, then fell as much as 8.1% to 159.44 after reports that its board was preparing for a possible leadership change. The same session, the Philadelphia Semiconductor Index rose 5.7% with all 30 components higher, CoreWeave jumped 24.7% on 168% revenue growth, and Micron added 11.9%. If AI demand were being repriced, AMD and Broadcom would not be bid. This is a governance discount on one name, and 204.54 — the 200-day average — is the level that settles it.
placeholder
Gold Extends Slide Below $4,000 as Rising US Inflation, Fed December Hike Bets Pressure Prices On November 12, gold prices (XAUUSD) extended their recent losses, falling further below $4,000 in early Asian trading to hit an intraday low near $3,968, marking a third consecutive day
Author  TradingKey
18 hours ago
On November 12, gold prices (XAUUSD) extended their recent losses, falling further below $4,000 in early Asian trading to hit an intraday low near $3,968, marking a third consecutive day
placeholder
Bitcoin lost its 200-day average — but the exit started three sessions before the CPIBitcoin fell 2.5% to about $73,400 on 11 November, an 11-week low, breaking its 200-day moving average near $73,100. The 24-hour liquidation tally reached $632 to $705 million, roughly 82% of it longs. But spot ETF outflows began three sessions earlier and USDT's market cap has shrunk $6.6 billion in 60 days — the leveraged length left before the inflation print did. A double top at 81,890 and 81,080 leaves a measured target of 69,366.
Author  Suzie
Yesterday 03: 22
Bitcoin fell 2.5% to about $73,400 on 11 November, an 11-week low, breaking its 200-day moving average near $73,100. The 24-hour liquidation tally reached $632 to $705 million, roughly 82% of it longs. But spot ETF outflows began three sessions earlier and USDT's market cap has shrunk $6.6 billion in 60 days — the leveraged length left before the inflation print did. A double top at 81,890 and 81,080 leaves a measured target of 69,366.
goTop
quote