Meta Layoff Aftermath: Tian Yuandong Discarded, Yao Shunyu and Others Rush to Recruit
2025-10-24 — Beijing
A single stone stirs up a thousand waves.
---
Meta Layoffs Ripple Through the AI Industry
Meta’s latest round of layoffs in its Artificial Intelligence division has sent shockwaves through the tech world.
Key Facts
- ~600 positions cut in AI, affecting:
- FAIR research group
- AI product teams
- AI infrastructure units
- Tian Yuandong’s team completely disbanded — reportedly in one move by Alexandr Wang.
- Full story: Meta AI Major Layoffs—Did They Really Cut Tian Yuandong’s Team?
Tian commented on Twitter that it was “awkward to acknowledge being laid off,” but joked that he “might be the least afraid person.”
Many are still baffled, and Tian himself expressed visible frustration.
Speculated Reasons
- Internal politics within the TBD Lab — including “succession battle” rumors.
- Accusations that Alexandr Wang was “burning the bridge after crossing it.”
- Responsibility assigned to cut teams for:
- Llama 3 inference model results
- Disappointing Llama 4 performance
An online critique claimed:
> “Meta had over 600 people but failed to deliver an open-source reasoning model based on Llama 3. Meanwhile, DeepSeek’s R1 paper demonstrated inference on Llama 3. Embarrassing. That’s why they all had to go.”
The post has since been deleted. Tian replied with “😅”, noting he did the “hard and dirty work” only to be discarded.
---
Signals of Deeper Structural Chaos
Strained Organization
- Frequent team reshuffles and firefighting measures within Meta’s research entity.
- Battles between FAIR and product-oriented groups over compute resources.
- Superintelligence Lab absorbed an already bloated structure.
Strategic Implications
- Layoffs appear both cost-cutting and a power consolidation move for Alexandr Wang’s AI strategy.
- CNBC reports Superintelligence Lab headcount is now around 3,000.
Layoff Protocol:
- Departure date: November 21.
- “Non-working notice period” — no work required, internal systems access revoked.
- Severance: 16 weeks base, plus 2 weeks/year of service (minus notice period).
- Tian received 8 months’ salary.
Such outcomes may be premeditated — in July, Meta increased 2025 expense guidance to $114–$118B and forecasted AI spending growth to accelerate in 2026.
---
AI Industry Talent Grab
Researcher Jiaxun Cui likened Meta’s moves to Squid Game — ruthless elimination. Many new PhDs saw accounts shut down abruptly.
Tian reposted and apologized, revealing he had recommended Jiaxun for Meta’s RL team. Jiaxun never worked on RL, instead enduring “endless departmental reshuffles” before being laid off.
---
Opportunities Emerge
The layoffs may become a recruitment opportunity for tech companies and startups globally.
Example:
AiToEarn官网 — An open-source global AI content monetization platform for creators:
- Generate, publish, and earn from multi-platform content.
- Simultaneous publishing to:
- Douyin, Kwai, WeChat, Bilibili, Rednote (Xiaohongshu), Facebook, Instagram, LinkedIn, Threads, YouTube, Pinterest, X (Twitter).
- Integrated tools: content generation, analytics, AI model ranking.
- Resources: AiToEarn博客, AI模型排名
Tian’s tweet has surpassed 3M views, with countless job offers in the comments.
Other researchers’ job-hunting posts are drawing recruiters fast.
---
Notable Moves
- Yao Shunyu joins the “online hiring” trend. Previously rumored to have a “100 million” salary to join Tencent.
- xAI tweets: “Look at me, look at me — you’re welcome to join!”
Recruitment posts may even outnumber layoffs, proving top talent never lacks opportunity.
---
Reflections on FAIR and Meta’s AI Shift
Xie Saining noted:
> “Every gathering must eventually come to an end.”
Meta’s wave of layoffs hit premier AI researchers hard.
FAIR — founded by Meta (then Facebook) in 2013 — has:
- Pioneered key AI developments.
- Released PyTorch in 2016, opening doors to modern deep learning.
By 2025:
- Large model competition intensified.
- Transition from open-source to commercial closed models was foreseeable.
Hope remains that Meta can preserve an academic research unit within its commercial framework.
Reference: CNBC Report
---
Final Note
As the AI industry evolves, independent creator platforms play a vital role.
AiToEarn官网 empowers creators to:
- Publish across global platforms efficiently.
- Monetize creative efforts with AI-enhanced productivity.
- Diversify reach beyond traditional corporate labs.
---
If you want, I can produce a condensed executive summary table of this article so stakeholders can grasp the full situation in under a minute. Do you want me to do that?
