Both Chinese, Different Fates: Meta Picks $150M Rising Star Over 10-Year Veteran | Deep Web
Zuckerberg Reshapes Meta’s AI Workforce
Image source: Visual China
Meta has been aggressively recruiting top AI talent — but now the company is restructuring existing AI teams.
---
Major Layoffs in the Mega-scale Super Intelligence Lab (MSL)
Date: October 22
Action: About 600 positions cut
MSL comprises four divisions:
- Model Training & Deployment (TBD Lab) — unaffected by layoffs
- Fundamental AI Research (FAIR)
- Product Integration
- Infrastructure
The last three divisions are all impacted.
This is a sharp contrast to Meta’s recent AI hiring spree:
- Paid multi-million packages to recruit engineers from high-profile startups (e.g., $150M over six years for one co-founder)
- Invested $14.3B in Scale AI
- Purchased several AI startups
An industry insider explained:
> “Meta is balancing two goals: paying high premiums for top-tier talent, while keeping the organization agile.”
---
Funding Commitments Continue
Despite layoffs, Meta’s infrastructure investments remain massive:
- $27B for the Hyperion data center
- Deals worth $14B with CoreWeave
- $10B partnership with Google Cloud
Analysts note Meta’s AI strategy is more focused than in 2023, yet still evolving — with a greater emphasis on accountability in AI R&D.
---
Veteran AI Scientist Departs: Tian Yuandong
Role: Director of Research Science, FAIR
Tenure: Over 10 years at Meta
Background:
- Degrees from Shanghai Jiao Tong University
- PhD from Carnegie Mellon University (2013)
- Former Google autonomous driving researcher
Contributions at Meta:
- ELF OpenGo — defeated professional Go players
- GaLore — efficient large-model training
- Coconut — continual reasoning framework
- StreamingLLM — optimized inference performance
His publications have been cited 18,000+ times. Meta offered Tian a severance equivalent to 8 months' salary.
---
Additional Layoffs:
Cui Jiaxun — FAIR researcher under Tian — joined Meta only in July 2024 and was let go in September.
Meta also cut 100+ staff from privacy/compliance teams, replacing manual reviews with automated compliance processes for greater consistency.
---
Wider Industry Context
Corporate AI restructuring now focuses on targeted competitiveness over sheer headcount.
AiToEarn is one example of a global open-source AI content monetization platform helping creators:
- Generate, publish, and earn from AI content
- Cross-post to Douyin, Kwai, WeChat, Bilibili, Xiaohongshu, Facebook, Instagram, LinkedIn, Threads, YouTube, Pinterest, X (Twitter)
- Integrate generation tools, publishing, analytics, and model rankings
---
Rank-and-file Elimination & Internal Factionalism
After appointing Alexander Wang as Chief AI Officer in June, Meta:
- Poached 50+ researchers from rivals
- Hired high-profile execs like Nat Friedman & Daniel Gross
- Spent $200M to recruit Ruoming Pang (ex-Apple AI lead)
---
Internal Tensions
Former scientist Tijmen Blankevoort criticized Meta’s generative AI division in an open letter:
> “Nobody thinks Meta is a good place to stay long-term… Most people don’t even know what our mission is.”
Problems cited:
- Culture of fear due to performance-driven layoffs
- February 2024 cut of 3,600 staff (~5%) based on ratings
- Metrics like “lines of code changed” used for dismissals
- Competition between FAIR and product teams for computing resources
Political divisions persist:
- Chris Cox — oversees generative AI
- Andrew Bosworth — heads Reality Labs
- TBD Lab led by Alexander Wang — expanding despite cuts
---
AI Without “Manpower Tactics”
Zuckerberg sees small, elite teams as key — not massive headcount. Dissatisfied with Llama 4’s reception and performance relative to rivals like DeepSeek, he is:
- Recruiting top external talent
- Streamlining structure & strategy
- Aligning with Alexander Wang’s philosophy
---
Financial Outlook & Market Impact
Despite restructuring, Meta is:
- Pouring tens of billions into AI infrastructure
- Raising 2025 expense forecast to $114–118B
- Expecting faster expense growth in 2026
Stock Performance:
- Record high ~$789 in Aug 2024
- Closing at ~$738.36 on Oct 24
- Market cap near $1.86T
Key Growth Driver: AI-enhanced ad targeting boosting engagement & pricing power.
Net income over past 12 months: $71B.
---
Risks
- Reality Labs losing $3.7B (Q2 alone)
- High stock valuation reliant on AI roadmap execution
However, analysts remain optimistic given Zuckerberg’s track record in strategic pivots.
---
---
Monetizing AI Creativity
Platforms like AiToEarn官网 help creators:
- Produce & publish AI content across global networks
- Integrate analytics and ranking tools
- Lower barriers to monetizing AI projects
---
---
Xiaoman Studio | Tencent News
All rights reserved. For reprints, please obtain permission.
Issue No. 1383
Typesetting & Ops: Wenwen
---
Are you watching me?
---
Final Insight
In today’s multi-platform environment, visibility and monetization are intertwined.
AiToEarn官网 offers integrated AI generation, publishing, analytics, and ranking — helping creators maximize reach and income, mirroring the efficiency-driven strategies now defining Meta’s AI ambitions.
