Meta Poaching OpenAI Talent via Face Detection on Livestreams

Description

A Twitter/X screenshot of a post by Yuchen Jin (@Yuchenj_UW) with a joke aimed at Sam Altman: 'Pro tip for Sam: Never show an Asian face on the OpenAI livestream. Meta's running a face detection on every frame. If it spots an Asian, 9-figure offer is instantly sent from Zuck's personal email.' The post quote-tweets Alexandr Wang (@alexandr_) from Scale AI welcoming Hyung Won Chung, @_jasonwei, and @EdwardSun0909 to the team, noting 'some may recognize them from various recent livestreams :)'. Hyung Won Chung's embedded post announces that after a great time at OpenAI, they recently joined @Meta Superintelligence Labs. A group photo shows three people standing together outdoors

Common Tags

AIResearch BigTechCompanies Openai TechRecruiting FAANGCompanies AIIndustryTrends JobMarketTrends CareerGrowth TechIndustryHumor SocialMedia

Comments

Anonymous ★ Top Pick Meta's recruitment pipeline: OpenAI livestream -> face detection model -> Zuck's personal Gmail -> 9-figure offer. Latency: 3ms. The only AI product at Meta that actually ships on time

Anonymous ★ Top Pick

Meta's recruitment pipeline: OpenAI livestream -> face detection model -> Zuck's personal Gmail -> 9-figure offer. Latency: 3ms. The only AI product at Meta that actually ships on time
Anonymous

Meta's new face detection model has a single classification layer: `is_poachable_openai_researcher`. It's the only model they have that's deployed straight to Zuck's outbox without a single PR review
Anonymous

Proof that Meta’s computer-vision pipeline has only two classes: background noise and ‘worth a blank-check offer.’
Anonymous

The real reason Meta pivoted to open-source AI wasn't philosophical - it was just cheaper than their previous strategy of buying every engineer who appeared in an OpenAI demo for more than their Series B valuation
Anonymous

The joke brilliantly captures the absurdist reality of AI talent wars where companies deploy the very technology they're building - real-time computer vision and automated decision systems - to poach each other's researchers. It's a meta-commentary on how the tools of AI research (face detection, automated triggers, instant notifications) could theoretically be weaponized for recruitment at scale. The '9-figure offer from Zuck's personal email' hyperbolically references both the astronomical compensation packages in AI research and the direct CEO involvement in recruiting top talent, a practice that's become standard at the executive level when competing for researchers who can advance AGI timelines. The underlying truth: in 2024's AI landscape, showing your team publicly is essentially painting targets on their backs for competitors running sophisticated talent intelligence operations
Anonymous

Kafka -> face-embedder -> Lambda(OfferService, amount=1e9): event‑driven recruiting with an SLO tighter than your stream’s P99 inference latency
Anonymous

OpenAI's face detector: superhuman accuracy from overfitting to the ICML best paper author distribution
Anonymous

Somewhere in PeopleOps: Kafka(openai.livestream.frames) -> CLIP embeddings -> ANN search against 'known_wizards' -> if similarity > 0.98 then OfferService.send(9_figures); finally, an HR system with a lower latency SLO than production
@volkov_s 1y

I won't believe any AI can distinguish asian faces
1. @deadgnom32 1y
  
  why not?
2. @deadgnom32 1y
  
  machine learning did it very well long before any GPTs came out. don't see a reason why they wouldn't now
3. dev_meme 1y
  
  Yeah, Chinese also confused how similar all people of your race look like 🌚
4. dev_meme 1y
  
  In 2018 solo face recognition was working perfectly (99%+ success) with only parts of face being visible That made protest in HK in 2020, where people used mask, especially absurd since they all were easily identifiable as long as eyes and oval of face is visible (which is always a case during daylight)
  
  @Box_of_the_Fox 1y
  
  You have some link where I could read more about it?