🐼 Silicon Valley Drama: Meta & Scale AI’s Billion-Dollar “Marriage” on the Rocks

 If you thought K-pop relationships were short-lived, wait until you hear about Meta and Scale AI. They signed a $14.3 billion deal in June, smiled for the wedding photos… and now it already looks like divorce court.


💸 Meta marries Scale AI… but maybe picked the wrong partner

Meta poured $14.3B into Scale AI, bringing in CEO Alexandr Wang and some top talent to run its brand-new Meta Superintelligence Lab (MSL).
The dream? “Combine data + brains and build the ultimate AI overlord.”
The reality? Within two months, key people started bailing like extras in a horror movie who know the haunted house is about to collapse.


⚡ Internal chaos, Silicon Valley style

  • Ruben Mayer (ex-VP at Scale AI) → joined Meta → quit after two months.

  • Several ex-OpenAI and ex-Scale folks also packed up and left.

  • New recruits complained about steering the Meta Titanic, while old Meta AI staff grumbled about their shrinking influence.
    → Basically, the lab feels less like a cutting-edge AI hub and more like a family business with daily shareholder drama.


📉 Meanwhile, Scale AI is… not scaling so well

  • Started with cheap crowdworkers for labeling data. But today’s AI needs specialized data from doctors, lawyers, scientists → and Scale AI is struggling to keep up.

  • Rivals Surge and Mercor went premium from day one, paying experts well → now they’re sprinting ahead.

  • Lost OpenAI and Google as clients right after the Meta deal.

  • Cut 200 staff in July, with the new CEO calling it a “market shift” (translation: clients ghosted us).


🤔 Did Meta really want Scale AI?

Some analysts think Meta’s move was less about Scale AI’s value and more about poaching Alexandr Wang.
But putting a data-labeling mogul (with no hardcore AI research background) in charge of a superintelligence lab feels a bit like giving your Uber driver the keys to a nuclear submarine.


🐢 Meanwhile at MSL…

  • They’re working on the next generation of Llama models, due later this year.

  • Problem: Llama 4 earlier this year landed with a thud — developers complained it couldn’t code well, reason well, or even follow instructions properly.


👉 Bottom line:

  • Meta: wants to dominate AI, but drowning in talent drama.

  • Scale AI: losing customers, staff, and maybe its shine.

  • Alexandr Wang: probably wondering whether running a shaky startup was easier than being Mark Zuckerberg’s in-house therapist.

Collect by Binlin

Comments

Viewed in recent months

Cambricon Goes Supersonic: 4,000% Revenue Boom, But Nvidia Still Laughs from Its Throne

🏃‍♀️ The 92-Year-Old Grandma Who Runs 200m in Under a Minute!

🌍 Google Spills the Tea: Your AI Prompt Isn’t Free (Environmentally Speaking)

🌲🌿 Brooo… They Found a Whole Weed Forest in California! 🌿🌲

Có nên đầu tư vàng thời điểm này?

👕 From Trendy Tee to Dinner Plate: The Wild Journey of Microfiber Fashion

🌙 Saigon (Vietnam) After Dark: Where the Night Comes Alive!

🧘‍♀️ The 10 Chillest Cities on Earth (2025 Edition)

🐭 New York vs. The Rat Army: High-Tech Showdown

🌦️ La Niña Might Crash This Fall… But Don’t Freak Out