Y’all, there’s trouble in Meta’s AI paradise.
Just two months after Mark Zuckerberg launched the Superintelligence Lab—(Meta’s moonshot project to “win” the AGI race)—the star-studded lineup already seems to be losing its shine.
Top researchers are bailing, big names are switching sides, and what was supposed to be a powerhouse team is starting to look more like a revolving door.
Here’s what’s going down:
Three top researchers peaced out. Two of them—Avi Verma and Ethan Knight—literally went back to OpenAI, the rival they ditched for Meta. Awkward, right? The other, Rishabh Agarwal, also resigned, citing a desire to “take on a different kind of risk,” which (if we’re being honest) sounds more like a polite way of saying “nah, I’m good.
Chaya Nayak, once Meta’s head of generative AI product, bailed—and yep, she’s now at OpenAI on ‘special initiatives.’ Ouch.
Shengjia Zhao (one of ChatGPT’s original minds) is still chief scientist at Meta, but morale? Let’s just say it’s looking shaky.
Meanwhile, Meta’s AI org just got chopped into four teams. Officially, this is “streamlining.” Unofficially, it screams internal turbulence.
And all this is happening even as Meta goes full tech-bro recruiter—tossing nine-figure salaries and cutting flashy deals with AI startups like Midjourney to fast-track growth.
Even Sam Altman called it ‘distasteful.’ And honestly? we get why. If your rival’s buying up your best people with big checks, the vibes are way off.
But this isn’t just a Meta problem. It’s a tech culture moment. The AI race is no longer just about GPUs and model size—it’s a people problem. And right now, Meta’s struggling to hold onto the very people it paid top dollar to attract.
Our Take:
Meta came out swinging, promising to build the Avengers of AI. But talent doesn’t always follow money—it follows vision. Right now, Meta feels like a giant experiment searching for its plot, while rivals like OpenAI project momentum and mission.
For tech leaders and startups, here’s the memo: AI breakthroughs aren’t born from fat checks alone. They come from teams that feel stable, aligned, and excited to be there.
If Meta doesn’t figure that out fast, all its hype and nine-figure packages risk turning it into the most expensive training program for OpenAI hires.
You can read more on this here.
