Why Meta’s Hungry for Brainpower - Xist4

October 6, 2025

Why Meta’s Hungry for Brainpower

Another Brain Drain to Big Tech

Last Friday, Andrew Tulloch quietly dropped a message to his team at Thinking Machines Lab: he’s off to Meta. Casual announcement. Big implications.

For those not deep in the AI ecosystem, Tulloch isn’t just another engineer — he’s been at the beating heart of open-source machine learning, from Torch to PyTorch to bootstrapping AI infra at Meta before. So yes, this is a bit like Pep Guardiola returning to Barcelona… if Barça paid him £10M and unlimited GPU credits.

But this isn’t just about one guy moving jobs. Tulloch’s leap tells us something bigger about where the talent market’s moving — and what scale-up founders need to get their heads around if they want to stand a chance of hiring (or keeping) the data and AI minds that drive growth.

The AI Talent War Is Heating Up — Fast

Let’s not sugarcoat it: top-tier AI and data talent is being hoovered up faster than a pub carpet after last orders. Big tech — Meta, Google DeepMind, OpenAI — are throwing everything short of naming rights at researchers and engineers.

You’re not just competing on salary anymore (though, yes, Meta’s rumoured $500K+ packages do make your startup equity pitch feel a bit... optimistic). You’re competing on narrative, tooling, mission, and yes — compute capacity.

Founders tell me all the time: “We’ve got a brilliant product, why wouldn't they join us?”

Here’s the harsh truth: brilliant products don’t hire brilliant people. Brilliant value propositions to brilliant people do. Different game.

Big Tech Plays the Long Game — You Need to Play Differently

Meta isn’t hiring Tulloch for next quarter’s metrics. They’re locking in brains for bets that won’t mature for 3–5 years. They’ve got the luxury of time and capital.

You probably don’t.

So you need a different strategy:

  • Look for high-performers before they’re household names — emerging stars, not just MVPs
  • Offer autonomy, not just stock options — Meta has layers of red tape. You can offer speed and agency
  • Design missions with real-world impact — AI-for-good isn't just a PR trick. It’s a magnetic pull for the right minds

Big Tech is Disneyland for engineers. You’ve got to be the independent film studio — riskier, but creatively irresistible.

This Isn’t About AI — It’s About Status

Tulloch going back to Meta isn’t just about compute firepower. It’s about working on bleeding-edge problems with peers who challenge you. That’s status, baby. And in high-calibre data and AI circles, status matters more than perks.

If your company can’t offer the status of being An Important Name on a CV, offer something else that hits the same nerve:

  • Founding team recognition — be generous with credit. Make their name part of the company lore.
  • Open-source contributions — helping them build a brand in public from inside your business
  • Community visibility — speaking slots, collaborations, even co-authoring posts. Get them seen.

Reputation is currency. Help your hires invest it wisely.

Forget Unicorns — Hire Zealots

I get it: you want a Tulloch of your own. But the moment someone’s CV looks like a Wired Magazine feature — they’re likely already in someone’s golden cage.

The smarter move? Hire your own version of 2017-era Tulloch. Hungry. Sharp. Maybe a little underpriced.

What you’re looking for isn't the unicorn — it’s the zealot:

  • They’re obsessed with the problem, not just the tech
  • They ship fast and ask smarter questions
  • They’d work on this even if they weren’t paid (but do pay them!)

Spot them early. Give them room to build. And don’t screw it up with a bloated interview funnel and six weeks of HR radio silence. (Yes, I’m looking at you.)

So, What Should You Do Now?

Here’s the reality founders and execs need to accept: you won’t outbid Meta. But you can outflank them.

Use Tulloch’s move as your wake-up call to rethink how you hire builders in the age of AI.

Ask yourself:

  • Are we selling our mission to talent like we sell our product to users?
  • Are we spotting raw stars early, or just chasing the headlines?
  • Can we offer more agency, visibility, and purpose than the big dogs?

If you’re stuck — call me. I’ve spent years helping scale-ups build their Data and AI teams without paying a king’s ransom, and without losing the plot in the process. And I promise — no buzzword bingo.

Final Byte

Andrew Tulloch just chose Meta.

You won’t win the war by chasing him — but maybe, just maybe, you’ll win by hiring the next one before he becomes “Tulloch.”

Play smarter. Hire earlier. Think different.



Back to news