Imagine being the architect of the world's most advanced AI, only to be pulled away by a whisper of a more revolutionary idea. That's exactly what's happening inside the hallowed halls of Anthropic, Google DeepMind, and OpenAI, as a mysterious new player, Core Automation, executes a stunning talent raid.
Founded by ex-OpenAI VP Jerry Tworek, the startup's first public move wasn't a product launch—it was a declaration of war on the status quo. Its goal? To build "the world's most automated AI lab" and create systems that "optimize and automate work, starting with research itself." But its most shocking achievement to date is its recruitment hit-list.
The "Nerdsnipe" That's Reshaping AI
The term "nerdsnipe"—to distract an engineer or scientist with a fascinating problem—has become Core Automation's unofficial recruitment strategy. Rohan Anil, a prominent researcher who worked at both Anthropic and Google DeepMind, publicly credited it for his departure. "Jerry Tworek nerdsniped me into starting this with him and others," he wrote, calling his former employer "one of the best places to work for a researcher."
He's not alone. Anmol Gulati, a research scientist from Google DeepMind who worked on the Gemini model, also announced he was "starting something new," expressing a fundamental disillusionment. "I've increasingly felt that the current research paradigm — scaling models, data, and static deployment won't get us all the way," he wrote, hinting at Core's pursuit of "new learning algorithms" and systems that automate their own creation.
Who Else Has Been Lured Away?
The brain drain doesn't stop there. The team now includes Joanne Jang, a former OpenAI general manager whose X bio simply states she's "trying to automate myself @coreautoai." The exodus also captures former Google DeepMind staffers Ehsan Amid and Avery Lamp, OpenAI's ex-head of people, Julia Villagra, and former research interns from Google and Meta.
On its website, Core boldly states its team consists of those who have "helped build frontier models" and "influential architectures." This isn't just hiring; it's a targeted extraction of institutional knowledge from the very companies defining our AI present.
Why Top Minds Are Betting on the Startup Dream
This marks a new front in the fierce AI talent wars that saw tech giants last year offering **multibillion-dollar acquihires and massive pay packages**. So why leave? According to recruitment experts, startups fight back with more than just competitive salaries.
"Equity is 'the big factor'," explained Shawn Thorne of True Search, helping offset the "opportunity cost" for top researchers. The allure is a potent mix of cofounder titles, access to critical computing power, time for blue-sky research, and the one thing a giant corporation can't offer: **total ownership and impact**.
Core Automation is following a path blazed by legends like Yann LeCun, who left Meta to start AMI Labs, pursuing "world models" over commercial scaling. It signals a growing belief among elite researchers that the next breakthrough won't come from simply making existing models bigger, but from reinventing the entire process of discovery.
For the rest of us, this quiet exodus could be the first tremor of a seismic shift. The race for Artificial General Intelligence (AGI) is no longer just between tech titans. It's now being driven by agile, ideologically-driven startups built by the very people who built the foundations. The labs that created ChatGPT and Gemini are now watching their own creators walk out the door to build what comes next.