AI startup Core Automation recruits researchers from Anthropic and DeepMind for new neolab

A new AI startup called Core Automation, founded by former OpenAI researcher Jerry Tworek, has quickly made waves by recruiting top talent from leading labs like Anthropic and Google DeepMind, in a competitive maneuver dubbed "nerdsniping." According to Business Insider, the company, launched just weeks ago, has already poached researchers from these elite organizations, intensifying the talent wars in the AI industry. This move highlights the growing fragmentation of expertise as startups lure away specialists from established players to pursue bold new directions.

Core Automation's aggressive hiring strategy comes amid a surge in so-called "AI neolabs," independent ventures aiming for breakthroughs that bigger firms might overlook. The Information reports that Tworek's startup is targeting between $500 million and $1 billion in funding, signaling ambitious plans to scale rapidly. Business Insider notes that "nerdsniping"—a term for intellectually captivating experts into switching teams—has allowed Core Automation to assemble a team with direct experience from the forefront of AI development, potentially accelerating its progress in core automation technologies.

This talent exodus underscores the fierce competition for human capital in AI, where researchers from Anthropic and DeepMind bring specialized knowledge in areas like alignment, reasoning models, and agentic systems. At Anthropic, internal surveys reveal engineers are already using their own AI tools like Claude for complex tasks such as debugging and code design, with productivity gains up to 50% and increasing autonomy in workflows. Meanwhile, Google DeepMind is advancing robotics intelligence through models like Gemini Robotics 1.5, enabling machines to handle multistep tasks with step-by-step reasoning, as detailed in Live Science.

The implications extend beyond individual labs, affecting the broader AI ecosystem. Startups like Core Automation are betting on human-led innovation to redefine automation, potentially reshaping business operations from incremental tools to fully autonomous systems, as Google DeepMind executives predict a structural overhaul in workflows over the next one to two years. Researchers jumping ship could accelerate this shift, introducing fresh perspectives on agentic AI—systems that learn domains like humans or execute long-horizon tasks independently.

For those affected, the moves mean disrupted teams at Anthropic and DeepMind, where talent flows heavily toward newcomers; dev.ua reports Anthropic pulling engineers from OpenAI at an 8:1 ratio and from DeepMind at 11:1. What happens next remains fluid: Core Automation's funding push could fuel rapid prototyping of new AI agents, challenging incumbents. Investors and companies alike are watching closely, as these neolabs position themselves to capture the next wave of AI dominance in an industry where human expertise remains the scarcest resource.