AI researchers continue to leave Google for its rivals

The departure of top AI researchers Jonas Adler and Alexander Pritzel from Google to Anthropic is the latest signal of a fundamental shift in the AI talent landscape. Following high-profile exits like Noam Shazeer (co-author of the Transformer paper) and John Jumper (lead of AlphaFold), this trend is no longer a trickle but a steady current. For an industry that prizes intellectual capital above all else, these moves carry significant weight.

What Happened

Adler and Pritzel were senior research scientists at Google DeepMind, contributing to foundational work in reinforcement learning and generative models. Their move to Anthropic, an AI safety-focused startup backed by billions in funding, mirrors a pattern: top-tier talent is leaving the established tech giant for smaller, more agile competitors. This is not about compensation alone—Anthropic offers competitive packages—but about research autonomy, mission alignment, and the opportunity to shape a company’s technical direction from an earlier stage.

Why It Matters

First, this is a brain drain from a company that has historically been the epicenter of AI research. Google’s “AI-first” strategy depends on retaining the minds who invented the very architectures powering today’s boom. Losing Shazeer (who co-invented the Transformer) and Jumper (who solved protein folding) is not just a PR hit—it directly weakens Google’s ability to lead in next-generation breakthroughs.

Second, the destination matters. Anthropic is explicitly building toward safe, interpretable AI systems. When researchers choose that mission over Google’s broader commercial portfolio, it signals that the “safety vs. capability” debate is now a real career decision, not just a conference panel topic. This could accelerate a bifurcation in the industry: one camp racing for scale, another prioritizing alignment.

Third, it raises questions about Google’s internal culture. Reports of bureaucratic friction, shifting priorities, and a slower pace of publication compared to startups have circulated for years. These exits suggest that even with vast resources, a large organization can struggle to provide the focused, high-agency environment that top researchers crave.

Implications for AI Practitioners

For engineers and researchers in the field, this trend offers a clear lesson: talent mobility is at an all-time high. The barriers to moving between big tech and well-funded startups have collapsed. Practitioners should evaluate not just salary, but the research culture, compute access, and mission clarity of their employers. The “Google resume stamp” is still valuable, but it no longer guarantees access to the most exciting work.

Additionally, the concentration of talent at a few firms (Anthropic, OpenAI, and a handful of others) means that the next wave of foundational models will likely come from outside Google. Practitioners should track where former Google stars land—it is a strong signal of where the next paradigm shift may originate.

Key Takeaways

• Talent is voting with their feet: Top researchers are leaving Google for startups like Anthropic, prioritizing mission and autonomy over brand prestige.
• Google’s AI dominance is eroding: Losing key inventors of foundational technologies weakens its long-term competitive position.
• Safety-focused labs are becoming talent magnets: The “alignment” camp is now a viable career path, not just an academic niche.
• For practitioners, mobility is a strategic asset: The current market rewards researchers who can identify high-agency, well-resourced environments beyond the traditional tech giants.