Google Restructures AI Coding Task Force to Close the Gap With Anthropic After Key Researchers Defect
Google is restructuring its recently established AI coding task force in an effort to narrow the widening technical gap with Anthropic, according to sources familiar with the matter.
The task force, formed just a few months ago, is shifting its training strategy to improve the model’s code generation capabilities while also strengthening its ability to produce other outputs such as presentations and slides, the sources said. The reorganization both formalizes what was originally a temporary team structure and broadens its mandate beyond pure coding.
The restructuring comes on the heels of high-profile departures that have shaken Google’s AI talent pool. Jonas Adler, who had been working on Google’s AI programming initiatives, and Alexander Pritzel, who was involved in training AI systems, are both planning to leave the company. Both researchers are widely regarded as key contributors to the Gemini model family — and both are heading to Anthropic.
They are not the only notable exits. Nobel laureate John Jumper has also jumped ship to Anthropic, and Noam Shazeer, co-lead of the Gemini project, is set to join OpenAI. The brain drain underscores the fierce competition for AI talent among frontier labs and the growing gravitational pull of Anthropic, whose Claude models have earned a strong following among developers.
Adding to the pressure on Google, the company’s next-generation frontier model — widely referred to as Gemini 3.5 Pro — has been delayed until July, according to a report from Business Insider. Google had previewed the model at its I/O developer conference in May, where CEO Sundar Pichai said it would arrive “next month.” That timeline has now slipped.
The convergence of talent losses, model delays, and an organizational pivot suggests a company under significant strain to maintain its position in the rapidly evolving AI arms race. With Anthropic and OpenAI both gaining ground in AI-assisted programming and general-purpose reasoning, Google’s restructuring may be its most consequential move yet in defending its AI ambitions.
