Google's AI Code Crisis: Brin Returns to Rally DeepMind Against Anthropic
Google DeepMind is in a state of mobilization. According to three people familiar with the matter, the unit has assembled a dedicated team of researchers and engineers to address a growing...
Google DeepMind is in a state of mobilization. According to three people familiar with the matter, the unit has assembled a dedicated team of researchers and engineers to address a growing problem: its AI models for writing software are falling behind. The effort, led by research engineer Sebastian Borgeaud, represents a direct response to the superior performance of Anthropic’s Claude and pressure from OpenAI.
Sergey Brin, Google’s co-founder, is personally involved. He has instructed staff to move quickly, emphasizing the need to advance in "agentic AI"—systems that can manage multi-step tasks such as constructing entire software applications. Brin views progress in this area as a stepping stone toward more self-improving artificial intelligence.
The urgency stems from a tangible performance gap. DeepMind researchers acknowledge that Claude currently produces better code than Google’s Gemini. While Google says AI assists with roughly half of its new code, some competitors are approaching near-total usage for certain projects. Internally, Google tracks engineer adoption of its coding tools, but uptake has been inconsistent, with many developers using them only for simpler tasks.
The new team is focusing on complex challenges, like interpreting large, existing codebases and generating complete files. They have a significant asset: access to more than two billion lines of Google’s own proprietary code for training, a dataset of unmatched scale. The mandate is to move beyond assistance and create AI that can function as a primary developer.
Recent internal tests, including a tool known as "Agent Smith," signal Google’s direction. However, public benchmarks continue to show rivals with an edge in autonomous execution. With Brin actively steering the effort, the project has become a top priority. The outcome will signal whether Google can leverage its vast resources to reclaim a lead in a field that has become central to the future of AI.
Source: Webpronews
Ready to Modernize Your Business?
Get your AI automation roadmap in minutes, not months.
Analyze Your Workflows →