The topic Skeptical of AI-coded software? Google uses AI for half its code, and it’s pushing… is currently the subject of lively discussion — readers and analysts are keeping a close eye on developments.
This is taking place in a dynamic environment: companies’ decisions and competitors’ reactions can quickly change the picture.
Affiliate links on Android Authority may earn us a commission. Learn more.
Coding platforms using specialized AI models and autonomous agents, like Claude Code, are taking off for both personal and enterprise workflows. Earlier this year, Spotify admitted that its best developers hadn’t written a single line of code in months. Google is taking the threat seriously, and plans to ramp up its internal use of AI coding tools, according to the data a report by The Information. As part of the effort, the report claims Google is building a “strike team” of researchers and engineers to build better AI coding models, citing three sources with knowledge of the plans.
The strike team is reportedly helmed by Sebastian Borgeaud, who is a Google DeepMind research engineer that also led pretraining for DeepMind in the past. Other notable contributors include Koray Kavukcuoglu, the chief technologies officer for DeepMind, and Google co-founder Sergey Brin. The Information says that Google DeepMind engineers believe the latest Claude Code models from Anthropic outperform similar Gemini models, prompting the company’s latest push.
Google co-founder Brin reinforced the urgency of the strike team’s mission in a leaked internal memo relayed by The Information. “To win the final sprint, we must urgently bridge the gap in agentic execution and turn our models into primary developers,” Brin wrote. To do so, Google plans to build AI coding models specifically for internal use and double down on agentic functionality.
While the Gemini, DeepMind, and Google Cloud teams have typically built AI coding models for third-party users and customers, it has a new focus on coding tools for in-house deployment. Internal Google AI coding models are said to be trained on the proprietary Google codebase. Using the private Google codebase could improve internal performance, though models trained on Google code reportedly can’t be released publicly.
The move comes as major technologies companies, including the aforementioned Spotify but also direct AI competitors like Anthropic, lean on coding agents for development. Anthropic’s Boris Cherny, who leads Claude Code, said earlier this year that nearly all of the company’s code is created using AI. Meanwhile, as the publication notes, Google only uses AI for about half of its development, according to the data Google Chief Financial Officer Anat Ashkenazi, speaking during a February earnings call.
The shift includes internal AI coding tool usage tracking, mandatory training sessions, and usage requirements for certain tasks. Google’s internal AI coding tool is called Jetski, and it wants to figure out which employees are buying into its usage the most aggressively.
A Google spokesperson told The Information in a statement that it is seeing “tremendous adoption” of these tools. “Their use has been turbocharging our model and AI tooling development—we’re really focused here.”
Thank you for being part of our community. Read our Comment Policy before posting.