Claude Code astonishes Google: AI generated a year-long project solution in one hour, showing rapid evolution of AI coding assistants.

A recent unexpected event when testing Claude Code was given by Jaana Dogan, a principal developer at Google who works on the Gemini API. According to her, Claude Code created a solution in roughly an hour after receiving a brief, three-paragraph problem description from the tool. This solution closely resembled the work her Google team had been doing for nearly a year.

The assignment concentrated on intricate systems that oversee and control several AI bots; her team had investigated numerous concepts in this field but had not yet produced a single design.

Demonstrating the Rapid Evolution of AI Coding Assistants, Claude Code

Building distributed agent orchestrators, which enable various AI agents to collaborate effectively, was the challenge, according to Dogan. She clarified that since last year, Google has been experimenting with a number of strategies, and internal teams were still divided over the ultimate course of action.

In order to adequately evaluate Claude Code, she made a reduced version utilizing only principles that were available to the public, avoiding the use of any private information.

Also read: LG CES 2026: Wallpaper OLED & Gallery TV Make a Stunning Comeback

Even with this little input, Claude Code produced a design that was very similar to what Google had done internally. Dogan communicated that the final product was not flawless and still required improvement.

But the result’s speed and quality, she said, were remarkable enough to alter people’s perceptions of AI coding tools. Since the true worth of these systems is revealed in areas where they already possess a solid technical understanding, she urged skeptics to test them there.

Claude Code Stirs Up Discussion About AI Coding Assistants’ Future

Dogan stated that Claude Code is only permitted for open-source projects and not for internal development when asked if Google utilizes it internally. In response to inquiries on when Gemini might attain a comparable level, she stated that her team is putting a lot of effort into both the models and the supporting systems.

Furthermore, Dogan emphasized that advancements in AI are not a zero-sum game. She stated that it’s critical to recognize competitors’ hard work and that Claude Code’s performance inspired her.

The speed at which AI coding has progressed—from writing single lines of code in 2022 to writing complete sections in 2023, multi-file projects in 2024, and now entire codebases in 2025—was something she considered. Millions of people saw her message, and many commented on how AI may boost individual creativity and lessen organizational friction.