Jaana Dogan, a senior engineer at Google, has stirred widespread debate in the tech industry after revealing that Anthropic’s Claude Code was able to recreate a complex system architecture in just one hour — a task her team at Google’s Gemini API had spent nearly a year planning.
In a viral post, Dogan explained that she gave the AI a high-level description of a problem her team had been grappling with: building “distributed agent orchestrators,” which she likened to a traffic control system for multiple AI agents working together. Using a prompt of only three paragraphs and avoiding any proprietary information, the AI produced a working prototype that closely mirrored the solution Google engineers had developed over months.
“I’m not joking, and this isn’t funny. We have been trying to build distributed agent orchestrators at Google since last year,” Dogan wrote on X on January 2, 2026. “I provided Claude Code with a description of the problem, and it generated what we developed last year in just an hour.”
She clarified that the AI-generated output was not production-ready and would still require human input and refinement. However, the incident highlighted a growing realization within the tech community: AI tools are increasingly bypassing the lengthy cycles of meetings, approvals, and consensus-building that often slow down large engineering teams.
When asked whether Google uses competing AI tools internally, Dogan said their use is limited to open-source projects. She also reassured followers that Google is actively working on improving its own models and infrastructure to remain competitive.
Dogan’s post, which garnered more than 4 million views, underscores how rapidly AI coding capabilities have evolved — from generating small code snippets in 2022 to restructuring entire systems by 2025. While AI may not yet be replacing human engineers, her comments suggest it is already transforming the planning and design phases that once consumed much of their time.
