A senior Google engineer says that Anthropic's Claude Code built in one hour a system that took Google one year to make. Jaana Dogan, Principal Engineer at Google, said that the agentic AI coding tool was tasked with generating a distributed agent orchestration system — which it managed to do in just one hour. Dogan said her own team at Google had spent one year building a similar system.
Humans vs AI
In a post shared on the social media platform X, Dogan wrote: “I'm not joking and this isn't funny. We have been trying to build distributed agent orchestrators at Google since last year. There are various options, not everyone is aligned…”
“I gave Claude Code a description of the problem, it generated what we built last year in an hour,” said the San Francisco-based techie.
Dogan said that Claude Code’s result was not perfect, but comparable to what the team at Google built over the course of one year.
“If you are skeptical of coding agents, try it on a domain you are already an expert of. Build something complex from scratch where you can be the judge of the artifacts,” the Google engineer advised.
Google techie provides more context
{{/usCountry}}“If you are skeptical of coding agents, try it on a domain you are already an expert of. Build something complex from scratch where you can be the judge of the artifacts,” the Google engineer advised.
Google techie provides more context
{{/usCountry}}Dogan later clarified that the underlying ideas behind such systems are not new and that her teams had already built multiple versions last year, each with trade-offs and no clear winner.
What has changed, she said, is how quickly coding agents like Claude Code can now turn well-understood ideas into working prototypes.
She stressed that the real difficulty lies in spending years learning, testing, and grounding ideas in real products, not in writing the code itself. Large organisations also face inertia because their infrastructure must work across many use cases, accumulating complexity over time.
Tools like Claude Code, Dogan said, make it easier to start from scratch without legacy baggage — and even without detailed prompts, she was surprised by the quality of the design recommendations it produced.
“What I built this weekend isn't production grade and is a toy version, but a useful starting point. I am surprised with the quality of what's generated in the end because I didn't prompt in depth about design choices yet CC was able to give me some good recommendations,” she concluded.