Today, Shanghai is so cold that I can only stay at home and binge-watch shows and browse Twitter.
I stumbled upon a piece of information that startled me. My first reaction was that I had read it wrong.
Google's chief engineer revealed: The project her team had been stuck on for a year was completed by her using Claude Code in just 1 hour.
Huh? You didn't read it wrong. 
A big shot from Google was secretly using a competitor's product, Claude, and it completely embarrassed her own team.
The developer community in Silicon Valley was instantly blown away.
This is so surreal that even I didn't dare to believe it at first.
01 A Year of Endless Debates vs. 1 Hour of Getting It Done
The cause of the incident is quite simple.
Jaana wanted to get a "distributed agent orchestrator" done.
We don't need to delve into the specific technical details. You just need to know that this is an extremely complex system engineering project.
Following the usual practices of major corporations like Google (and of course, including major domestic companies, you know the deal)—
- First, hold a briefing meeting;
- Then, write hundreds of pages of PPTs to align on the details;
- Team A says use this solution, Team B says that solution is better;
- An endless cycle of arguments, compromises, and process approvals...
The result was: a year passed, and the project was still "in the process of trying to build."
This "big company disease," I bet everyone reading this has deeply experienced it, right?
But this time, it seems that Jaana is a short-tempered person. She just couldn't stand it anymore.
On a whim, she threw the requirements directly to Anthropic's Claude Code.
Surprising, yet logical—
Just one hour later, Claude Code directly handed her a working prototype of the system.
Although not 100% perfect, this was something her team had fumbled around with for a whole year and couldn't produce!
This sounds almost magical.
After all, Google is not a small workshop, and those engineers are no pushovers. Can a single Claude Code really outperform an elite team?
Facing such doubts, this short-tempered小姐姐 (little sister) directly retorted:
"Stop arguing, go try it yourself. Find a field you are most skilled in, and let it write a complex thing from scratch, then you'll know what I'm talking about."
Besides the skepticism, of course, there were many people being sarcastic: Doesn't Google have a bunch of 'own sons' tools like Gemini CLI and Antigravity?
Why didn't she use them?
Facing such sarcastic remarks, her reply was very magnanimous:
You see, this is the heart of a top-tier technologist.
And just as commenter YuChen Jin said: Providing engineers with the best AI programming tools is one of the best decisions you can make.
Allowing employees to use competitor products is the magnanimity of a top-tier company.
02 Even the Elites Are Panicking: Your Worldview Might Be "Obsolete"
At this point, some might say: "Hmph, it must be because the code was too simple, or this person has no experience, right?"
Such conclusions often appear in my comment section. Whenever I post about similar programming models (or Agents), there's always a group of people jumping out to spray them as "toys" or "gimmicks."
But, Jaana Dogan is definitely not some fresh graduate intern.
She is a Google chief engineer, a legendary figure in the Go language community, and a top-tier expert ranked in the top 1% globally.
Even an expert of this level feels that the code written by current programming agents (especially Claude Code) is not just "usable," but has reached an "expert-level aesthetic" in terms of architecture and logic.
Does this mean the singularity is truly coming?
Actually, this kind of "cognitive earthquake" is sweeping through the top circles of Silicon Valley.
These past few days, not just Jaana, but even Igor Babuschkin, a core member of competitor Musk's xAI, had to admit: "Opus 4.5 is quite good."
AI circle "top god" Andrej Karpathy directly added another blow:
This is the reality.
Last night, Boris Cherny, the founder of Claude Code, casually posted a simple tutorial, and it instantly attracted over 2 million people to watch and learn.
On one side, the elites are all shouting "the world is changing," while on the other side, millions of people are cramming overnight.
Looks like the world is really about to change?
03 In Conclusion: Stop Arguing, Just Use It
After reading this incident, my biggest feeling isn't how embarrassed Google is, but rather the blunt truth said by netizens:
"Anyone who thinks AI is a bubble definitely hasn't used Claude Opus 4.5."
Now there is a large group of people online who are busy with all kinds of skepticism every day: When they see AI demonstrating writing a frontend, they argue "can it write the backend logic?"
When they see AI generating images, they spray it saying "that's just stitching together, not design."
But the problem is, after all this skepticism, they probably haven't even tried it themselves once.
This incident also reveals a more essential truth: a tool is just a tool.
Whether it's Google's or Anthropic's, whether it's a domestic champion or an overseas big model, as long as it can solve the problem and improve efficiency, it's a good thing.
It's 2026. Put aside your biases and put down your arrogance.
In this era, there are only two types of people: "doers" who use AI (even competitors) to get things done, and "commentators" who stand on the shore still arguing about whether the water is cold or not.
Which one do you want to be? Try it out, it's always better than just talking.
If you agree, give a 👍, a ♥️, and share it with your friends. This would be a great help to us.