As Silicon Valley is struggling in a future, where AI agents do most of the software programming, a new problem is being created: Finding the Bugs produced by AI before production. Even Openai deals with such issues, a former employee described.
Newly formed start Player Create a solution: Use the AI agents who are trained to find and correct problems before the code production, says the start -up CEO and the only founder of Animeesh Koratana.
Koratana created the Playerzero while at the Stanford Dawn Laboratory for mechanical learning under his adviser and laboratory founder, Matei Zaharia. Zaharia is, of course, a renowned Databricks programmer and co -founder. He created his fundamental technology while working for his own doctorate.
Playerzero announced on Wednesday that it has increased a $ 15 million series, led by Ashu Garg of Foundation Capital, an early Databricks supporter. This is followed by a $ 5 million seed led by Green Bay Ventures and several notable angels, including Zaharia, CEO Dropbox Drew Houston, CEO Figma Dylan Field and CEO by Vercel Guillermo Rauch.
During her time at Stanford Dawn, Koratana, now 26, worked on AI modeling technology and “exposed to language models too early,” he says. He met with developers who created some of the first AI encoding tools.
He then hit him that “there is this world in which computers are going to write the code. They are not going to be people anymore,” says Koratana Techcrunch. “How will the world look like at this point?”
He knew before the term “ai slop”, even popular that these agents would produce code that broke things like their human supervisors do.
TechCrunch event
Francisco
|
27-29 October 2025
This problem will also deteriorate by so many agents who have sprung much more code than ever written before. It will not always be practical for people to control all AI written code for errors or illusions. And the issue becomes even more intense for the large, complex bases of business -based code.
Playerzero trains models “that really understand the code bases and we understand how they are manufactured, the way they are architecturally,” says Koratana.
His technology studies the story of the errors, issues and solutions of a business. When something breaks, his product can then “understand why he can correct it and then learn from these mistakes to prevent them from happening again,” says Koratana. It drives its product to an immune system for large code bases.
The landing of Zaharia, his adviser, as an angel, was a first step in raising capital, but the moment he really validated his idea was when he showed a demo to another famous programmer: Rauch. Rauch is the founder of Triple Unit tooling Company Vercel and creator of the popular javascript open source frame Next.js.
Rauch watched Koratana’s demo with interesting but skepticism, asking how “real”. Koratana replied that this was the code “running in production. Then its investor who will soon be the angel investor replied:” If you can really solve this way you can imagine, it’s a great deal. “
Of course, Playerzero is not alone in trying to solve the error problem created by AI. Just last week, Anysphere’s runner Started Bugbot To detect coding errors, as just one example.
Still, Playerzero is already gaining attraction for its emphasis on large codes. While designed for a world where agents are the encoders, it is currently used by several large companies using co-pilots. For example, Zuora subscription billing company is one of Marquee’s start -up customers. Zuora uses technology in all its mechanical groups, including watchdog of the most valuable code, its billing systems, he said.
