PlayerZero raises $15M to prevent AI agents from shipping buggy code 

While Silicon Valley racets to a future where AI agents perform most of the software programming, a new problem is created: finding the bugs generated by AI before they are put into production. Even OpenAi has to do with such problems, a former employee has described.

New financed startup Player has made a solution: use AI agents who are trained to find and solve problems before the code is put into production, the founder of the startup, Animesh Koratana, tells WAN.

Koratana created Playerzero while he was in the Stanford Dawn Lab for Machine Learning among his adviser and Lab -founder, Matei Zaharia. Zaharia is of course a famous developer and the co-founder of Databricks; He created his fundamental technology while working on his own doctorate.

Playerzero announced on Wednesday that it collected a Serie A of $ 15 million under the leadership of Ashu Garg of Foundation Capital, an early data tabricksbacker. This follows a seed of $ 5 million led by Green Bay Ventures and various remarkable angels, including Zaharia, Dropbox CEO Drew Houston, Figma CEO Dylan Field and Vercel CEO Guillermo Rauch.

During his time at Stanford Dawn, Koratana, now 26, worked on AI model compression technology and “was exposed to language models very early,” he says. He met the developers who made some of the first AI coding tools.

It then touched him: “There is this world in which computers start writing the code. It will no longer be people,” Koratana told WAN. “What will the world look like at that point?”

He knew that before the term “ai slop” it was even thought that these agents would produce code that broke things, just like their human overseers.

That problem would also be exacerbated by so many agents who push so much more code than ever written before. It will not always be practical for people to check all code written by AI for bugs or hallucinations. And the problem becomes even more intense for the large, complex code bases on which companies trust.

Projectzero trains models “that understand code bases very deeply, and we understand how they are built as they are architated,” says Koratana.

His technology studies the history of the bugs, problems and solutions of a company. When something breaks, his product can “find out why and repair, and then learn from those mistakes to prevent them from ever happening again,” says Koratana. He compares his product with an immune system for large code bases.

Landing Zaharia, his adviser, as an angel was a first step in fundraising, but the moment that his idea was really validated when he showed a demo to another famous developer: Rauch. Rauch is the founder of Triple Unicorn Developer Tool Company Vercel and maker of the popular open-source JavaScript framework Next.js.

Rauch watched the demo of Koratana with interest but skepticism and asked how much of it was ‘real’. Koratana replied that this code was “working in production. Like, this is a real example. And he was quiet,” says Koratana. Then his upcoming investor of the Angels replied: “If you can actually solve this as you imagine, it is a very big problem.”

Playerzero is of course not only to solve the bug problem generated by AI. Last week, the cursor of Anysphere Launched Bugbot To detect coding errors, as just one example.

Nevertheless, Playerzero already gets a grip on his emphasis on large code bases. Although it was conceived for a world where agents are the coders, it is currently used by various large companies that use coding co-pilots. For example, the Zuora billing company is one of the Tentup customers of the startup. Zuora uses the technology in its technical teams, including its most valuable code, to watch his billing systems, said it.