Google is reportedly engaged in a new mysterious project: training artificial intelligence (AI) to write code, as well as subsequent bug fixes and code updates, people familiar with the matter said today. Analysts say the project could reduce the need for human software engineers in the future.
Sources close to the project said the project, which is being developed under the code name “Pitchfork” within Google, was originally set up in the “X” research division of Google’s parent company Alphabet and has now made its way to Google Labs, meaning the project Pitchfork is now in Google Labs, which means the project is being taken seriously by senior management.
Pitchfork is now part of a new division of Google Labs called “AI Developer Assistance,” led by Olivia Hatalsky, who has worked on the project since its inception. Hatalsky, who was responsible for Google Glass and other “moonshot projects.
According to internal documents, the purpose of the Pitchfork project is “to train code to write code, and to repair and update itself, using machine learning (ML)”. Pitchfork will first learn programming styles and then write new code based on what it has learned.
A Google spokesperson said in response, “The Pitchfork team is working closely with the company’s research team to explore different use cases to help developers.”
The Google spokesperson also said that the initial goal of Pitchfork was to build a tool that could update Google’s Python programming language codebase to newer versions. But over time, the goal of the project shifted to a general-purpose system: one that could reduce the need for humans to write and update code, while maintaining code quality.