
Oracle has issued a new rule prohibiting the submission of AI-generated code or text to the OpenJDK project. This story is currently going viral on Hacker News (401 points).
Key Points:
• Code, text, and images in Git, GitHub pull requests, emails, and the Java Bug System are all banned.
• AI can still be used privately to read and debug code — only submitting the generated output is forbidden.
• Reasons: Code reviewer burden, security, and intellectual property issues. Oracle stated that code that “looks correct but is actually wrong” threatens critical systems.
• This contrasts with Larry Ellison, who previously claimed that Oracle’s code today is written by AI models.
[Insight: Even the company pushing AI the hardest doesn’t dare let AI touch the code that the world relies on. A lesson for Cambodian developers: You can use AI as an assistant, but you must fully understand the code you put your name on.]
Kun Long
Author at The Sharing KH