AI NewsProductivityReported

A developer argues all AI-generated code should be open source

An essay says code written by AI should be released as open source or dedicated to the public domain. It is on the Hacker News front page with 783 points.

AI News

Editorial2 min read

LinkedInX

Why it mattersThe essay changes the question a team asks before using code from an assistant. The question is no longer only quality. It is also who can own it.

A founder closing a Series A fills in a due-diligence sheet. One box asks which parts of the codebase were written by an AI assistant. The next asks who owns those parts. Most teams skip the second question. An essay on Hacker News is making the case that the human does not. The piece, "All AI-generated code should be Open Source," is by the developer Jakub T. Jankiewicz on his own blog. It carries 783 points on Hacker News and sits on the front page.

The argument, in two parts

Jankiewicz makes two points. The first is about where the models came from. The code written by the assistants was learned from open-source projects. It was also learned from answers on Stack Overflow. "That code was simply scraped from the internet and used without permission or proper attribution," he writes. Stack Overflow's licence requires attribution. He argues the licence is broken when the training data is sold back as a proprietary feature.

The second point is about who can own the output. He cites the Monkey Selfie case. A monkey cannot hold copyright on a photo it took. The US Copyright Office has applied the same rule to work made by an AI alone. He argues that code produced from a one-line prompt sits on the same side of that line.

What other bodies already do

Wikimedia Commons tags every image that is only AI as public domain, he notes. UK law grants a limited copyright to computer-generated work. That copyright does not exist in the US. The author's proposal is for developers and companies to release assistant-written code as open source. Or to dedicate it to the public domain under the Creative Commons Zero licence. He aims the proposal at cases where the human adds little beyond the prompt.

Why this is on the front page

The post is an opinion piece. It cites no new ruling. It is on the Hacker News front page because the question matters to any team selling software. Who owns code written by a model is now unavoidable. The answer most legal teams give today is that nothing is settled. The matter will be decided case by case. The essay's claim is that the honest reading of current law and policy is already leaning one way. A team that treats assistant output the same as hand-written code is making a bet, not a judgement.

Teams that want a clean answer have three options today. Mark assistant-written files as public domain. Have a human rewrite the parts the company wants to own. Or keep records detailed enough that a buyer can tell which lines are which. The essay's contribution is not a legal answer. It is making the question harder to ignore.

Source

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.

Share
LinkedInX
Start a project