April 28, 2026
Copyright cage match!
Who owns the code Claude Code wrote?
Leaked robot-written code, DMCA fireworks, and a comment cage match over who gets to claim it
TLDR: A high-profile leak of AI-written code sparked takedowns and a bigger question: if a robot wrote most of it, can anyone own it? Commenters split between “money wins,” “no one cares,” and “edit it and it’s yours,” while jokesters ask who owns the timber—showing why this fight matters to every coder.
A jaw-dropper: after Anthropic’s AI coding tool “Claude Code” allegedly leaked, GitHub lit up, a Python rewrite went mega-viral, and the takedowns flew. Then came the real brawl: can you even claim ownership over code that a robot wrote? Under US rules, only humans can own copyright, and “meaningful human authorship” is the magic phrase. Translation: if the machine did most of the creative work, you might not have legal protection. That’s the plot twist powering this week’s drama, and the comment section is the arena.
The hot takes are sizzling. One camp is pure cynic: “Courts side with money,” says one user, predicting big companies will keep the crown no matter who typed it. Another crowd rolls their eyes at the whole debate, basically saying nobody treats code like crown jewels anyway—devs copy, paste, and remix daily. Meanwhile, the optimists argue that if you tweak AI output enough, the new version is yours—like adding sea monsters to Jane Austen. The thread’s meme of the moment? A woodshop analogy: a machine cuts wood, but the carpenter owns the table—so who owns the code cut by Claude? International watchers ask if Europe will play by the same rules. Verdict from the crowd: bring popcorn—this legal season’s just getting started. Read the full legal breakdown by Sena Evren for the fine print.
Key Points
- •Ownership and protection of AI-assisted code depend on meaningful human authorship, employment IP assignments, and potential GPL-license contamination.
- •On March 31, 2026, Anthropic’s Claude Code source (~512k lines) was accidentally published, mirrored on GitHub, rapidly rewritten in Python, and faced DMCA takedowns.
- •The incident raised whether Anthropic can claim copyright if Claude largely wrote its own code, questioning the basis for DMCA takedowns.
- •The U.S. Copyright Office (Jan 2025) and Supreme Court (Mar 2026, Thaler appeal) confirm AI-generated works without meaningful human input aren’t copyrightable in the U.S.
- •“Meaningful human authorship” isn’t quantified; courts seek evidence of creative choices, making protection uncertain in agentic coding workflows.