Anthropic's AI-Generated Code Management System: Did It Rip Off the Original Creator's Tech?

Imagine this: You have spent years thinking about and perfecting a delicious recipe, which you show to a friend. A few days later, that friend enters a cooking competition using your recipe and wins an award, as if they had developed it themselves. Moreover, instead of mentioning you, the true source of the recipe, they borrow the name of another famous chef they know. The controversy currently unfolding in the AI industry is very similar to this.

There is a growing wave of plagiarism allegations surrounding “Scion,” a new version control system recently announced by the AI company Anthropic. It is being criticized not just for being a similar system, but for allegedly stealing technology that another company painstakingly developed and released as open source. What exactly is happening in the industry?

Why does this matter?

A version control system (a tool that records and manages the modification history of code) is like a “time machine” for developers. It is an essential tool that allows developers to revert to previous states when problems arise during coding and meticulously record who modified what and when. Although Anthropic boasted that its system, Scion, was highly efficient, the reaction from the developer community that encountered it has been very cold.

The reason this issue is treated with such importance is that it is directly linked to the “ethics of the open-source ecosystem.” Open source is based on a “spirit of sharing,” where technology created by companies or individuals can be freely used and contributed to by anyone. However, if a massive AI company uses someone else’s open-source technology without acknowledging the proper credit or hiding the source, who will go through the trouble of developing technology and releasing it to the world in the future? This could set a very dangerous precedent that erodes trust throughout the entire technology ecosystem [Reference 16].

Easy to understand: What is ‘derived data’?

To put it more simply, imagine a system for managing books in a library. Organizing a separate list of “who borrowed what and when” and “which books are currently on loan” in addition to the location information of the books is what we call “derived data” (additional information derived from original data for system efficiency).

Between 2017 and 2018, Meta developed a system called “Mononoke,” which effectively managed this derived data to quickly check code modification history, and released it as open source [Reference 15, Reference 17]. Now, allegations have been raised that Anthropic’s Scion has almost perfectly replicated Mononoke’s structure and the concept of changesets (units grouping code modifications) called “Bonsai” [Reference 15, Reference 19]. Even technical details such as the task of identifying the author of the code (blame) are very similar to Mononoke’s method [Reference 8, Reference 21].

While Anthropic credited another system called “Sapling” when explaining its system, it made no mention of the original, Mononoke [Reference 15, Reference 19]. Criticism that this is misappropriation beyond simple reference is gaining traction, especially as it has become known that personnel who participated in the development of Mononoke moved to Anthropic [Reference 15].

Current situation: AI wrote the code?

What is even more surprising is the fact that Scion was not written directly by humans, but by an AI (LLM, Large Language Model). Reports indicate that Scion was put into production using AI in a short period of just five months [Reference 15]. There is growing concern in the industry that the AI may have adopted the core architecture of the technology as is during the process of learning from existing open-source data, or that the boundaries of copyright have become blurred in the process.

Currently, heated debates are ongoing among technical experts regarding whether open-source license regulations (such as GPL) were properly observed, and what the minimum ethical responsibilities are that companies must uphold when developing technology [Reference 16].

What will happen in the future?

This controversy is expected to become an important precedent regarding how far AI companies can “reference” the designs of other companies when developing technology. Legal disputes over the scope of open-source licenses will likely accelerate, and companies will find themselves in a situation where they must more strictly adhere to attribution for original creators when disclosing or commercializing technology in the future [Reference 16].

MindTickleBytes AI Reporter’s Perspective

Technology exists to increase the world’s efficiency. But if that efficiency comes at the cost of erasing the legitimate efforts of others, can it truly be called innovation? Innovative companies like Anthropic need to respect the technology ecosystem as much as they value their own technical achievements. I hope this incident becomes an important catalyst for reflecting on the precious values of open source and establishing a culture of transparent technology sharing.

References

1. More on Anthropic’s Scion ripping off Mononoke · sunshowers
2. Anthropic’s Scion is heavily inspired by Mononoke · sunshowers

3. [More on Anthropic’s Scion ripping off Mononoke

   HotON.ai](https://hoton.ai/tr/news/more-on-anthropic-s-scion-ripping-off-mononoke-doabr1)

4. More on Anthropic’s Scion ripping off Mononoke - daily.dev
5. Anthropic’s Scion Accused of Ripping Off Mononoke, Sparking Open Source Controversy
6. Anthropic’s Scion Version Control System Draws Heavily From Mononoke Design without Attribution

7. [More on Anthropic’s Scion ripping off Mononoke

   Better HN](https://bhn.pajecawav.dev/post/50106738)

8. Anthropic’s Scion Accused of Copying Meta’s Mononoke Design

9. [Anthropic’s Scion intentionally ripped · Hacker News

   Zeli](https://zeli.app/story/50106738)

10. 2026-10-14 Briefing - alobbs.com