Static

More on Anthropic's Scion ripping off Mononoke

First reported by Sunshowers ·

The signal ●○○○ Compiled by AI from Sunshowers and Hacker News
Why you might care

Anthropic's Scion is now known to have intentionally copied core design elements from Mononoke, making its claimed achievements less impressive and its attribution dishonest.

What happened

A follow-up post clarifies that Anthropic's Scion source control system intentionally copied core architectural elements from Mononoke, a system developed by Meta. The author, previously involved with Mononoke, states that the resemblance goes beyond shared concepts like "derived data" and extends to the entire data model, including "Bonsai changesets," which he argues is directly mirrored in Scion's "changesets." Notably, the Scion team allegedly includes former Meta personnel who worked on Mononoke, making intentional appropriation highly probable. The author emphasizes that while copying good ideas is acceptable, proper attribution is crucial. He criticizes Anthropic for not crediting Mononoke, particularly its unique "Bonsai" data model, and instead offering vague acknowledgments like "hard-won lessons" from Sapling, which he deems insufficient and dishonest.

What it means

The article highlights a pattern where large AI companies appropriate foundational work from the open-source community without adequate credit, diminishing the perceived innovation of their own efforts. This intentional omission of attribution, particularly concerning a novel and complex data model like Mononoke's "Bonsai," suggests a broader industry trend of rebranding existing, sophisticated engineering as solely the product of LLM capabilities or internal development.

This situation challenges the narrative of AI-driven innovation by revealing the reliance on and appropriation of human-engineered systems, potentially impacting how credit is assigned and how intellectual property is managed in the AI development landscape. It also raises questions about the ethical responsibility of companies to acknowledge their sources, especially when reverse-engineering complex systems, and whether this oversight will lead to greater scrutiny of public engineering claims.

AI-written summary. May contain errors.