TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Anthropic’s description of derived data in its Scion version control system has prompted a former engineer from Meta’s Mononoke team to point to similar design ideas in the earlier system. The engineer says Scion’s public account credited Sapling but did not mention Mononoke; Anthropic has not responded in the material provided, and the extent of any direct influence is unclear.
A former engineer who helped build Meta’s Mononoke source-control system says Anthropic’s new version control system, Scion, uses design ideas similar to Mononoke’s, and argues that Anthropic’s public description should have credited the earlier work. The comparison centers on derived data, a technique both systems use to make operations such as blame more efficient; the source material does not establish whether Mononoke directly influenced Scion.
Anthropic described Scion’s use of derived data in a post published the day before the comparison appeared, according to the author of the original report. The technique lets a version control system store or compute information derived from source history so that requests such as blame—which identify who changed particular lines—can be handled more efficiently. The source provides no measured performance figures for Scion.
The author says they described Mononoke’s use of the same concept in 2022, in materials prepared while seeking a role at Oxide. They argue the resemblance goes beyond two teams independently arriving at similar ideas. They also say Anthropic’s account credited Sapling for “hard-won lessons” but did not mention Mononoke, which they had helped build at Meta.
The author characterizes the omission as bordering on plagiarism, but also says they do not object to Anthropic adopting Mononoke’s design. They describe Mononoke as open source and say spreading useful ideas is part of the point. Those are the author’s views; the report includes no response from Anthropic and does not offer evidence establishing how Scion’s designers developed the system.
Credit for AI-Assisted Design
The dispute highlights a growing attribution problem as AI tools assist with software architecture: systems may reproduce or adapt established designs without making their origins apparent. The author argues that models trained on public code and writing can draw on earlier work, making it more important—not less—to identify relevant prior art when describing a new system.
For developers, the issue is not simply whether two systems share a technique. Derived data may be a general approach, and similarity alone does not prove copying. But clear attribution helps readers distinguish an original contribution from an established design and gives credit to teams whose work shaped later tools. The material does not establish that Claude designed Scion or that Anthropic knowingly omitted Mononoke.
As an affiliate, we earn on qualifying purchases.
Mononoke was developed by a Meta team, including the report’s author, as part of work on source control. The author places the team’s key design work in 2017–18 and says it drew on prior art, including Google’s SrcFS and database write-ahead logs. The account presents Mononoke’s open-source release as a way for others to use and build on its ideas.
The comparison is specific to the shared use of derived data to speed operations such as blame. The report itself points to a Cursor blog post by Vicent Marti as an example of a system with related insights but important design differences. That example underscores why a shared concept does not, on its own, establish that one system copied another or that the systems are otherwise alike.
As an affiliate, we earn on qualifying purchases.
What the Similarity Establishes
The available material does not show whether Scion’s developers consulted Mononoke, whether Claude generated or shaped the relevant design, or whether the overlap arose independently. It also does not include Anthropic’s explanation for omitting Mononoke from its account, or clarify the scope of its acknowledgment of Sapling.
No technical comparison of the systems’ implementations is provided. The shared derived-data concept is the focus of the author’s criticism, but the report does not establish that Scion and Mononoke use identical architectures or that any legal or licensing rules were breached. The plagiarism characterization remains the author’s allegation, not a confirmed finding.
AI-assisted software development tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Whether Anthropic Responds
The source material describes no planned technical review, formal complaint or response from Anthropic. The next development to watch is whether Anthropic addresses the comparison, explains the sources behind Scion’s design, or revises its account to acknowledge Mononoke. Until then, the relationship between the systems and the reason for the omission remain unconfirmed.
open-source version control systems
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is Scion?
Scion is Anthropic’s version control system. The report focuses on its use of derived data to make operations such as blame more efficient.
What is Mononoke?
Mononoke is an open-source source-control system built by a team at Meta. The report’s author says they helped develop it.
What is the alleged similarity?
The author says both systems use derived data to speed operations such as blame. The source does not provide a full technical comparison or establish that the implementations are identical.
Has Anthropic been shown to have copied Mononoke?
No. The report’s author argues that the similarities go beyond shared insights, but the material does not establish direct influence, deliberate copying or a licensing violation.
Why is the author raising the issue?
The author says Anthropic’s account credited Sapling but did not mention Mononoke. They say they do not object to others using open-source ideas, but argue that the earlier team should receive credit.
Source: hn
Evergreen bestsellers Picks
bestsellers
As an affiliate, we earn on qualifying purchases.
