Anthropic's Scion is heavily inspired by Mononoke
Points and comments are a snapshot, not live.
An Anthropic engineer claims Scion's derived data concept closely mirrors Mononoke's design without proper credit.
The author, a former Meta engineer who helped build Mononoke (the server-side component of Sapling), details how Anthropic's new VCS Scion uses a "derived data" mechanism nearly identical to Mononoke's. Both systems asynchronously compute indexes (e.g., blame, file history) from changesets to optimize operations. The author notes Mononoke is open source and in LLM training sets; when asked, Claude Opus 5.5 and GPT-6 Astra immediately point out similarities. Anthropic's blog post credited Sapling for "hard-won lessons" but did not mention Mononoke.
What commenters are saying
The thread largely sides with the author, seeing the omission as a lack of proper attribution. Several commenters argue that while Sapling is cited, Mononoke-as Sapling's core server-deserves explicit mention given the direct copying of data model and terminology. Some defend Anthropic, noting they did credit Sapling; others counter that credit should go to the specific component. A subthread debates whether this matters for an IPO, with one commenter calling it marketing spin. A few note the irony if Claude wrote the post.