Databricks engineer questions Anthropic's attribution for Scion version control system
A detailed technical critique alleges Anthropic's new source control tool borrows heavily from Meta's Mononoke without sufficient credit.
What to know
- Anthropic's Scion version control system draws heavily on Meta's Mononoke without adequate attribution, according to technical analysis by a Databricks engineer.
- The controversy reflects broader concerns about AI companies and developers using AI tools to build systems inspired by prior work, then advertising them without clear credit.
- Technical community discussion highlights the absence of established norms for attribution when systems are developed with AI assistance or inspired by academic/industry work.
- Core design patterns in question trace back to Durham Goode's work at Facebook on Mercurial scaling and user experience, which influenced both Mononoke and now Scion.
The dispute Whether Anthropic deliberately copied Mononoke without attribution or independently arrived at similar solutions through legitimate technical reasoning, and where to draw the line between acceptable reinvention and unethical credit-taking. · positions read across 20 posts and comments
Anthropic should have credited Mononoke and prior work more clearly when publishing marketing materials about Scion.
-
“There's an obvious incentive for AI companies to credit their own model, over the people whose work the model digested.”
andyc · Lobsters ↗
The analysis is well-reasoned and clarifies important distinctions between Scion and competing approaches that press coverage missed.
-
“I feel like I learned a lot about both Scion and the rest of the source control space from this blog post, especially in comparison to what I learned from the Anthropic one.”
borntyping · Lobsters ↗
There are no clear norms yet for attribution when using AI or when systems are developed independently but inspired by prior work; establishing standards is the real issue.
-
“I've been using AI to help with a lot of system designs work at $WORK lately…am I ripping someone else off without credit…I think the system I'm building on is different enough…but...”
quasi_qua_quasi · Lobsters ↗
typesanitizer Databricks Source Control engineerAnthropic AI companyRain Paharia Technical writer/analystDurham Goode Former Facebook Source Control team lead
How it unfolded 5 developments, newest first · click a bar or a number to jump postscomments
-
5
Commenters distinguish between independent reinvention and deliberate copying
Discussion clarified that the concern centers on Anthropic publishing a blog post advertising Scion while allegedly copying existing work, with one commenter noting suggestions that this "wasn't a case of Anthropic's model reinventing stuff without attribution so much as them deliberately copying existing work without attribution."
“But a clear line is when you start writing a blog post advertising your work, like Anthropic did…”
— andyc -
There are some other negative incentives with a monorepo, which show up whenever Google people touch a project with downstream dependencies. In a monorepo, API providers and API consumers are in the same place. If you want to refactor a component then you can, and then CI will catch any places you failed to fix consumers and you can fix them…
2 more of the top 3 · 15 posts in this stretch
-
> I recall being frustrated with that, and half-jokingly proposing that it would be good to move to a more custom storage solution or have custom indexes next to Git’s storage itself. There's surprisingly very few options for databases with good structural sharing. I think Pijul's…
-
> A part of me wonders if ex-googlers have just kinda poisoned the ranks of staff engineers all over the world and gotten everyone to copy the monorepo + bazel over and over again. I think it's more that if you do multiple repositories you have to figure out policy and tooling for how they're all going to be coordinated. A monorepo means you can…
-
-
4
Technical community reflects on attribution standards when using AI
Discussion expanded to broader questions about attribution when developers use AI tools for system design, with commenters noting the absence of established norms for crediting prior work discovered through AI.
“I've been using AI to help with a lot of system designs work at $WORK lately… am I ripping someone else off without credit.”
— quasi_qua_quasi -
The matter aside, this post is really really well written.
2 more of the top 3 · 4 posts in this stretch
-
I feel like I learned a lot about both Scion and the rest of the source control space from this blog post, especially in comparison to what I learned from the Anthropic one.
-
Certainly biased here, but seconded. Thanks so much for writing this, ~typesanitizer.
-
-
3
Commenter raises broader pattern of AI company attribution practices
Lobsters user andyc noted an "obvious incentive for AI companies to credit their own model, over the people whose work the model digested" and referenced the Navier-Stokes proof controversy involving OpenAI and Anthropic.
“There's an obvious incentive for AI companies to credit their own model, over the people whose work the model digested.”
— andyc -
Thanks for writing this up. Your background and experience convinced me :-) There's an obvious incentive for AI companies to credit their own model, over the people whose work the model digested. If there isn't already, there should be a list of these incidents / marketing programs: - The Navier-Stokes proof and unethical behavior by OpenAI - IIRC…
-
-
2
Post gains traction on Hacker News and Lobsters
The analysis reached audience on multiple platforms, with the Lobsters post scoring 82 points and 11 comments, and concurrent discussion appearing on Hacker News.
-
1
Databricks engineer publishes comparison of Scion and Mononoke
Typesanitizer from Databricks's Source Control team published a detailed technical analysis comparing Anthropic's recently announced Scion centralized version control system to Meta's Mononoke, recommending readers also review Rain Paharia's posts titled "Anthropic's Scion is heavily inspired by Mononoke" and "More on Anthropic's Scion ripping off Mononoke."
“hg smartlog gives you a focused view of your branches and how they relate to the development trunk…the critical thing here is not that it can be done; it's that it becomes the default view of their source-control environment.”
— Durham Goode, Facebook source control team lead (quoted in typesanitizer's analysis) · source
What people are saying 13 voices from 1 site · best of 20 · verbatim
- Nov 2
-
From the depth on the post: > This makes Anthropic’s achievement significantly less impressive Didn't we all know that the whole thing was anyways generated? Claude is a thinking database, for a given set of parameters it found the perfect match and then unzipped it into a new project. It had to know the source though, thus the attribution was due…
-
I keep thinking "oh, it would be nice to have something like Manifold" (which, for my purposes, I understand to provide lower-latency-than-S3 access to immutable, content-addressed objects). Then I realize that, as a member of the source control team, it's probably _my_ job to build it 😅
-
> In this case, you could say that Mononoke depends on magic, infinitely-scalable object storage, but this is no longer remarkable because everyone has access to it these days! Fun fact, Mononoke was one of the first users of Meta's S3 equivalent, Manifold. Though we multiplexed object writes across Manifold and MySQL (the latter used as a…
-
Is that something that we can blame on monorepos, though, or is it Google's engineering culture and its incentives, and they took advantage of a monorepo to make it possible? Facebook had a few repos and steadily switched to a monorepo following the same internal incentives, and I don't think it's the wrong choice for either of them. If well…
-
This post quoting Anthropic's post: > where tools built for human scale start to strain. I can see two kinds of strain here. Human style use of tools on a magnificent scale, or some new, AI-style use of tools that requires previously-unattainable levels of resources. Too bad this isn't clarified. I bet it's human-style-use on a large scale. "AI"…
-
> To see the impact of this in practice, compare the effort involved in building a ten-year-old Android and iOS app for the latest OS release. I thought that Android specifically is managed quite differently at Google? Is it still part of the monorepo, just synchronized with the separate public git repos?
-
You'd need to make sure that you have sign-off gates for anything that changes things outside of a specific team's responsibility, but ideally you'd also want CI that cloned two different versions of the monorepo and did testing with different subsets against others. But it's not really about what is possible, it's about what is *easy*. Things…
-
> you have to do the design up front because the cost of changing is high. Interesting point. If that’s the only/a major downside, could you enforce this somehow in a monorepo? Probably not, right?
-
I remember seeing a blogpost that made the claim "if you don't have a monorepo, you don't have 'continuous integration' - you have, at best, 'very frequent integration'"
-
Oh yeah, if we ever decided to write a blog post or a journal article (this is an area where we could at least plausibly publish a paper) you bet I would do a thorough literature search. Though IIRC there was some talk I saw implying that this wasn't a case of Anthropic's model reinventing stuff without attribution so much as them *deliberately*…
-
I think this is bound to happen, and there are no norms right now. But a clear line is when you start writing a blog post advertising your work, like Anthropic did ...
- Nov 1
-
Thank you for this write up! I love your writing style and how in depth it is! I do think the "7 people full time on a major migration (with tokens)" is an interesting strategy. That's probably enough people to really have a good idea of all the parts, that can tackle a _lot_ of different tiny issues, but (if they have the right autonomy) they can…
-
I've been using AI to help with a lot of system designs work at $WORK lately (not directly, just "give me ideas for X, pick one, build it out, evaluate whether it works, etc etc etc") and this kind of thing has been on my mind a lot lately: am I ripping someone else off without credit. Like, I think the system I'm building on is different enough…