Anthropic’s Scion is closely impressed by Mononoke · sunshowers


Discuss on Hacker News or Lobsters.

Here is Anthropic describing (archive link) yesterday how their new model management system, Scion, makes use of an idea referred to as “derived knowledge” to make operations like blames extra environment friendly:

How Anthropic describes Scion’s derived knowledge.

Here is me, again in 2022, describing in my Oxide candidate supplies how the Mononoke supply management server I helped construct makes use of an idea referred to as “derived knowledge” to make operations like blames extra environment friendly:

Derived data. While Bonsai changesets work great for pushes, they work less well for things like diffs. Git diff commit1..commit2 is an O(n log n) operation where n is the number of changed files – with Bonsai changesets, you’d have to look at every single commit between commit1 and commit2 to compute the list of files changed. For a large repository this can be orders of magnitude slower (and compounds the pointer-chasing problem). To solve this, we moved the diff-friendly representation to being derived data (i.e. an index). The server creates a separate Merkle DAG with a Git-like tree/blob model, with pointers back to Bonsai changesets as required. Note that: Derived data can be computed and stored asynchronously, outside of the main critical section. The pointers between commits and trees are flipped around. With the Git data model, the text of a commit contains the hash of the root tree. With the Bonsai model, the text of the root tree contains the hash of the commit that introduced it. This gave us the ability to build out other derived indexes for whatever other operations we wanted. For example, git blame is quite slow on large git repositories, requiring a traversal of the entire repository (not just commits that touch the file). With Mononoke, we built out a blame-friendly index where each version of each file was a separate node in the Merkle DAG, with its parent set to the previous version of the file, and with pointers back to Bonsai changesets. This meant that blames became O(number of commits that touch the file) rather than O(number of commits in the repository). Over time, the team has added more indexes to handle specific use cases, and Mononoke now has 8 separate indexes. Adding a new index is a relatively cheap operation that doesn’t involve any fundamental changes to data structures.
How I described Mononoke’s derived knowledge.

I believe the extent of imitation goes properly past having related insights. The proven fact that Sapling was solely credited for “hard-won classes”, and that Mononoke was not talked about in any respect, borders on plagiarism. The submit comes off as Anthropic claiming that Claude designed all this, whereas the truth is that our (extraordinary) crew at Meta got here up with most of the insights that Claude dutifully utilized nearly a decade later. (We checked out loads of prior artwork for Mononoke and Sapling too, from Google’s SrcFS to write-ahead logs in databases.)

I’m very glad that the imaginative and prescient our crew had again in 2017-18 was so highly effective and enduring — whenever you go in opposition to the grain, you’re all the time a bit uncertain if what you’re doing is genius or insanity, and I believe Mononoke’s design has stood the take a look at of time. And, to be clear, I should not have an issue with Anthropic copying Mononoke’s design. (This is a part of why we put out open supply — to profit humanity by means of the unfold of fine concepts.)

But as LLMs proceed to help in designing new programs, they’re prone to index closely on present designs. Mononoke is open source and in each LLM’s coaching set. If you ask Opus 5.5 or GPT-6 Astra (thanks Matthew Fowle) about prior artwork, they each instantly level out the similarities to Mononoke.

Giving credit score to people who got here earlier than us is now simpler — and extra vital — than ever. There is not an excuse.



Source link