The ML glossary that outgrew its only maintainer

What it does

ML Glossary is a community-built reference site—hosted on Read the Docs—that defines machine learning terms with concise explanations, citations, and (ideally) visuals, code snippets, and LaTeX equations. Think of it as a wiki that wants to be a textbook, but shorter. The project is built with Sphinx and written in reStructuredText, which the authors chose over Markdown for its richer documentation features.

The interesting bit

The most human thing in the repo is the maintainer’s apology: he’s been “heads down at Cruise” building self-driving ML infra and has left the project unattended despite 54,000 monthly active users. He’s now openly recruiting maintainers with merge privileges, and casually jokes about inventing a “cheatsheet cryptocurrency” to reward contributors before admitting he has no money for that. The ambition is clearly Distill.pub-style clarity, but the reality is a solo-maintained docs repo gasping for help.

Key highlights

• Entries require concise explanations and citations; excellent ones add diagrams, Python/numpy code, and equations.
• Explicitly prioritizes accessibility and visual/interactive diagrams over dense academic prose.
• Claims 54k monthly active users but is actively seeking new maintainers to review submissions.
• Built in RST/Sphinx rather than Markdown, targeting “large and complex documentation projects.”
• Rough drafts are accepted with the expectation that the community will enhance them over time.

Caveats

• The sole maintainer admits to prolonged non-responsiveness, so the project currently lacks active curation.
• The “cheatsheet cryptocurrency” reward system is clearly aspirational (and likely a joke), so contributor incentives are currently just goodwill.

Verdict

Worth bookmarking if you’re learning ML and need quick, visual definitions of terms like ELBO or backpropagation. Skip it if you’re looking for an interactive library or an actively curated academic journal—the content quality will depend heavily on which entries the community has polished.