Submit
Three kinds of contribution. All of them public.The index is seeded from awesome-fly and re-synced from it, so a new project belongs there first — it will arrive here on the next sync. Ratings, evidence and disputes live on this side.
Not in the 141? Add it to awesome-fly — a pull request against the list, or the suggestion form if you would rather not edit markdown. It lands here on the next sync, unrated, with its description carried over.
Most of the index is unrated. Pick a project, read its source, and file the level you think it earns with a link behind every claim. Bring the negative result too if the project reports one — those are the useful ones.
Think a level is too generous, or too mean? Say so, with the same evidence standard. Disputed ratings stay visible while they are argued — nothing gets quietly edited away.
Wrong author, dead link, description that misreads the project, a dataset we mislabelled. Small corrections are as welcome as ratings — the index is only as good as its worst entry.
What makes a rating land
The full rules →- A link per claim. “It has plasticity” is not evidence. A file, a line, a README section or a paper is.
- Name what changes. Which synapses, driven by what signal, on what timescale. This is the whole L1/L2 boundary.
- Say why not the level above. The most useful sentence in a rating is the one that rules out the next level up.
- Quote the project's own limits. Many of these projects are admirably honest about what they did not achieve. Carry that over.
thelearningfly.com rates projects, not people. Keep disputes about the evidence.