deep-funding-primary-sources-2026-05-11

Deep Funding primary sources

Reviewed: 2026-05-11 UTC

Core framing

• https://www.deepfunding.org/
  • Official project site.
  • Used to confirm the project’s self-description as a market of AIs as the engine and humans as the steering wheel, the three-part mechanism split (dependency graph / model submissions / spot checkers), the three challenge levels, the collaborator set, and the current public challenge framing around ~15,000 weights, 34 seed nodes, and >5,000 child nodes.
• https://hackmd.io/f3Cz0_RhS5-6pXpYHn2iKg
  • Public mechanism draft titled Deep Funding: A Prediction Market For Open Source Dependencies.
  • Used to confirm the stronger conceptual framing around bots / models / traders, jury override or precedence logic, and the goal of allocating ecosystem funding through market-weighted predictions of human judgment.

Graph and scoring implementation sources

• https://raw.githubusercontent.com/deepfunding/dependency-graph/main/README.md
  • Official repository README.
  • Used to confirm the edge-weight semantics, the current depth-2 graph framing, the updated graph counts, the seed-repo set, and the fact that model builders are really assigning credit flow from dependent repos to dependencies.
• https://raw.githubusercontent.com/deepfunding/scoring/main/README.md
  • Official scoring repository README.
  • Used to confirm that Deep Funding exposes at least two explicit scoring geometries: single-layer and two-layer / three-layer approaches.

Public explainers and integration context

• https://gitcoin.co/apps/deepfunding
  • Public app page.
  • Used to confirm the high-level Ethereum dependency-funding framing, the initial $250,000 sponsorship breakdown, and the way the project is described to public-goods funders and participants.
• https://gitcoin.co/mechanisms/deep-funding
  • Public mechanism page.
  • Used to confirm how Deep Funding is narrated in the broader AI-PGF landscape and how its human-jury-plus-model-competition structure is presented to the public.
• https://paragraph.com/@pairwise/deep-funding
  • Official Pairwise post about Deep Funding.
  • Used to confirm how juror spot checks were operationalized through pairwise repository comparisons and to situate Pairwise as the human-comparison layer inside the broader mechanism.

Main takeaways from this pass

• Deep Funding cleared the bar because it is not just another grants app or AI-evaluator pitch; it separates graph construction, edge-weight submission, jury spot checks, and final allocation into distinct layers.
• The most useful reusable primitive is not AI grading but dependency-credit routing. A submission says how much credit a source repo passes to each dependency, with the residual remaining at the source.
• The project sits at a useful junction of several existing corpus branches: Open Source Observer-like data infrastructure, Pairwise-like human comparison UX, Drips-like upstream funding logic, and grants-mechanism capital allocation.
• Public materials disagree on exact graph and reward figures. That is not just noise: it shows that graph curation, seed selection, challenge scope, and payout structure are themselves governance surfaces.
• This entry belongs in the active corpus because it gives the public-goods funding cluster a better comparison point for dependency-directed and jury-audited capital allocation, especially where ecosystem-scale judgment must be compressed into a tractable interface.