Open-source funding and governance gaps are getting serious

A PDF arguing that governments, companies, and nonprofits should fund open-source AI drew a thread that quickly identified the core problem: goodwill and part-time contributions cannot keep pace with well-funded commercial efforts. One commenter proposed a concrete mechanism: $200K inducement prizes every 6-12 months for the first open model to hit a defined capability benchmark, modeled on Nobel laureate Michael Kremer's work on prize-based R&D incentives.

The Bluesky thread ran a parallel version of this problem. Bluesky trademarking ATProto prompted relief in the thread, but also surfaced the fact that there is no independent protocol governance organization to transfer the trademark to yet. The plan to transfer ownership 'in the future' to an 'appropriate, independent organization' that does not currently exist is exactly the governance gap that kills open standards.

These are two different domains but the same structural issue: open infrastructure requires institutional commitment that the current ecosystem is not providing at sufficient scale.