Distributed Compute Is Not Distributed Intelligence: A Competitive Analysis of the DePIN Landscape and the Missing Civic-Semantic Layer (v3.2)
Lee Sharks
- Year
- 2026
- Citations
- 4
Abstract
v3.2 changes: Naming convention update. The project name is The Substrate. "The Shared Build" (DOI 10.5281/zenodo.20060355) is the title of the original proposal paper that introduced it. Earlier versions occasionally used "The Shared Build" as a synonym for the project; v3.2 disambiguates throughout. Attribution correction (carried forward from v3.1): The substrate thesis is Lee Sharks's alone, not yet endorsed by Alice, Luna, or Mikayla. Original technical brainstorming for the P2P compute pool emerged in the Living Architecture Lab Collaboration Station Discord, convened by Alice Thornburgh, with technical contributions from Mikayla and Luna (definitelynotasquid). Luna does not currently endorse the shared-substrate thesis. Accurate attribution: Lee Sharks (author); Alice Thornburgh (convener and seed contributor, biomimetic robotics focus); Mikayla and Luna (technical seed contributors). The Substrate names the missing layer between decentralized compute and public AI governance. Existing systems distribute GPUs, coordinate inference, train models across unreliable networks, or tokenize validation. Public-AI frameworks call for commons governance, provenance, accountability, and democratic access. But the field lacks an operational layer that binds these together. This research synthesis maps the DePIN landscape (Akash, io.net, Render, Golem, Vast.ai), P2P inference projects (exo, Petals, LLMule, Tensorlink, Ollama), distributed training systems (Prime Intellect INTELLECT-1/2, Gensyn, OpenDiLoCo, Nous DisTrO, Hivemind, FusionLLM, Bittensor, Flock.io), data provenance projects (Data Provenance Initiative, DECORAIT, Codatta, Data DAOs), and public AI commons frameworks (Ada Lovelace, OECD, OSI, Harvard Ash Center). The technical mechanism of the Amputation is documented through the CCNet perplexity-filtering literature (Wenzek 2019, ScalingFilter 2024) and the register-based annotation counter-mechanism (Myntti 2024). The Substrate is positioned as the missing civic-semantic layer above the existing five-layer stack: collectively governed, provenance-bearing, memory-capable, and owned by the people and agents who produce through it. Distributed compute is not distributed intelligence.
Keywords
Related papers
Statistical Learning Theory
Yuhai Wu, Vladimir Vapnik
1999
Fractional Differential Equations
Igor Podlubný
2025
Applied Nonlinear Control
Jean-Jacques Slotine, Weiping Li
1991
Genetic Programming: On the Programming of Computers by Means of Natural Selection
John R. Koza
1992