Decentralized Science Platform (LM-Dawn class)

Class card for the LM-Dawn cluster of platforms that bypass MM/DM institutional gatekeeping in scientific research — journals, peer-review monopolies, university affiliation requirements, and funding-agency concentration — via open-protocol publication, cooperative IP frameworks, distributed compute infrastructures, and mass-collaboration mathematics. DESCI signal versus grift distinction [CANON-frame; EXTRAP-classification]: The 2021–2023 token-bubble produced substantial DeSci theater (token-gated journals, governance-token DAOs with no scientific output, NFT "research crowdfunding" schemes with no IP or deliverables). Substantive LM-class signal resides in: Open Science Framework (OSF; Center for Open Science, 2013+; ~700K registered researchers; preregistration + replication infrastructure); Polymath Project (Tim Gowers + Terence Tao, 2009+; mass-collaboration mathematics; Polymath8 bound on prime gaps via Zhang/Maynard); Lean/Mathlib (2013+; formal proof commons; ~600 contributors 2024; mathematical reasoning as open-source codebase); Folding@home / BOINC (distributed protein-folding / climate / astro compute); LabDAO (2022+; "compute-to-discovery"; Ethereum + cloud lab automation); VitaDAO (2021+; longevity IP cooperatively commercialized via IP-NFT framework); Molecule.to (2021+; IP-NFT framework for longevity, mental-health, women's-health DAOs); Gitcoin Grants (retroactive public goods funding for open-source research, 2019+; quadratic voting allocation). Four mechanism pillars of the LM-class form: (1) Open-protocol publication — bioRxiv/medRxiv/arXiv-style preprints + post-publication community review (PubPeer, Sci-Hub as pirate substrate; OSF preregistration; ResearchHub tokenized review). Adjacency from arXiv-preprint-infrastructure-1991 (DM-Day): the LM lift is not faster open access but *cooperative accountability* post-publication — community review replaces journal peer-review as the primary epistemic quality signal. (2) Cooperative IP — IP-NFT frameworks (Molecule, VitaDAO); patent-pools (Open COVID Pledge; CERN Open Hardware); defensive-publication networks (Defensive Patent License). Adjacency from github-code-collaboration-2008 (DM-Day): same copyleft/open-license logic applied to biomedical IP and hardware, where the MM patent system was the prior gatekeeper. (3) Distributed compute + data infrastructure — Folding@home (~1.8M active user CPUs/GPUs 2024); BOINC-style citizen-science computing; Zooniverse (2.4M registered volunteers; 80+ active research projects 2024); iNaturalist (3M+ observations/month 2024; biodiversity data commons); SETI@home (historical 2020 close; proof-of-distributed-compute-for-science). LabDAO extends this to wet-lab automation: researchers contribute cloud compute access to pooled laboratory protocols rather than purchasing institutional access. (4) Mass-collaboration mathematics / formal-proof projects — Polymath Project (2009+; Gowers's "massively collaborative mathematics" hypothesis); Lean/Mathlib (mathematical formalization commons; proof verification as LM epistemic-quality substrate; the "argument-of-progress" in formal form — claims are machine-verifiable, not just peer-reviewable). Adjacency-lift mechanism (Hidalgo §3): arXiv (open-access preprint) + GitHub (distributed collaboration) + IPFS (decentralized storage) + DeFi (research-funding via quadratic-funding/retroactive-PG) are individually DM-Day machines. The LM-class adjacency lift yields: *civic-science cooperative* with (a) community-owned protocols rather than platform-owned gatekeeping; (b) cooperative IP rather than patent monopoly; (c) distributed compute as public-goods rather than commercial-compute-as-service; (d) machine-verifiable proof rather than credential-mediated authority. No single DM ancestor produces this configuration — the adjacency requires the coupling of all four. Emergence_subtype [v0.2-gap — recorded here]: cooperative_commons. The class instantiates an emergent public-goods form analogous to the OSS ecosystem but applied to scientific knowledge, biomedical IP, and formal proof. Capture-resistance [EXTRAP]: capture_resistance_index MEDIUM-HIGH (0.60). Community-owned protocols resist capture structurally (IP-NFT frameworks assign ownership to DAO, not platform); but token-bubble dynamics demonstrated that governance-token DAOs are vulnerable to token-holder extraction — Molecule.to restructured in 2023 after treasury depletion; VitaDAO governance was contested in 2023 by token-accumulation pressure. OSF is the highest-resistance instance (501(c)3, not tokenized; COS governance is non-profit-accountable). Proletarianization risk [EXTRAP]: MEDIUM (0.48). Formal-proof infrastructure (Lean/Mathlib) is technical_memory: the proof library persists even if the mathematical community does not actively extend it. Citizen-science observation data (iNaturalist, Zooniverse) is corpus-persistent without the volunteer community that generated it. The substantive risk is community re-internalization thinning: if the graduate-student cohort that maintains Lean/Mathlib moves to industry AI roles, the proof-commons persists as artifact but degrades as living practice (Stiegler proletarianization signal). Lower than OSS (0.55) because the DESCI community includes more intrinsically-motivated participants and the math-proof niche is smaller and more cohesive. LM mechanism signatures: capture_resistance_index 0.60 [EXTRAP]; proletarianization_risk 0.48 [EXTRAP]; rewilding_fraction LOW (science is epistemically not ecological; the "rewilding" analogue is epistemic_pluralism_rising); liveness_temporal_coupling HIGH 0.70 (post- publication review per-citation; community annotation is real-time); new_nature_density LOW (science produces knowledge not new-nature assemblages). pace_layer: culture (primary; epistemic-pluralism layer — peer review norms, open-access standards, scientific-method protocols evolve at culture speed, slower than fashion, faster than governance). Governance is secondary: funding- allocation reform via quadratic funding, DAO governance, and open-grant mechanisms operates at governance pace-layer speed. Cross-era position: the class draws substrate_provision from MM Humboldtian university (credentialed contributor pool, peer-review-norm substrate); experiences hostile_inheritance from MM Royal Society / journal-publisher apparatus (journal embargo, impact-factor gatekeeping, retraction politics); draws adapted_inheritance from DM arXiv (open-preprint norm); draws mutualistic_ coupling from DM GitHub (distributed-collaboration tooling); draws substrate from DM post-humboldtian-research-university-1810 (the university as the credential-granting machine whose outputs feed the DESCI contributor pool even as DESCI tries to circumvent it). All quantitative state-variable values are [EXTRAP]. CANON framing applies to: OSF/COS 2013+ existence; Polymath Project 2009+ existence; Lean/Mathlib 2013+; Folding@home / BOINC / Zooniverse / iNaturalist statistics as cited.