Standalone record of the Viridis Canon — a direct extension of the core Intelligence Bound (the two-body generalization). The core IB canon spine is unchanged (frozen at v9, record 20724138) ; this record links to the spine via isDerivedFrom the concept DOI 10. 5281/zenodo. 19317982. The Intelligence Bound dI/dt ≤ P·D/ (kB T ln 2) has, in each of its thirteen prior self-applications, constrained a single intelligence (this is the fourteenth). The central thesis of conservation technology is that artificial intelligence and the living biosphere are not master and resource but partners. This paper supplies the rate law for that partnership. Fusing bipartite information thermodynamics (Horowitz–Esposito, with the 2024 information-arbitrage refinement), the thermodynamics of prediction (Still–Sivak–Bell–Crooks), and the Good Regulator Theorem (Conant–Ashby), it proves the Symbiotic Intelligence Bound (SIB): two coupled dissipative learners can each raise the rate at which they accumulate permanence-intelligence above their standalone ceiling by harvesting the predictive information the partner carries about the shared world, dIA/dt ≤ PA DA/εL + Φ_ (E→A). Three laws govern the subsidy: (i) No free lunch — at the coupled fixed point the information one partner consumes is, bit for bit, what the other regenerates (İA = −İE), paid in dissipation; (ii) Complementarity — the rate subsidy vanishes at equilibrium and is nonzero only across distinct fluctuation sources, a thermodynamic derivation of ecological niche differentiation (cooperation requires complementarity, not similarity) ; (iii) Threshold — a sharp Symbiotic Transition partitions couplings into mutualistic (super-additive learning) and parasitic. The extraction efficiency is the universal cos²Θ Fisher–Rao alignment factor. The paper gives the Good Regulator Theorem its first thermodynamic, bidirectional form — each partner is the other's good regulator, capped by the Intelligence Bound — naming the biosphere-as-Regulator the 14th Intelligence-Bound self-application. Scope. Lean 4 (Aristotle-verified, zero sorry, axioms ⊆ propext, Classical. choice, Quot. sound) certifies the structural results — the joint-bound recovery, NESS reciprocity, no-perpetual-learning dissipation positivity, the mutualism/parasitism threshold, and the bounded cos²Θ extraction efficiency — not empirical magnitudes. 18/18 numeric checks pass. Working paper; not peer-reviewed.
Hart et al. (Fri,) studied this question.