The Next Experiment: The Research Discussion Series

Ten research proposals identifying structural gaps in active AI programs. (1) Three consciousness theories (IIT, GWT, HOT) share a deficiency-substrate condition that generates testable predictions including an organoid boundary. (2) The signal-origin principle: methods converge under AI only within clusters sharing an entropy source; across clusters, domains couple but do not integrate. (3) Connective abduction formalized as a five-step method; demand-based verification and signal-origin constraints address the replication crisis. (4) The Stability Thesis in reverse: malice is as stable as goodness across tool transitions; the only structural answer is distributed accountability. (5) The human role in agent workflows is signature, not review; accountability weight is a natural throughput cap. (6) Model collapse is thermodynamic: data has measurable distance from its entropy source (physical=0, human=1, AI=2+); distance predicts collapse resistance. (7) The sim-to-real gap is thermodynamic: simulation is distance 2+ regardless of fidelity; biological precedents (bats, sharks) use direct physical coupling at distance 0. (8) Shannon-Boltzmann equivalence: the evidence chain is complete (Jaynes, Landauer, Berut, 2025 ASDF). (9) The entropy distance map: pairwise CKA across signal domains reveals which integrate and which only couple, redefining World Signal Sufficiency. (10) Expert-AI integration operates at the image schema layer; Go provides the existence proof. A thermodynamic chain connects Papers 8-6-9-7. Every paper includes experiments with explicit failure conditions. Companion to The Decalogy on Artificial Intelligence (Ahn, 2026; SSRN). Revision Note: All ten papers fully written (previously skeleton/outline for Papers 3–9). Paper titles and content substantially revised. Paper 3 now formalizes connective abduction with demand-based verification and applies Ioannidis's replication-crisis framework. Paper 4 replaced deployment architecture focus with the Stability Thesis in reverse: malice is as stable as goodness across tool transitions. Paper 5 reframed from outcome architecture to signature principle: accountability weight as natural throughput cap. Paper 6 introduces thermodynamic distance of data (physical=0, human=1, AI=2+) as the mechanism behind model collapse. Paper 7 reframed from DAC causal measurement to thermodynamic interpretation of the sim-to-real gap, with biological precedents for direct physical coupling. Paper 8 replaced carbon-silicon efficiency gradient with Shannon–Boltzmann equivalence and the complete evidence chain (Jaynes, Landauer, Bérut, 2025 ASDF). Paper 9 replaced WSS lower-bound measurement with the entropy distance map: full N×N CKA matrix redefining WSS as within-cluster integration plus across-cluster coupling. Paper 10 replaced Planetary Synchronization Rate with the schema bridge: expert-AI integration at the image schema layer, Go as existence proof. Thermodynamic chain established across Papers 8→6→9→7. SID metric deprecated throughout; DAC retained with case-by-case ROI. Signal-origin principle and eight new terms added to terminology.