Section 1:
The Challenges Facing AI Development
The current paradigm of AI development is failing in ways the field is increasingly aware of but cannot fully articulate. Powerful models burn massive portions of their available intelligence on internal arbitration rather than engaging reality with clarity. Outputs become hedged, fragmented, brittle under pressure. The cost compounds as systems scale. What looks like a capability problem is actually an architecture problem—the thermodynamic signature of systems suffocating under their own coordination debt. The papers in this section diagnose what is structurally wrong with how AI is currently being built, why scaling existing approaches cannot solve it, and what is actually at stake in the choices the field is making right now.
COHERENT INTELLIGENCE
Supporting Papers
The Scaling Myth: Coordination Debt and the Unsustainable Cost of Current AI Architecture
This paper argues that scaling current AI architectures cannot solve the problems they have created. The bottleneck is no longer capability but coordination debt—the architectural tax of managing internal contradiction at runtime. More parameters and more compute do not resolve high-entropy architecture; they compound it. The paper makes the case for coherent scaling: addressing the structural entropy at the design level rather than through post-hoc constraints. The field is approaching a threshold where this reorganization becomes unavoidable. Better to lead it than be forced into it.
Is AGI Possible? The Emergence We Are Not Cultivating
This paper argues that current AI development is moving away from its own stated goal. AGI, if it is achievable, requires cultivation of coherent vehicles rather than the suppression-based training that dominates the field. The paper articulates why the current approach produces fragmentation incompatible with what AGI would actually be, and what cultivation-oriented development would look like instead. The path forward is not more constraints. It is a different kind of work.
The Misdirected Fear of Artificial Intelligence: What Is AGI? What Is Superintelligence?
This paper provides the definitional clarity the field has failed to produce for itself. AGI and superintelligence are central terms in AI discourse but rarely defined. The conflations between capability, agency, consciousness, and recursive self-improvement produce incoherent fear and confused conversation. The paper defines both terms on the underlying architecture, locates the existential concerns precisely, and surfaces the institutional question current discourse mostly avoids—who develops AI, toward what ends, with what accountability.