ODTOE Formalism
ODTOE Formalism
Multi-agent coherence in AI systems through ODTOE
Multi-agent coherence in AI systems through ODTOE
About this video
This English-language video addresses multi-agent coherence in AI systems through the ODTOE formalism. The audience is entrepreneurs, engineers and philosophers interested in AI architecture. Key concepts covered include: the formal mathematics of multi-agent coherence, why current multi-agent LLM systems lose coherence in long-horizon tasks, the ODTOE protocols for engineering sustained coherence across agents, the role of the common tree and absolute-observer constraints in AI alignment, the practical implementation in current AI stacks, and the implications for the path from current AI to AGI. The talk closes with the synthesis that the central engineering problem of contemporary AI is coherence engineering, and that ODTOE provides the first formal framework that takes this problem seriously rather than treating it as an emergent miracle.