Jim talks with Ben Goertzel about a paper he co-wrote, "OpenCog Hyperon: A Framework for AGI at the Human Level and Beyond." They discuss the way Ben defines AGI, problems with an economically oriented definition, the rate of advancement of a society, the history of OpenCog, mathematical models of intelligence, Jim's early use of OpenCog, a distributed Atomspace, Atomese vs MeTTa languages, knowledge metagraphs, why Ben didn't write a custom programming language for the original OpenCog, type theory, functional logic programming, moving from weirdly ugly to weirdly elegant, technical debt, grounding of Atoms, interfacing Hyperon with LLMs, nourishing a broader open-source community, hierarchical attention-based pattern recognition networks, heuristic induction, cognitive synergy, why scalability requires translating declarative representation into procedural form and vice versa, retrieval-augmented generation, predictive-coding-based learning as an alternative to back-propagation, the possibility of an InfoGAN-style transformer, and much more.