Ben Goertzel: Artificial General Intelligence

The episode begins with Ben Goertzel discussing the influence of science fiction on his understanding of intelligence and the potential for intelligences beyond human comprehension. He cites works like Stanisław Lem's Solaris and the original Star Trek series as pivotal in shaping his vision of AI. Goertzel's fascination with the unknown and the incomprehensible drives his motivation to pursue AGI, suggesting that current human understanding may be limited by our cognitive biases and cultural constraints.

Goertzel argues that the current focus on neural networks in AGI research is misguided, emphasizing the need for a diverse approach that includes symbolic learning and other methodologies. He highlights OpenCog's hypergraph-based architecture as a promising alternative, capable of dynamically modeling complex systems in ways that static neural networks cannot. This approach, he suggests, could lead to more adaptive and scalable AI systems, crucial for achieving true AGI.

Lex Fridman does not directly challenge Goertzel's assertions about the limitations of neural networks, but the conversation implies a tension between the mainstream AI community's current trajectory and Goertzel's vision. The lack of direct pushback on Goertzel's critique of neural networks suggests a broader acceptance of his argument or a missed opportunity for deeper exploration of counterarguments.

The discussion concludes with Goertzel's reflections on the philosophical implications of AGI, particularly the potential to redefine human values and ethics. He advocates for significant investment in longevity research, positing that future generations may view the inevitability of death as an outdated concept. The conversation shifts towards the potential societal impacts of AGI and the ethical considerations that must guide its development, leaving open questions about the balance between technological advancement and human values.