EP 211 Ben Goertzel on Generative AI vs. AGI

6 snips

Nov 28, 2023

AI researcher and author Ben Goertzel joins Jim Rutt to discuss the strengths and weaknesses of modern LLMs, why LLMs alone won't achieve AGI, OpenAI's integrative system, solving LLM hallucinations, and the duality of heuristics and abstractions. They also explore the limitations of LLMs, the challenge of scalability, and the future of AGI with the OpenCog project.

Ask episode

Chapters

Transcript

Episode notes

Introduction

00:00 • 3min

The Role of Large Language Models in AGI Systems

03:05 • 6min

AI Hallucinations and Reality Discrimination

09:24 • 9min

The Funding of Anthropic and the Race Towards AGI

18:24 • 21min

The Process of Generating Ideas and the Limitations of AI

39:02 • 7min

Advancing Non-Backpropagation Training Algorithms

45:53 • 8min

OpenCog, Hyper-On, and the Future of AGI

53:25 • 15min

Jim talks with recurring guest Ben Goertzel about the ideas in his paper "Generative AI vs. AGI: The Cognitive Strengths and Weaknesses of Modern LLMs." They discuss the exponential acceleration of AI development, why LLMs by themselves won't lead to AGI, OpenAI's integrative system, skyhooking, why LLMs may be useful for achieving AGI, solving LLM hallucinations, why Google hasn't replicated GPT-4, LLM-tuning lore, what differentiates AGI from other forms of AI, conceptualizing general intelligence, Weaver's theory of open-ended intelligence, multiple intelligence, the Turing test & the Minsky prize, what LLMs aren't good at, the danger of defining AGI as whatever LLMs can't do, the derivative & imitative character of LLMs, banality, doing advanced math with GPT-4, why the human brain doesn't form arbitrary abstractions, the duality of heuristics & abstractions, adding recurrence to transformers, OpenCog Hyperon, using a weighted labeled metagraph, orienting toward self-reflection & self-rewriting, the challenge of scalability of infrastructure, acceleration on non-LLM projects, and much more. Episode Transcript JRS Currents 072: Ben Goertzel on Viable Paths to True AGI "Generative AI vs. AGI: The Cognitive Strengths and Weaknesses of Modern LLMs," by Ben Goertzel "OpenCog Hyperon: A Framework for AGI at the Human Level and Beyond," by Ben Goertzel et. al. Dr. Ben Goertzel is a cross-disciplinary scientist, entrepreneur and author. Born in Brazil to American parents, in 2020 after a long stretch living in Hong Kong he relocated his primary base of operations to a rural island near Seattle. He leads the SingularityNET Foundation, the OpenCog Foundation, and the AGI Society which runs the annual Artificial General Intelligence conference. Dr. Goertzel’s research work encompasses multiple areas including artificial general intelligence, natural language processing, cognitive science, machine learning, computational finance, bioinformatics, virtual worlds, gaming, parapsychology, theoretical physics and more. He also chairs the futurist nonprofit Humanity+, serves as Chief Scientist of AI firms Rejuve, Mindplex, Cogito and Jam Galaxy, all parts of the SingularityNET ecosystem, and serves as keyboardist and vocalist in the Jam Galaxy Band, the first-ever band led by a humanoid robot.