BT Clip: AGI Pioneers: ChatGPT Won’t Get Us to AGI.. Here’s What Will (Ben Goertzel)
The Beyond Tomorrow Podcast with Julian IssaShow Notes
Are large language models like ChatGPT and Claude enough to create AGI?In this clip, Ben Goertzel explains why today’s LLMs are incredibly powerful, but still fundamentally different from true artificial general intelligence.He argues that simply building a bigger version of ChatGPT or Claude is unlikely to produce human-level AGI. Instead, reaching AGI will require new AI architectures that can go beyond their training, generalize to unfamiliar problems and make genuinely novel leaps in reasoning.Ben also explains why current models can already outperform humans in areas like coding, mathematics and knowledge retrieval, while still struggling with the kind of imagination and general intelligence humans use to push beyond existing knowledge.
Full episode here: https://open.spotify.com/episode/7wGN1x0vQQgdzoktHipqBK?si=454702357d55415e
More from The Beyond Tomorrow Podcast with Julian Issa
Is “AI will destroy humanity” a real risk, or a story we tell ourselves because we haven’t done the hard engineering?
Do we really know enough about GLP-1 drugs? Mike Israetel breaks down the risks and trade-offs, from anhedonia and calorie-deprived diets to cognitive distortion and the myth of a “zero-harm” default.
What if the fastest way to a longer life isn't a miracle drug, but a wartime mentality against aging?
Are we living in a simulation?
What happens when millions of people hand their lives to AI agents, and nobody has built the rules of the road?