Episode 2289: Gary Marcus on how Artificial General Intelligence (AGI) is, in the long run, inevitable

Dec 31, 2024

Gary Marcus, a prominent AI skeptic and former CEO of Geometric.AI, discusses the inevitability of Artificial General Intelligence (AGI) by 2100. He highlights the urgent need for AI regulation before technology surpasses human control. The conversation dives into the complexities of the current AI landscape, the alignment problem, and the rivalry among tech giants like OpenAI and Google. Marcus critiques the focus on generative models and emphasizes that the path to AGI might require yet-to-be-invented approaches, urging a balance of methodologies in future development.

Ask episode

AI Snips

Chapters

Books

Transcript

Episode notes

ANECDOTE

Historical AI Approaches

Marcus highlights two AI approaches: symbolic (System 2) and neural networks (System 1).
Both have historically failed to deliver reliable AI, necessitating a combined approach.

INSIGHT

OpenAI's Challenges

Gary Marcus predicts OpenAI's potential downfall due to financial losses and lack of competitive advantage.
He suggests the winning AI technology might not yet exist.

INSIGHT

Facebook's Open-Source AI

Marcus questions the value of Facebook's open-source AI, particularly its fixability.
He believes Zuckerberg's unilateral decision to release it, especially to China, could be detrimental.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

Gary Marcus is amongst the world’s leading skeptics on the AI revolution. So it’s worth taking note when Marcus admits that “of course we are getting to AGI eventually”. No, he says, artificial general intelligence (AGI) won’t take place in 2027 or perhaps even 2050. But it will happen, he confidently predicts, by 2100. So that only underlines Marcus’ argument, made in his acclaimed 2024 book Taming Silicon Valley, of the desperate need to regulate AI before it regulates us. And it also contextualizes our short term preoccupation with corporate pioneers of generative AI technology like OpenAI, Anthropic, DeepMind and xAI. As Marcus argues, it’s likely that the dominant AI technology that will get us to AGI by the end of the 21st century hasn’t even been invented yet.

Gary Marcus is a leading voice in artificial intelligence, well known for his challenges to contemporary AI. He is a scientist and best-selling author and was founder and CEO of Geometric.AI, a machine learning company acquired by Uber. A Professor Emeritus at NYU, he is the author of five previous books, including the bestseller Guitar Zero, Kluge (one of The Economist's eight best books on the brain and consciousness), and Rebooting AI (with Ernest Davis), one of Forbes's seven must-read books on AI.

Named as one of the "100 most connected men" by GQ magazine, Andrew Keen is amongst the world's best known broadcasters and commentators. In addition to presenting KEEN ON, he is the host of the long-running How To Fix Democracy show. He is also the author of four prescient books about digital technology: CULT OF THE AMATEUR, DIGITAL VERTIGO, THE INTERNET IS NOT THE ANSWER and HOW TO FIX THE FUTURE. Andrew lives in San Francisco, is married to Cassandra Knight, Google's VP of Litigation & Discovery, and has two grown children

Keen On is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit keenon.substack.com/subscribe