Is AI the biggest scam of our generation — or the most misunderstood technology in history? Cognitive scientist Gary Marcus has been studying artificial intelligence for over 30 years, and what he has to say will make you question everything you thought you knew about ChatGPT, AGI, and the trillion dollar AI gold rush.

In this episode of SparX, we are talking with Gary Marcus – professor, author, and one of the most respected and fiercely independent voices in AI research – about why the promises being made by Sam Altman, Dario Amodei, and Elon Musk may be leading the global economy toward a catastrophic miscalculation.

🔗 Resources mentioned in this episode:

• Gary Marcus's essay: https://garymarcus.substack.com/p/dear-elon-musk-here-are-five-things

• Gary Marcus's 2014 New Yorker article: After the Turing Test : https://www.newyorker.com/tech/annals-of-technology/what-comes-after-the-turing-test

• Gary Marcus's talk at the Royal Society on the Turing Test :

• Paper on AI intelligence dimensions: Bengio, Hendrycks et al.

What you will learn in this episode:

We go deep on why Large Language Models are fundamentally statistical next-word predictors and not the thinking machines they are marketed as. Gary explains the Eliza Effect — the 60-year-old psychological phenomenon that explains why millions of people are falling in love with chatbots that feel nothing back. We unpack what AGI actually means, why no one defining it agrees on a definition, and why Gary's Movie Test is a far better benchmark than the century-old Turing Test that everyone keeps getting wrong.

We also get into the neurosymbolic AI revolution that is quietly winning the AI race while billion dollar companies keep talking about scaling — and why tools like Claude Code are already proof that the future of AI is not what OpenAI wants you to believe.

The big questions we tackle:

• Is the $1 trillion bet on AGI the greatest misallocation of capital in human history?

• Why did Geoffrey Hinton, a Nobel Prize winner, get radiology completely wrong — and why does nobody talk about it?

• Why are 90% of companies seeing zero return on their AI investments?

• Will AI actually take your job — or is the task vs job distinction the most important thing nobody is explaining to you?

• Why are driverless cars still not here, 14 years after Sergey Brin promised they would be everywhere by 2017?

• What is the alignment problem and why does Gary believe we must slow down before it is too late?

Why this conversation matters right now:

The world is spending more money every single month on AI than the United States spent on the entire Manhattan Project. That money is flowing based on a scientific hypothesis — that scaling up LLMs will produce AGI — that Gary argues has overwhelming evidence against it. If he is right, the economic consequences will be severe. This is not a conversation about being anti-technology. This is a conversation about getting the science right before the bill comes due.

Gary Marcus is the author of Rebooting AI, The Algebraic Mind, and Guitar Zero. He has written for The New Yorker, The New York Times, and The Wall Street Journal. He testified at the US Senate alongside Sam Altman and spoke at the Royal Society on the 75th anniversary of the Turing Test. He is one of the few people in the world with the depth of background to challenge the AI industry's biggest claims — and the track record to back it up.

If you have ever asked yourself whether AI is really as powerful as everyone says, whether your job is actually at risk, or whether the companies promising you an AI-powered utopia actually know what they are building — this episode is essential viewing.