We’ve discussed the financial aspects of the AI boom and bubble in previous episodes, but today we’re trying to demystify the technology itself.

This week we’re joined by Gary Marcus, scientist and author of many books about AI and the Marcus on AI Substack. We discuss how chatbots like ChatGPT work, how they’ve advanced in recent years by harnessing large language models to more traditional symbolic AI systems, and also how the technology still feels stuck in the mud. We also discuss the economic problems these technologies have to overcome, whether there’s really demand for AI compute, get into the recent hype for AI “hacks” and math breakthroughs, and finally get into regulatory solutions, based on Marcus’s congressional testimony on AI.

Listen via Apple, Spotify, or wherever you get your podcasts.

David Dayen is the executive editor of The American Prospect. He is the author of Monopolized: Life in the Age of Corporate Power and Chain of Title: How Three Ordinary Americans Uncovered Wall Street’s Great Foreclosure Fraud. He co-hosts the podcast Organized Money with Matt Stoller. He can be reached on Signal at ddayen.90.

Matt Stoller is research director at the American Economic Liberties Project and the author of Goliath: The 100-Year War Between Monopoly Power and Democracy.