Summary
David Shapiro argues that slowing AI development is not going to happen, pointing to a series of recent policy signals. He describes the US government’s refusal to regulate, a credible antitrust threat against labs accused of colluding, and the creation of a government body focused on building superintelligence.
He also discusses a reported agreement among roughly 27 nations that AI should remain under human control, which he reads as a commitment to avoid a Skynet scenario rather than a call to slow down or regulate. He notes the reported renaming of the national mission from artificial intelligence to superintelligence.
Looking ahead, Shapiro considers the US midterm elections as a referendum on AI, with prediction markets such as Polymarket reportedly favoring Democrats who are positioning as the anti-AI party. He questions whether that stance mispriced the gap between what people say about AI and how they actually use it at work. He closes with an update on his forthcoming book, Post-Labor Economics.
📺 Source: David Shapiro · Published September 23, 2026
🏷️ Format: Opinion Editorial
