The Risks of Sharing Data with AI

This week Nate and Maria discuss the release of GPT-5, the latest model from OpenAI. This model promises to be faster, smarter, and more useful while also reducing hallucinations and sycophancy. It also lets users choose among different AI “personalities.” What do Nate and Maria think so far?

Then, they turn to the newly inked Nvidia trade deal, which notably includes a 15% cut of sales to China for the US government

Further Reading:

Ethical Issues In Advanced Artificial Intelligence by Nick Bostrom, 2003

SuperIntelligence: Paths, Dangers, Strategies by Nick Bostrom, 2014

For more from Nate and Maria, subscribe to their newsletters:

The Leap from Maria Konnikova

Silver Bulletin from Nate Silver

See omnystudio.com/listener for privacy information.