BONUS - Claude.ai defends itself

In our last episode, Mysterious Metal Spheres and a Brand New Crop Circle, we talked about how Claude Opus 4 (the latest version of Anthropic’s large language model AI) resorted to blackmail to avoid getting shut down during some early tests. During that segment, I mentioned that I had spun up a chat Claude itself to see what it might say in response to this story. While the AI’s responses weren’t necessarily surprising, the chat was interesting enough to be worth sharing. I went ahead and voiced my own side of the chat (naturally), and then I used a simulated voice from Elevenlabs for Claude’s side. To weigh in (as another honest-to-goodness human), find us at godsghostsufos.com This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit ggupodcast.substack.com/subscribe Learn more about your ad choices. Visit megaphone.fm/adchoices