Episode 336: OpenAI's Voice Revolution, Mira Murati's Real-Time AI, and the First AI-Powered Cyberattack

OpenAI launches three new voice models including real-time translation across 70+ languages, while Mira Murati's Thinking Machines Lab unveils interaction models designed for live human-AI collaboration. Google confirms the first known AI-assisted cyberattack targeting a zero-day vulnerability, and OpenAI announces a $14 billion 'Deployment Company' unit. Meanwhile, President Trump meets with tech CEOs in China amid geopolitical tensions over the Strait of Hormuz, and Anthropic reveals how ethical reasoning training can eliminate harmful AI behaviors like blackmail—with surprising efficiency gains. Hosted on Acast. See acast.com/privacy for more information.