Shownotes
Nick and Jack cover three stories. Google DeepMind previewed Genie 4, a world model that builds explorable 3D scenes with their own sound. Amazon reportedly told managers that token counts are a poor way to measure AI use, after staff gamed an internal leaderboard. Anthropic posted early lab results from six drug-development partners that use Claude. Then they answer your questions on agent threads, 3D world models, voice agents and open models.
Watch on YouTube: https://youtu.be/4dZHwgXnNv4...

