Thinking Machines, founded by former OpenAI CTO Mira Murati, announced a new approach called 'interaction models' that process audio, video, and text simultaneously to enable real-time AI collaboration. Unlike current models that wait passively for users to finish typing or speaking, interaction models will perceive and respond to user actions continuously, mimicking natural human-to-human collaboration.
Why it matters: This represents a significant architectural shift in how AI systems perceive and interact with users, potentially reshaping the conversational AI experience and setting a new standard for real-time multimodal AI applications.