• Thinking Machines introduces “interaction models” built for real-time human-AI collaboration across audio, video, and text. • The core problem is that today’s AI is mostly turn-based, which makes collaboration feel slow and rigid. • Their model uses 200ms “micro-turns,” so it can listen, watch, speak, interrupt, and react continuously. • A separate background model handles deeper reasoning, tools, browsing, and longer tasks while the interaction model stays responsive. • Early benchmarks show strong responsiveness and interaction quality, but limits remain around long sessions, connectivity, safety, and scaling larger models. #ThinkingMachines #AI #LLM