Thinking Machines Lab says it’s building full duplex AI , which means an AI system can take in what someone is saying while generating a response. In plain English, it’s closer to a phone call than a walkie-talkie. The startup, founded last year by former OpenAI CTO Mira Murati, announced interaction models, starting with TML-Interaction-Small. It says the system can respond in 0.40 seconds, a pace that puts it near ordinary human back-and-forth. There’s a catch for anyone hoping to try it today. This remains a research preview, with limited access planned in the next few months and a broader release expected later this year. A faster kind of AI exchange The core idea is easy to understand, and the change is meaningful. Instead of waiting for someone to finish speaking before working on an answer, the model processes incoming speech while preparing its response. Unsplash That delay matters because pauses make AI assistants sound artificial.…