Thinking Machines Lab, the AI startup founded last year by former OpenAI CTO Mira Murati, announced on Monday a new development called "interaction models." Essentially, this technology aims to enable AI to simultaneously listen while it talks.
Currently, all AI models operate on a turn-taking basis: you speak, it listens; it responds, you listen. Thinking Machines Lab is challenging this paradigm by developing a model that processes user input and generates a response concurrently. This approach is designed to make AI conversations feel more like a real-time phone call than a sequential text exchange.
Technically termed "full duplex," the company claims its TML-Interaction-Small model achieves a response time of just 0.40 seconds. This speed is roughly on par with natural human conversation and is significantly faster than comparable models from industry leaders like OpenAI and Google.
However, it's important to note that this is currently a research preview, not a public product. Thinking Machines Lab has indicated that a "limited research preview" will be available in the coming months, with a wider release planned for later this year.
The underlying idea that interactivity should be an inherent feature of a model, rather than an add-on, is intriguing, and the benchmarks are impressive. Nevertheless, whether the real-world user experience will fully align with these technical claims remains to be seen until the technology becomes more widely accessible.
