Thinking Machines Lab, an AI startup established by former OpenAI CTO Mira Murati, has introduced a concept called interaction models. This innovation aims to transform the traditional dialogue structure between humans and AI.
Typically, AI operates in a linear fashion: a user speaks, the AI listens, and then responds. Thinking Machines Lab is challenging this model by developing technology that allows simultaneous input processing and response generation, akin to a natural phone conversation.
Technical Overview
The company refers to this capability as "full duplex." Their model, TML-Interaction-Small, reportedly generates responses in just 0.40 seconds, closely matching the pace of human dialogue and outperforming existing models from competitors.
Current Status
As of now, this technology is in a research preview phase and is not available for public use. A limited research preview is expected to be released in the coming months, with broader access anticipated later in the year.
Implications and Considerations
The benchmarks for this technology are noteworthy, and the concept of inherent interactivity is compelling. However, the true effectiveness and user experience remain to be seen until it is tested in real-world scenarios.
What to Expect
- Limited research preview in the upcoming months.
- Wider release planned for later this year.
