Home / Technology / AI Learns to Interrupt: A New Conversation Style
AI Learns to Interrupt: A New Conversation Style
12 May
Summary
- New AI interaction models allow simultaneous input and output.
- Company claims response times mimic natural human conversation.
- A limited research preview is planned for the coming months.
Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has revealed its development of interaction models. This innovative AI aims to change how we communicate with machines by allowing for simultaneous input and response processing, a feature termed 'full duplex.'
The company's TML-Interaction-Small model reportedly achieves response times of 0.40 seconds. This speed is comparable to natural human conversation and reportedly surpasses current models from major AI players like OpenAI and Google.
While impressive in its technical claims, this is currently a research preview and not yet available to the public. Thinking Machines plans to release a limited research preview in the coming months, with a broader release anticipated later in the year.
