- Thinking Machines Lab has released its first AI model, aiming to break away from the traditional question-and-answer paradigm by processing audio, video, and text in parallel. The model processes data in 200-millisecond chunks and is designed to outperform OpenAI‘s GPT Realtime 2 and Google’s Gemini Live in terms of interaction quality.
- The company argues that current voice AI models like those from OpenAI focus too heavily on Q&A, which they claim limits the model’s ability to engage in dynamic, interactive conversations. Their new model is designed to facilitate more natural, conversational exchanges.
- This shift towards interactivity could potentially democratize access to advanced AI voice technologies and enable them to be used more effectively across various applications where real-time interaction is crucial.
Source Read original →
