Here’s what Mira Murati’s AI company is up to

2026-05-11 · Source: The Verge · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Fundamental Awareness, quick

Summary

Thinking Machines, an AI company founded by former OpenAI CTO Mira Murati, announced its development of "interaction models." These models aim to enable real-time, continuous collaboration between humans and AI by processing audio, video, and text simultaneously, allowing the AI to "think, respond, and act in real time." The company states that current models operate in a "single thread," limiting the bandwidth for human-AI interaction. Thinking Machines provided examples of interaction models in action, including identifying animal mentions in a story, real-time speech translation, and posture correction. A limited research preview is planned for the coming months, with a wider release targeted for later this year. Murati founded Thinking Machines in February 2025.

Key takeaway

For AI scientists and developers focused on human-AI collaboration, exploring Thinking Machines' interaction models could inform your next-generation interface designs. Your current models' "single thread" perception limits user input and AI output; consider how continuous, multimodal input could enhance real-time responsiveness and natural interaction in your applications. Prepare for the limited research preview to evaluate its potential.

Key insights

Interaction models aim to enable real-time, multimodal human-AI collaboration by continuously processing diverse inputs.

Principles

• AI should meet humans where they are.
• Real-time multimodal interaction increases collaboration bandwidth.

Method

Interaction models continuously take in audio, video, and text, then think, respond, and act in real time, overcoming the "single thread" limitation of current AI.

In practice

• Real-time speech translation.
• Contextual awareness (e.g., posture correction).

Topics

• Interaction Models
• Thinking Machines
• Mira Murati
• Real-time AI
• Multimodal AI

Best for: AI Scientist, Research Scientist, Director of AI/ML, AI Product Manager, General Interest

Related on AIssential

• New AI Model SHOCKED Me — Thinking Machines Labs' new model offers real-time, multimodal, and context-aware conversational AI.
• Mira Murati’s Thinking Machines Lab Introduces Interaction Models: A Native Multimodal Architecture for Real-Time Human-AI Collaboration — Native multimodal architectures can enable true real-time human-AI collaboration by integrating interactivity into model weights.
• Mira Murati steps back into the spotlight, carefully — Thinking Machines Lab is developing "interaction models" for real-time, continuous human-AI communication beyond turn-based prompts.
• Thinking Machines shows off preview of near-realtime AI voice and video conversation with new 'interaction models' — Thinking Machines' interaction models enable real-time, full-duplex human-AI communication by processing multimodal inputs and outputs simultaneously.
• [NOTICE] This feed is no longer maintained — Native interaction models enable real-time, multimodal human-AI collaboration by integrating interactivity directly into the AI's core architecture.
• Interaction Models: A Scalable Approach to Human-AI Collaboration — Native, real-time multimodal interaction within AI models enhances human-AI collaboration beyond turn-based systems.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Verge.