Editorial illustration for Thinking Machines develops AI that processes input and replies simultaneously
Thinking Machines develops AI that processes input and...
Thinking Machines Lab dropped its new model on Monday. Founded just last year by ex-OpenAI CTO Mira Murati, the startup calls it an "interaction model." The goal is explicit: to capture the messy, overlapping cadence of real talk, where a reply often begins before the first speaker has finished.
Murati's launch directly challenges the polite, turn-taking etiquette of incumbents like ChatGPT. The lab is betting users will prefer an AI built for interruption—a core, if sometimes grating, feature of actual conversation. It’s a rejection of the artificial pause. The argument from Thinking Machines is simple: strict query-processing creates a stilted, unnatural rhythm.
Related Reading
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
From Archives Digg relaunches as AI news aggregator, targeting busy users but faces adoption doubts
From Archives Method uncovers hidden coalitions in multi‑agent AI using mutual‑info graph
Common Questions Answered
What is Thinking Machines Lab's new 'interaction model' designed to do differently from ChatGPT?
Thinking Machines Lab's interaction model is designed to process input and generate replies simultaneously, mimicking the natural overlapping cadence of real human conversation where responses often begin before the speaker finishes talking. This directly challenges the turn-taking etiquette of incumbent AI models like ChatGPT, which follow a more formal query-response pattern that the startup argues creates a stilted and unnatural rhythm.
Who founded Thinking Machines Lab and what was their previous role in AI?
Thinking Machines Lab was founded by Mira Murati, who previously served as the CTO of OpenAI. The startup was established just last year, bringing Murati's expertise from one of the leading AI companies to this new venture focused on conversational AI.
Why does Thinking Machines believe users will prefer an AI built for interruption?
Thinking Machines argues that strict query-processing in traditional AI models creates an artificial pause that makes conversations feel unnatural and stilted. The company believes users will prefer an AI that captures the messy, overlapping nature of real talk, where interruptions and simultaneous processing are core features rather than limitations, making interactions feel more genuinely human-like.
How does the interaction model's approach to conversation differ from the 'polite, turn-taking etiquette' of current AI systems?
The interaction model rejects the artificial pause inherent in traditional AI systems that wait for a complete query before processing and responding. Instead of following strict turn-taking rules, Thinking Machines' model enables simultaneous input processing and reply generation, allowing for the kind of natural interruptions and overlapping speech that characterize actual human conversations.
Further Reading
- AINews Thinking Machines' Native Interaction Models — Latent Space
- Thinking Machines Lab Unveils 'Interaction Models' for Real-Time Multimodal AI — TechCrunch
- Mira Murati's Thinking Machines Lab Previews Full-Duplex AI Interaction — The Verge
- Defeating Nondeterminism in LLM Inference — Thinking Machines Lab
