Multimodality · 2026

Interaction Models: A Scalable Approach to Human-AI Collaboration

Thinking Machines Lab

Replaces discrete turn-taking with a time-aligned ~200ms 'micro-turn' architecture over continuous audio/video/text streams, enabling simultaneous speech, interjections, and visual proactivity in a 276B MoE model.

Editorial record

Plain-language summary

Replaces discrete turn-taking with a time-aligned ~200ms 'micro-turn' architecture over continuous audio/video/text streams, enabling simultaneous speech, interjections, and visual proactivity in a 276B MoE model.

Knowledge graph

Relationships

Descendants

Source record

Provenance

Record ID
P-660
Record created
2026-07-13
Last reviewed
2026-07-14
Record version
2

Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.