OpenAI, Thinking Machines and China's Telos are all building full-duplex models you can interrupt
kimmonismus · x · 2026-10-23
- A thread on full-duplex models: AI that keeps listening while it speaks, so a user can correct it mid-answer ("actually, use the other file") and it adapts within the same interaction, no turn-taking required.
- The author names three players: OpenAI, Thinking Machines Lab, and China's Telos.
- Core argument: for real-time agents, training the timing of interaction matters as much as the answer itself.
Related event: OpenAI, Thinking Machines and Telos Race Toward Full-Duplex Models(2 posts)→
More from Models
- ChatGPT web can actually listen to audio, but the same model can't via API or Codex — Possible_Statement84 · 2026-10-23
- MiniMax M3.1 Flash free for 3 days on 0G, with 1M context and 512K output — MiniMax_AI · 2026-10-23
- TLA+ expert corrects Grok: TLC only exhausts finite state spaces, not all behaviors — tianyin_xu · 2026-10-23
- OpenAI blocks manual reset with 13% usage left: 'no reset needed right now' — AccomplishedSugar490 · 2026-10-23
- Meta's 816B Muse Spark Open scores 39, losing to 180B Qwen3.8-Flash-Next at 40 — BTA_Labs · 2026-10-23
- Gemini refuses to generate image in Docs, blames document size limit — viksit · 2026-10-23