Thinking Machines、リアルタイム割り込みに対応する「フルデュプレックス」AIを発表
Thinking Machines Labは、入力の処理と出力の生成を同時に行うことで0.40秒の応答時間を実現する研究モデル「tml-interaction-small」を発表した。
thinking machines lab, founded by former openai cto mira murati, announced 'interaction models' designed to process input and generate output concurrently, eliminating the turn-taking latency of current systems. the company claims its tml-interaction-small model responds in 0.40 seconds, significantly outperforming comparable models from openai and google by operating like a phone call rather than a text chain. this approach makes interactivity native to the model architecture rather than a bolted-on feature. the company plans a limited research preview in the coming months ahead of a wider release later this year.