Mira Murati Presents Interaction-Model Vision for Thinking Machines

Mira Murati, CEO of Thinking Machines Lab, made her first major public appearance in 18 months during a Bloomberg interview, previewing the company's "interaction models" that process continuous streams of audio, text, and video in 200-millisecond intervals. The company has spent the past year and a half raising capital, hiring researchers, and shipping one product, Tinker, an API for fine-tuning open-source AI models. Murati framed the interaction-model work as an early step without providing a release date, and addressed the November 2023 OpenAI board episode that employees called the "blip.

Mira Murati Presents Interaction-Model Vision for Thinking Machines TechCrunch reports that Mira Murati made her first major public appearance in roughly 18 months in a Bloomberg interview, where she previewed Thinking Machines Lab's technical direction. Per TechCrunch, the company has spent about a year and a half raising capital, hiring researchers, and shipping one product, Tinker, an API for fine-tuning open-source models. In the interview Murati previewed what Thinking Machines calls "interaction models," which she described as systems that process continuous streams of audio, text, and video in 200-millisecond intervals to capture conversational texture; she framed the work as a first step and did not give a release date, according to TechCrunch. The piece also notes Murati addressed the November 2023 OpenAI board episode, which employees called the "blip." What happened TechCrunch reports that Mira Murati, CEO of Thinking Machines Lab , made her first major public appearance in roughly 18 months during a Bloomberg interview in San Francisco. TechCrunch writes that Thinking Machines has spent the past year and a half raising capital, hiring researchers, and shipping one product, Tinker, an API for fine-tuning open-source AI models. In the interview, Murati previewed the lab's concept of "interaction models," which she described as designed to process continuous streams of audio, text, and video in 200-millisecond intervals. TechCrunch reports that Murati framed this as an early step rather than a finished product and declined to provide a specific release date. The article also records that she answered questions about the November 2023 week at OpenAI, an episode employees reportedly called the "blip," and said she felt clear about her decisions in each moment, per TechCrunch. Editorial analysis - technical context Interaction models as described require low-latency, multimodal streaming inference, and tightly integrated front-end capture audio, vision, text with model-stack state management. Companies pursuing similar designs commonly invest in streaming ASR, continuous vision pipelines, chunked attention or stateful memory layers, and orchestration for sub-200-ms responsiveness. Those architectural choices increase operational complexity, raising compute and engineering costs compared with turn-based, batch prompt flows. Industry context Reporting places Murati's public reappearance against a backdrop of intense competition from OpenAI , Anthropic , and xAI , all of which continue to dominate headlines and hiring. Industry observers have noted that founders and labs emerging from stealth or spinouts often use controlled public appearances to move from quietly building to signalling capabilities and attracting partners, talent, and customers. What to watch Indicators that will clarify Thinking Machines' technical progress include latency and throughput benchmarks for live multimodal streams, published demos showing interruption and mid-thought handling, developer adoption metrics for Tinker, open models or checkpoints, and any partner integrations with real-time platforms. Observers should also watch for more detailed technical disclosures that show how the lab manages persistent state, cost per minute of interaction, and privacy controls for streaming audio and video. Note: Quotations and product details above are reported by TechCrunch, which summarized a Bloomberg interview with Murati. Thinking Machines has not provided a dated public release timeline in the reporting. Scoring Rationale Notable industry development because a high-profile founder outlined a concrete technical direction for a new lab, but the disclosure is early-stage and lacks release timing or benchmarks, limiting immediate practitioner impact. Practice interview problems based on real data 1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with. Try 250 free problems /problems