Mira Murati, a figure known more for her behind-the-scenes engineering prowess than public pronouncements, recently stepped back into the spotlight. As the former CTO of OpenAI and now CEO of her own venture, Thinking Machines Lab, her media appearances have been notably rare. So, when she granted an interview to Bloomberg in San Francisco on June 4, 2026, it marked her first major media engagement in roughly 18 months, signaling a new phase for her innovative company.
Murati, who previously served six years as OpenAI’s CTO, has deliberately kept Thinking Machines Lab operating in stealth mode. Over the past year and a half, the company has focused intensely on core development: securing capital, recruiting top-tier researchers, and shipping its initial product, Tinker, an API designed for fine-tuning open-source AI models. This quiet period allowed for foundational growth, but the rapidly evolving AI landscape now calls for a more visible presence.
Thinking Machines Steps into the Ring
The timing of Murati’s re-emergence is no coincidence. The AI industry has seen an explosion of activity, with rivals like OpenAI constantly dominating headlines. Anthropic’s momentum is a frequent topic of discussion, and Elon Musk’s xAI, now integrated into SpaceX, is gearing up for a significant public offering, drawing considerable attention and investment. In such a competitive environment, maintaining a low profile eventually yields diminishing returns, necessitating a strategic move to remind the market of Thinking Machines’ existence and innovation.
Murati used her Bloomberg interview to achieve precisely this, and little else. She offered a compelling preview of Thinking Machines’ upcoming innovation: what they call “interaction models.” These models, she explained to interviewer Emily Chang, represent a fundamentally different approach to AI interfaces, moving beyond the traditional turn-based prompt-and-response dynamic prevalent today. This groundbreaking technology is designed to process continuous streams of audio, text, and video, adapting in rapid 200-millisecond intervals.
Unveiling Next-Gen “Interaction Models”
The core idea behind these interaction models is to allow AI to truly understand the nuances of human communication in real-time. This means picking up on the natural texture of conversation, including interruptions, mid-thought corrections, and even pauses for contemplation. Such capabilities promise a more natural, fluid interaction with AI, blurring the lines between human and machine communication. Murati, however, carefully positioned this as a crucial “first step,” refraining from offering a specific release date for the full product.
Beyond product announcements, Murati also addressed the tumultuous events of November 2023 at OpenAI, an episode internally dubbed “the blip.” This period saw the firing of CEO Sam Altman and Murati’s brief, critical tenure as interim CEO. She asserted her clarity of purpose during that chaotic week, stating that her decisions were consistently guided by the mission to protect the company and its team, which made her choices “obvious” even amidst external pandemonium.
Reflecting on Leadership and AI Governance
Murati firmly believes that her involvement during that intense five-day stretch and its immediate aftermath prevented OpenAI from “imploding.” However, she acknowledged that clear intent doesn’t always equate to clear consequences. In retrospect, she expressed a desire to have pushed harder for more information, a clearer transition plan, and greater transparency, though she remained coy about her overall assessment of how things ultimately transpired. When pressed about her trust in former boss Sam Altman, she skillfully redirected the conversation.
Instead, Murati pivoted to a broader concern she holds deeply: the increasing concentration of consequential decision-making within a limited number of hands across the entire AI industry. Her worry, she elaborated, isn’t primarily about the character of individual leaders, though that certainly plays a part. Her greater apprehension lies in the absence of robust structural checks and balances. Murati emphasized that even well-intentioned individuals can make poor judgments, and organizations, however virtuous, can drift off course without proper oversight. She strongly suggested that too much emphasis has been placed on individual virtue and far too little on effective governance mechanisms.
Murati also tackled questions regarding the recent departures of several high-profile researchers from Thinking Machines, a topic she has largely avoided in public. She downplayed the significance, attributing it to the inherent volatility of building a frontier AI lab from the ground up, compressing years of organizational flux into mere months. While acknowledging the allure of the “nine-figure packages” that have become standard in the fierce war for AI talent, she suggested that compensation rarely tells the whole story.
The Human Hand on the Wheel of AI’s Future
With a touch of humor, Murati shared her own competitive philosophy: “When I wake up in the morning, I am not thinking about how to kill the competitor.” This highlights her focus on innovation rather than direct rivalry, a strategy that resonates with the unique challenges of developing cutting-edge artificial intelligence. Her nuanced perspective on industry dynamics further cements her reputation as a thoughtful leader in a field often characterized by bold, aggressive claims.
Naturally, the conversation broadened to the future of AI itself, including its impact on humanity. Murati, born in Albania and speaking with a slight Eastern European accent, offered a measured and thoughtful response. She rejected the simplistic framing of an inevitable AI dystopia or utopia, asserting that neither outcome is predetermined. Instead, she argued that the present moment is the critical juncture that will shape the trajectory of AI’s future. Her consistent message throughout the interview was clear: if humans relinquish control too soon, the future will undoubtedly look different, and not for the better. The ongoing human hand on the “wheel” is paramount.
Source: TechCrunch – AI
