Thinking Machines Lab CEO Mira Murati stated that frontier AI must function like a tandem bike, requiring continuous collaboration between humans and machines throughout development. Speaking at Bloomberg Tech 2026 in San Francisco on Friday, the former OpenAI CTO argued that the industry needs to move beyond the concept of simply keeping humans in the loop.

Murati described the tandem bike analogy as a system designed for collaboration. She explained that both humans and AI must be actively involved, with the stronger entity pedaling harder during difficult tasks, but both retaining control over the direction. This approach contrasts with the idea of human oversight as a mere checkpoint.

Governance Over Individual Authority

Murati emphasized that AI's future is not a predestined outcome and dismissed predictions of either dystopia or utopia as oversimplified. She called for institutional checks and balances to supersede individual authority, asserting that governance and transparency are as critical as the character of the technology builders.

Reflecting on her experience as interim CEO during OpenAI's November 2023 board crisis, Murati noted that weak institutional design carries significant costs. She argued that the conversation around AI leadership often focuses too much on individual character rather than decision-making structures.

Real-Time Interaction Models

Murati pointed to Thinking Machines Lab's recently unveiled interaction models as an early demonstration of this philosophy. These models continuously process audio, text, and video in real time, facilitating a working relationship where humans and AI operate together rather than independently. She warned that removing humans from AI development now would make alignment harder to achieve as systems become more capable.