Mira Murati’s Thinking Machines Lab has finally stepped into view with a clear message: the future of AI may not be only about autonomous agents running tasks alone, but about systems that collaborate with humans in real time. Its new “interaction models” are designed for voice, video, and text working together in a live, streaming loop. That makes this launch important, because it challenges the current industry mood that treats agentic AI as the only serious direction.
At a basic level, this is a different philosophy of AI. Instead of asking the machine to disappear into the background and work on its own for long stretches, TML is trying to make AI behave more like a responsive teammate. It can listen, see, answer, pause, adapt, and even react to what is happening on screen or in conversation. For many everyday uses, that may actually be more practical than a fully autonomous agent.
This also reminds us of a larger truth about general-purpose technologies: they rarely move in just one direction. Even when one design becomes fashionable, parallel paths keep emerging because users have different needs. Agentic AI may be powerful for long workflows, but collaboration-first AI may prove better for teaching, meetings, creative work, healthcare, support, and live decision-making. In that sense, Murati is not rejecting the agentic wave so much as questioning its monopoly.
The real test now is not the idea, but the execution. If Thinking Machines can make this system truly fast, reliable, and useful without feeling gimmicky, it could define a major new category of AI. If not, it risks becoming another elegant demo in a field full of them. Either way, the message is clear: the AI future is not one road, but many.
THE WINNER WILL NOT JUST THINK SOLO — IT WILL LEARN TO WORK WITH US.
Sanjay Sahay
Have a nice evening.