Thinking Machines on Real-Time AI Interaction Models

The reason why we even started Thinking Machines is because we wanted to build a frontier AI lab that's really focused on, um, really human AI collaboration piece, and that means a lot of things, and there are specific research bets that go into that.

But we wanted to showcase, um, our work in one of the first bets through the interaction models.

And the interaction models are a new kind of model.

If you consider the types of models that we work with today, they're very, um, turn-based. Mm.

And so you talk, they talk, then they go off and think.

Um, once you've given sort of a prompt on what it is that you want to do, and while they're thinking, it's almost like they're deaf and blind. Mm.

They cannot perceive anything else about what's going on, and then it's your turn, and while you're talking, um, they really cannot perceive anything about how you're talking.

It's not happening in real time. And by contrast, our interactions with each other are very rich.

There is a lot of information in our interactions when we are silent, when we're thinking, when we're interrupting one another.

And so interaction models are able to capture all of these neurons.

They're not turn-based. They're more like time-based- Oh ... interaction, where they're continuously taking in audio, text, video, and continuously providing output.

Now, we cut this up in chunks of twenty milliseconds, and this enables you to actually, uh, catch these things like interruptions and simultaneous speech, um, and really create a rich, high bandwidth interaction between humans and machines- Mm ... which we think is critical to actually enable,

um, agency and, uh, and, and enable people to be more in this loop- Mm-hmm ... um, as we advance AI further and further.