We paid a bit more of a fixed cost up front, but we designed the Gemini models to be very multi-modal from day one. There were areas where the strengths started showing. Nano Banana was an example of this—you were able to see it all together. It's an amazingly dynamic frontier.
There are two to three labs pushing each other pretty vigorously. At any given month, we might feel like we've done something well, then realize there are a couple of things we're behind on. The picture will be dynamic again in a few months. The frontier is intense, as you would expect it to be. That's how I think about it.
When I talk to researchers who aren't at Google or the other labs, one of the things they commonly mention is that they feel the difference between the two or three other labs and the Google team is that Google is not as 'AGI built.' In other words, there's less of a belief at Google that AGI is right around the corner, and less focus on accelerating through it. The people at Google are obviously thinking deeply about that question.
Do I think that's true, and does it impact how people think about the future and what they build? I think we have probably scaled our CapEx from $30 billion to approximately $100 billion—$80 billion. You don't do that unless you think about the curve a certain way. I see it as largely semantics, maybe because we are a larger company with a lot of products that touch so many areas.
