(无标题)
We’re working to build as much compute as we can, as quickly as we can, over the next few years.
We believe that the world continues to underestimate AI demand: we’re already compute bottlenecked on launching new features, and exponential progress of model capability continues.