If you had 20 engineers, 100,000 computers, and 3 months, what would you build?

Leveraging the resources of 20 engineers, 100,000 computers, and a 3-month timeline, I would focus on building a scalable, distributed system for large-scale data processing or machine learning. This could involve developing a cloud-native platform for AI model training and deployment, or a robust data analytics pipeline capable of handling massive datasets. The key would be to design for performance, reliability, and cost-efficiency, enabling rapid iteration and deployment of data-intensive applications.