Fluid Compute

This post reads like a chart in an Apple keynote presentation. Tell us what this really is.

Do they just stop billing you for `await` time & keep the lambdas warm?