Before CloudFront I had a setup that used custom HTTP server code to capture requests directly as S3 objects, then wait on an SQS queue for the response which would indicate the location of an S3 object containing the entire response verbatim. (Yes, each server could handle exactly one request at a time, call it a learning experience.) This was the closest I ever felt I got to a “static” website.

The goal was to eliminate redundant computation, and it did that, but there are much better approaches these days. What I ended up with was just bespoke, persistent caching.