[zgiber]

I feel that the comparisons are quite one dimensional (crushes Rust, Go etc..) picking one or two metrics. Why is the memory footprint almost fully ignored in the article? Application speed is rarely the bottleneck in platforms I work with, but it does matter a lot how many instances you can run in a cluster where memory is often a limitation. For me personally it's also important how easy is to write concurrent code to work around IO bottlenecks.