Thanks a lot for the article! I have a few notes: "No compromises": there's always compromises, but thanks for the marketing :) "Startup sub-millisecond": nbb

Thanks a lot for the article! I have a few notes: "No compromises": there's always compromises, but thanks for the marketing :) "Startup sub-millisecond": nbb and Node.js don't start sub-millisecond, but around 50ms for Node.js + another 50ms for nbb (on my Macbook Air M1). So that's the cost for Node + nbb for a cold started lambda + whatever baseline startup you'd get. "It has comparable performance to compiled ClojureScript programs": this really depends on your program and workload, so if performance matters a lot to you, always measure your specific workloads. In Ray's example, most of time was spent in calculations done inside the cryptographic libraries so that's why the performance was similar to that of JS directly. I thought I'd be the pedantic commenter before anyone else did it ;).