The incorrect alternative I meant is the idea that you can "just switch languages" or frameworks and get an instant speed boost. I mean, yes, sometimes your web server, caching proxy or CDN can act as a silver bullet, and rewriting can provide its own benefits if it simplifies the code, but assuming one language is better than another because "it ranks higher" is only one way of looking at the problem. It's more important to consider other factors, from low-level Linux to high level system architecture, perhaps with something like OpenTelemetry to pinpoint distributed hotspots, for example. Yes, you can learn something from these benchmarks, but alone benchmarks and frameworks can't speed up your app. :)