I know this comparison isn’t exact, but I didn’t attempt to engage with the supposed latency measure. That kind of measurements are really variable and hard to

I know this comparison isn’t exact, but I didn’t attempt to engage with the supposed latency measure. That kind of measurements are really variable and hard to pin down. You’d need percentiles to make even an uneducated judgement, not a single random sample. Still… a 100KiB bundle is not that bad of a starting point. It objectively isn’t, if you measure it up to other popular frameworks.