I'm a bit concerned about at least one measurement:

Read 16-byte file 100000 iterations 2.1s for 100k reads 21219.3ns per read

This appears to be testing... the filesystem cache? It's certainly not seeking on a rotational disk, "SSD" isn't on the page, so... not sure where he's getting 21 microsecond reads from.

It's measuring the overhead of a set of open()/read()/close() syscalls - exactly the situation where you would want to use the cache to avoid physical disk latencies swamping the language variation.

Then it should say so, not suggest any rereading is being done and that it's just kernel plus memcpy.