[WhyNotHugo]

It's really hard for new designs to thrive outside hobby/research usage.

Linux/BSD have massive hardware support, where a new OS needs to start from scratch. The same is true of software support: a new OS that truly has new primitives either adds a POSIX layer (and all POSIX-like software can't benefit from many of its improvements).

Even once you get past hardware and software support (both herculean tasks) performance needs years of hard work to be comparable to mainstream OSs.

And don't even get me started on the ever-growing complexity in standard protocols: HTTP2/3? QUIC? TLS? All "doable", but also huge sinks for time and effort when you already have a huge list of things to deal with.