First, at least some of the issues linked have been addressed. Second, they don't strike me as severe. So, all in all, the above take seems like an exaggeration from my POV.
To what degree do these affect the commenter above? Other people? I've not noticed them myself. Maybe there is more to the story?
> I recommend you don't subject yourself to software that is this level of unreliable and unbothered about being so unreliable.
Can the above comment say more about "this level of unreliable"?
Think of all the people using Nushell. Out of say N hours of usage of Nushell, what percentage correspond to a user feeling something like "Nushell is unreliable and/or buggy in a way that notably affects me"? Of course I don't have data, but I would guess it would be very small, probably 0.5% or less.
My feels would be that people who "leave" or "bounce off" Nushell mostly do so for other reasons, such as (a) not POSIX compatible; (b) it feels weird to some due to its design decisions (maybe pipelining for example); (c) it has a learning curve because it is different; (d) things change quickly; there are breaking changes. But I would be surprised if there was anything close to ~0.1% of people feeling like Nushell is anywhere in the ballpark of being even one of these three: (i) deeply-flawed; (b) bug-riddled; or (c) led by people unbothered by unreliability.