Rationale: if one build fails, the other one will still continue running.
Especially since we already have some (and will presumably add more)
proprietary software to nixda which may require manual intervention
using nix-prefetch-url, this seems like a good idea.
An alternative would be to just specifiy `--keep-going`, but this seems
likely to lead to unreadable output.
However, I'm not sure what exactly happens when we try to run builds for
both hosts at once on the same nix store — it may be that this isn't the
best idea, but I guess we can just make the pipelines sequential instead
if necessary.
