Well the aim is to sanitize the code and move some of the unnecessary boot procedures to a later stage of start up. I don’t think boot times differences would be noticable.

It depends on what your bottleneck is. For example on my system I get

$ systemd-analyze                                                       
Startup finished in 11.976s (firmware) + 3.879s (loader) + 2.013s (kernel) + 157ms (initrd) + 6.354s (userspace) = 24.382s 
graphical.target reached after 6.316s in userspace.

The kernel boot process is only responsible for 2s of my boot time. So even if this does end up improving boot times, there’s very little it can do. The real improvement for me would be to choose a faster-booting m/b. You can run systemd-analyze on your setup to see if the kernel boot time is more significant for you.