I don't think that the cathedral vs bazaar part is interesting (even if I cannot help thinking so Plan9 never got a good desktop, so much for cathedral development) some of his concrete points are interesting though, some not so much:
1) lack of portability to FreeBSD:
==> why aren't those patches upstream?
2) the software in the FreeBSD ports collection contains at least 1,342 copied and pasted cryptographic algorithms.
==> if true, given the security aspect, that is a big issue!
3) libtiff is used by one library even though Firefox cannot render TIFF images
==> someone made a mistake in the dependencies (most probably) or there is a bug/dead code in FF, a very minor issue.
4) one package needs both Perl and Python
==> an even more minor issue
5) you will find that you need three different versions of the make program, a macroprocessor, an assembler, and many other interesting packages.
==> code reuse and dev freedom have drawbacks..
6) libtool, which tries to hide the fact that there is no standardized way to build a shared library in Unix. Instead of standardizing how to do that across all Unixen—something that would take just a single flag to the ld(1) command
==> That is pure BS, what about existing systems??
6) he hates autoconf
==> the implementation is bad, I think that everybody agree here, but some here prefer autotools to the other tools (for example cmake) so what this means is that we still doesn't have "very good" build tools, a sad situation indeed.
So (2) is a big issue, (1) is an issue too, (6) is bad too..