The example he uses in the paper is from operating system design, where doing the "right" thing was too hard to implement. The "wrong" way is simpler to implement, and you get your OS written faster. Your OS now has a documented design flaw, and people have to work around it - but everyone does because you got your OS out first.
That is, he was more addressing design choices made when solving a problem. The full discussion is in his paper. It starts with the sentence "Two famous people, one from MIT and another from Berkeley (but working on Unix) once met to discuss operating system issues.": http://www.dreamsongs.com/WIB.html
http://www.computerworld.com.au/article/313989/-z_programmin...
By the way, Dan Weinreb has an interesting piece in which he reveals that he was the MIT guy and Bill Joy the Berkeley guy:
http://danweinreb.org/blog/the-worse-is-better-idea-and-the-...
I recommend Kay Schluehr's comment on that page, which argues that arguments like WiB are drawing overly general philosophical conclusions from instances of success or failure in the market.
Gabriel's full account of the essays he's written is on his website: http://www.dreamsongs.com/WorseIsBetter.html
I don't begrudge him for being unable to make up his mind whether he agrees with "worse is better". There are disadvantages to both approaches. But there's no need to be a drama queen about it ("One fellow was seriously nervous that I might have a mental disease") or inflate it into an artificially large "philosophical" question.
And since it's always a moving target, getting something that works "well enough" and then moving on to the next thing will always keep you ahead of people polishing their gleaming cube of perfection for ever more marginal returns.
Consider company BigCo, which starts a massive $10mil project that will revolutionize the industry. It is epected to finish in five years. A small startup builds a little product for $1000 in a year, which implements one of the core features of BigCos vision. This is clearly a "worse" product. Using the revenue of $0.1mil, they enhance their product. This yields $1mil revenue after two years. Two iterations later the implemented BigCos vision in four years with just $1000 starting capital. When BigCo releases their final product it is already obsolete. This is how worse is better.
In the (Free) software world it isn't about money, but about users and developers. Gabriel tells a Lisp vs C example in "Worse is better", where "1987" is equivalent to the end of the five year period in my example: In concrete terms, even though Lisp compilers in 1987 were about as good as C compilers, there are many more compiler experts who want to make C compilers better than want to make Lisp compilers better.