The point of the analogy is that implementation -- essentially what CS is all about -- is not interesting just as the way you construct a pen or use it to write grammatical sentences is not interesting.
Computer Science is not about how to build a computer.
Computer Science is about what you can do with a computer. Google's pagerank is computer science. Recommendation systems, like Amazon and Netflix use, are computer science. Computer science is about discovering new ways to design bridges -- not about building them.
Implementation here refers to software as well.
Computer Science is about what you can do with a computer.
Not really. The vast majority of CS research papers are not about new uses of a computer.
Let me guess: you're not a big fan of Knuth either?
But clearly, we need to go beyond this and think about apps too.