This quote means that if you are used to design web-sites and produce web-content, you may be an expert in structuring and presenting information, but might not necessarily be aware of how to add information about the information.
This does not mean that you have to be a knowledge engineer to succeed. But it means that you have to learn additional languages, to document the meaning (semantic) of the information you put on the web.
The relation between the web and the semantic web, is somewhat similar to the relation between object oriented programming in a programming language, and modeling in UML the objects and their relation. You can be a perfect programmer (implementation), have a perfect understanding of the high-level design that you have implemented (abstraction), but if you do not know the UML language, you’ll have to learn it if you want to express the knowledge about your code.
This analogy is not perfect: programing language are very well formalized. There are tools that can reverse engineer code to deduce UML diagrams. You’d then only have to complete missing information (e.g the meaning of an association, or some characteristics that are not obvious in the code, etc…). In the semantic web, the content can be highly unstructured, so it’s much lore difficult to automate the identification of the concepts and their relation. And consequently, the addition of this (optional) information requires more effort.