This is just the first release and I'll continue to improve the whole thing, but most RSS feeds are a mess. Different charsets, all sorts of tags in different feeds, invalid XHTML etc...
Tidy helped me with most of the HTML problems that those feeds have but charset is still an issue, sometimes I cannot get the charset directly and need to detect it. Sometimes iconv fails to transform from one charset to another etc... That's why you get this message on some feeds. This problem is not random at all, it depends which charset is used for the feed. For the Wall Street Journal, I guess that they must be using different charset encoding then...
I haven't experienced this problem with the feeds that I'm using (for example:
http://www.feedbooks.com/newspaper/114 ) but I know that it exists. In the upcoming weeks I'll improve the part of the website dedicated to RSS/newspapers with new pages to browse, search and display your content. Along with these improvements I'll see if I can get rid of some of these charset errors through better charset detection.