The WYSIWYS bias is based on several things.
First, it's a snobby bit of coder's superiority, sort of like "I understand the HTML code and you just cheat with a program that does everything for you." It's just another way of saying "I know something you don't know." This is very prevalent among techie/coder types... "Even though you have a tool that lets you do some of the same thing I can do, you don't understand what's underlying that tool so I'm better than you."
Second, I would say a very strong argument could be made that WYSIWYG tools create inferior, clumsy, often bloated HTML code. A web designer who doesn't understand HTML might come up with a block of code that looks just the same as another guy's code, but it might be much, much longer than it needs to be, might not be truly standards-compliant (though it might display OK in most browsers), might have problems that aren't immediately obvious.
For example, I was looking into the HTML source of the web site of one of the labels whose CDs we sell on the Hypnos store. The label owner, whose web site looks perfectly good in the web browser, encouraged me to "grab" the URL of his hosted MP3 audio clips, and link to them in his label's CD listing on the Hypnos store. In other words, I was just going to go to his site, hit "view source," and find the URLs for the MP3 files so could point to them, and not have to create my own MP3 sample clips for those CDs. Nothing strange there... but when I looked into the HTML code for this label's web site, I found the kind of bizarre HTML artifacts that can only be created by a WYSIWYG web design program.
Where my own HTML code might have said:
<font color="yellow">Artist name - Album title</font>
.... this other label's HTML code for the equivalent single line of text display, might have read something like this:
<font></font><font color="yellow"></font><font color="yellow"></font>
<font color="yellow">Artist name - Album title</font><strong></strong>
This web site's HTML will display basically the same thing as my HTML but it's easy to see the difference. Multiply this kind of sloppy, flabby HTML across every line of text on a web site and it's easy to see how a WYSIWYG application can generate a 30kb file of HTML that could be coded by a person who understood HTML, using Windows Notepad or Mac Textedit, in a 1kb file.
In the example I give above, there's nothing really wrong with the second block of HTML other than its bloat, but there are many cases where a WYSIWYG app actually gives you text that's just plain wrong. I briefly took a look at Nvu and I won't make that mistake again.