I think what may have happened is that developers and tools got much better: developers got much better at designing and testing robust dynamic code (through working with server-side languages like Groovy and Ruby) and IDEs (I'm talking about Eclipse, NetBeans and IntelliJ) got much better at working with and debugging dynamic code. Overall this means that typos are less than they used to be and are much less likely to make it into customer hands.

As someone who visits companies to tell them how to fix their software before releasing it, I can tell you categorically that this is not true.

Companies that produce code in dynamic languages like PHP, Ruby, Python compared with languages like C#, Java or even C++ tend to have lower quality formalized testing and lower quality overall software.

Even worse is the fact that languages where the distinction between code and data becomes blurry (such as whoever thought you should upload files and put source code to be executed both on the filesystem should be shot - as should the person who thought SQL ever being a string that you can glue together at runtime, or the guy who thought HTML should be uncompressed, unencrypted, human readable and have badly defined parsing rules) are vastly more likely to have critical security bugs.

I hate to say it, but when I review sites by equally crappy developers written in C#/.NET versus PHP, the ones written in dynamic languages tend to fall faster and harder with less effort on my part.