In this case, their CAPTCHA page intentionally included code to DoS a particular blog, sending a request to search for a random string every 300ms (search is very CPU-intensive). This was regardless of the archived site you were trying to view.