The thing is, this is actually about reCaptcha, and that is not the same as the ordinary Captcha such as Google uses for their own log-ins. reCaptcha uses two words, side by side. See [google.com...]
It's a project where one of the words is the real entrance test and the other is a scan that has previously stumped OCR software. The guesses on the "stump the OCR" version then make up a kind of crowd-sourcing for the OCR project, but entrance does not depend on success - the word is still an unknown.
reCaptcha characters are relatively easy to read, and the application also offers accessibility to blind users. In my opinion, it is worth working through the technical challenges to use.