Nice life if you can get it ...
Me, I've spent the day correcting forty two pages of page proofs, and filling out immigration forms. Honestly, I'd prefer to be filling out tax forms!
Back in the day—that would be 1997—when I first encountered page proofs as a research assistant, some journals still seemed to be re-keying the text. As you might imagine, this led to somewhat more errors. Perversely it made correcting the page proofs easier. You knew you'd find at least an error a page (and people with more experience and years in the business can tell us about the times when one error a page was a good thing), so there was some reward for your time.
Nowadays with many presses sliding the text in from your word processing document, there are fewer errors at the final stage. But it's still the final stage. Last chance to save yourself from error. Who wants something out there under their name with "forth two" for "forty two." Not me. Coincidentally, there were as many pages in the article as there were articles that we reviewed. Now, hopefully, I just wait for the October issue to arrive ...