I'm guessing many people's first response to this situation in real life would be to call out to coworkers - hey can anyone read this?, yet most of us would stick to the solo problem solving mechanics in our answer - search the # online, ocr + babelfish, call the number cold, whatever.

I understand fermi tests are intended to show broad problem solving abilities, but imho, these tests are for lazy people. Talk to a candidate for 20 minutes. If you can't tell if he/she is suited for the role, or at least for a follow up interview, one of you is not qualified.

Slightly pedantic, but this question was not a Fermi problem - Fermi problems are numerical approximation problems, and this one is more behavioral.

That said, I don't think they're lazy questions necessarily. However, I'm a physicist, enjoy Fermi problems (actually compulsively do the calculations when I hear a new one - a million golf balls!), and do think that order of magnitude approximations and quick unit changes are important skills to have developed.

I do think it's very lazy when they come out of a book or are riddles, because there are so many short approximations we can do at any moment. How much will the electric bill be for this office when it's August? or I just made a pot of coffee and forgot to place it back on the burner - when will it be too cold to enjoy? No accurate answer is needed or cared about, but the ability to predict important factors and come up with a list of contributing components is important.