Have we evolved to be bad proofreaders?

I came across this fascinating article. It suggests a reason why we tend to see patterns that aren't really there: for example, faces in rock formations, etc. Essentially, it seems to be that seeing something that isn't there tends to have less serious consequences than not seeing something that is there. So running up the nearest tree when you see a log that looks like a crocodile confers an evolutionary advantage because, very occasionally, the log will turn out to be a crocodile...

What's that got to do with proofreading? Well, it made me wonder if the same mechanism is at play when we misread words. Could this be why we can stlil raed wrods eevn wehn teh lettres are in the worng order? (See also one of the entries to our recent writing competition, which was based on this idea: http://www.edittrain.co.uk/perfect_start_kersti_h.php).

If so, this tendency to make patterns out of apparently random features means that proofreaders are really fighting against the tide.

But it also made me wonder: if proofreaders do have to suppress this instinct, are they also less likely to fall for other optical illusions? For example, are proofreaders less likely to see faces in the shadows of Martian landscapes?