That’s a good example for showing that despite changes that would produce major differences in sound, by sight the basic meaning of the sentences is fairly easy to determine. I suspect that the originator chose those particular words to make doing that as easy as possible, though. I was curious to see if other scrambled words would be as easy to read, so I tried doing that to the first long sentence in the first post of this thread (using the same technique — keeping the first and last letters in place). The result seems a good bit harder (to me anyway, and I did the rearranging).
“The sec-load ertexeps who erenineged tihs denlice hvae swohn a dinomec cenvserels in atkacintg the cmoomn snese of pihoncs, wlihe pliing up shotpisires taht spopudsely jifsuty the haox of Wlhoe Wrod or Shigt Wdros.”
Note that I intentionally tried to change the interior shapes of the words, though, and it’s the shapes along with the context that help us when reading by sight. If the letters were rearranged at random, the words would be easier than in my example. The basic point still holds that fast readers can recognize the meaning of many words just by glancing at them and without confirming that all the letters are in the proper order. (The fact that we often don’t spot typos, even those that would change the pronunciation in a conspicuous way, is a confirmation of this.)
It does seem to get more difficult as the average length of the words increases. I did a quick word count of your example sentence, looking for short words. Short words being words with 1, 2, or 3 letters only so that they cannot be rearranged. Your sample sentence was 29% short words. The original piece I posted was 46% short word, which was probably a big part of what made it easier to read. The short words provide the grammatical context and make it easier to guess the meaning of the whole.