Almost every general statement about how Tibetan is pronounced is a statement about the speech of Lhasa, and Lhasa is not Tibetan.
The Tibetic varieties spoken across the plateau and the Himalaya differ from one another at least as much as the Romance languages do, and they are not mutually intelligible across the range. The central varieties around Lhasa, which supply the standard, have lost the initial consonant clusters the spelling records and have become strongly tonal in compensation. The Amdo varieties of the northeast have gone the other way: they retain a great deal of the old cluster system and are correspondingly less tonal, so a word written with a prefix or a superscript may still be pronounced with the consonant the spelling shows. The Khams varieties of the east differ again, and Ladakhi and Balti in the far west retain archaic features of their own.
The written language is shared. All of these varieties are written with the same thirty letters and the same spelling, which is one reason the spelling has not been reformed to match any single pronunciation: it is the common property of speakers who could not otherwise read one another.
So the silent letters are not silent everywhere. What Lhasa speech has spent, other Tibetan speech is still using, and the alphabet is closer to some living varieties than to the one usually described.