I am wondering if the loss in popularity of the dipthong ("squished-together vowels" - a delightfully apt description!) is, at least in part, due to the relative difficulty in producing them on word processing software. They are fairly rare in English, anyway, so the temptation to reduce them to a single vowel is quite intense.
If this is so, it is rather a pity, as yet another example of how our culture is increasingly being defined by our technology.
A probable second reason - and one which was apparent before the advent of mass-PC/Mac ownership - is uncertainty among many people of how to pronounce words with dipthongs.