Your question could be rephrased, "Why don't we use keyboards that are extraordinarily well designed for programming?" But that makes me ask, "For what task are current keyboards well designed?" The answer is none. Certainly not touch typing.
End users are oblivious to the issue as are many technical users. The small fraction who notice see it as a minor problem with no clean solutions. Combine them with the small fraction who prize standardization as the primary design concern and the set of willing voices on the topic amounts to a chorus of improvement-nay-sayers. Not only do they want the status quo for themselves, they would like you to stick with it too.
At the top, < 2% use alternate keyboard layouts. These are the people who see a problem and try to choose the best available solution even when it means lots of hard work. Ironically, when external forces banish some of these people back to the land of querty, they become the most ardent standardization evangelists.
But DVORAK offers no help for the left hand key staggering that goes against your hand. And COLEMAK is powerless to save you from the home key bumps that brute force your finger tips off center so that your little finger has one-column reaches on the left and three-column reaches on the right.
The same programmers who understand the rationale for normalization in a database and the DRY principle in software mostly see nothing wrong with a keyboard where every modifier key is duplicated.
And then there is the < 1% who actually go out and try to improve the situation. Kinesis Contoured, Data Hand, TypeMatrix, a dozen others. All of which are improvement in some regard, but all of which have other design problems that make them unworthy of unflinching advocacy.
Bottom line, the vast majority of us care nothing for keyboard innovation so if you want something better you'll have to make it yourself, but don't expect kudos.