I think he means the Bell Labs people were wrong, i.e Robert Morris did not enjoy math, for him it was about the programming. He later mentions that it was tough keeping in touch with the people at bell labs after ignoring their advice about math.
My reading: Bell Labs folks would have ordered it Programming > Math >> Computer Science. RTM was good at programming, was being taught CS at college, and took too long to learn that neither that nor math were for him and that Bell Labs had it right all along.
If I understand you correctly, I think you have it backwards -- as I read it, Bell Labs saw math as having "lasting value," while computer science was just a practical application. Morris didn't want to look foolish, so he persisted with math long after he realized that he really wanted to be a programmer; unfortunately, after he gave up on mathematics as a vocation, he didn't have the fortitude to keep in touch with those who advised him otherwise.
I'm fairly certain it's this, given the history of CS and programming. It started as a branch of mathematics. As an academic field, even in the 1980s/1990s and in some places today, it was/is taught as a branch of math. And comparatively, programming/operating computers was not high prestige and was seen more as grunt work.
Interestingly, in some ways the people in Bell Labs were right: the mathematics is what objectively /lasts/. Consider: how much mathematics from the 1950s is still in use and relevant, versus how much written code is still in use and relevant? Mathematics (and algorithms, etc.) have a permanency to them that simply written programs don't.
Don't understand this part, what's the obvious thing about math he was late in admitting? Were the Bell labs people right in his opinion or not?