I think what made it change was a combination of things:
Seeing the damage caused to students (I work at a secondary school) by policies of selecting a G&T cohort, and 'labelling' students as G&T. The damage is caused to students who are labelled as G&T and students who aren't labelled G&T.
I've done a lot of reading about the subject, and the approach of Barry Hymer, Jack Whitehead, Marie Huxtable and others, is the one that rings true to me in my private/personal/family experience and in my professional experience.
Much of the work I do is with underachieving students and 'low ability' students. Yet when I read BH's writing on G&T, and that of others, I learn much there that affects how I teach these students.
I think Maths is a particularly problematic area when it comes to G&T. I don't personally see advantages in pursuing programmes of significant acceleration, certainly before secondary level. But we've been fortunate that ds1 (and ds2) to an extent, whilst exceptionally able at Maths, have not been particularly passionate about Maths, in the way they are with some other areas (reading, writing, science, history) and so they have been able to pursue other things rather than just whizzing ahead in Maths.
I agree completely btw that some students learn at vastly different speeds, even within a wholly 'normal' range. But the things that determine effective learning are skills that can be taught and/or stifled and restricted, or simply not encouraged and allowed to wither and die: resilience, enquiring, making connections, team work, independence, creativity, motivation, organisation, etc.
Treating 'ability' as a fixed item rather than a variable would make a mockery of much that we attempt to do in education.