This has been noted starting (I think?) in the early 2000's-- I recall some people doing comparisons of "fresh-from-the-box" elements, stacking up 10 plates of older LEGO and newer LEGO. The stacks from older moldings lined up very well, the stacks from newer elements didn't line up as well. I think I noticed this personally for the first time in 2006 or 2007, by doing the 'finger test' described above-- build a flat wall, and run your finger along it, noticing the brick seams.
Certainly more recently, LEGO has consciously made cutbacks to quality in order to survive financially. Prior to Jørgen Vig's promotion to CEO, there was a huge focus on quality, and a very small focus on monetary result. As Jørgen reported, people used quality "as a crutch". "You can't do X because of quality" was a phrase he constantly heard when trying to restructure the supply chain, and it was a mindset he supposedly had to get people out of, because the relatively small improvements in quality that were being provided were costing a LOT of money when all tallied up. Hence, these small bits of above-and-beyond quality that we saw from LEGO in years past have been increasingly eliminated.
To give one example as it was relayed to me-- when LEGO was looking to make their micromotor, they scoured motor manufacturers for the best small-scale motor they could find. Not just "best in the toy industry", either-- it could literally compete with professional robotics motors and other industrial use motors. As a result, the motors cost an exorbitantly high amount to create.... But they were really awesome!
LEGO did things like that all the time, supposedly; in line with the old motto "only the best is good enough". But now, that's out the window. Now it's about maintaining profitability without sacrificing too much quality.