> Apologies if this was long and off-topic, but I hope it's left some kind of thought somewhere.

Yeah, tl;dr - confirmation bias and sample of n=1

There are data points available for median wages of bricklayers and they are nowhere near the figures you described.