That paper, which follows 106 people born in Denmark from 1975 and 1979, says there may be correlation coefficients of 0.48 or 0.50 for IQ groupings between ages (1) 4 and (2) 14 and 44 years. 0.48^2 is 23%, 0.5^2=25%, so if you accept everything else about this, this would mean 23%-25% of variance in the older years may be explained by IQ at 4... that strikes me as not enough to make your conclusions ("IQ is pretty stable from preschool onward) even if the rest of the study is concluded to be accurate.