Right, the comment surely should have said "substantively significant" and not statistically significant, since statistical significance is a product of not just effect size but also sample size and estimator efficiency (and, indeed, chosen alpha -- anything can be significant if you make the alpha large enough)