Frequencist methods when they are used to make predictions and get useful information out of experiments have hidden implicit priors that bias inferences in opaque ways.

The very honest frequencists will admit that their procedures are only rejecting hypotheses so small as to have no practical utility. Others will use weird and dishonest doublespeak where they call rejecting an insignificantly small hypothesis "statistical significance".

But I suppose they do redeem themselves a bit with the wording "null hypothesis" which candidly conveys the sense of having significantly rejected _nothing_.