The formulation given is a fairly weak one, and allows for the possibility that what has remained uniformly the same for some time need not remain so.

Something like uniformity of nature seems a necessary assumption to do empirical science, in as much as one cannot even test some proposed regularity without assuming other regularities.