That's already happened with NG DNA assembly. PIPES cloning vs. Gibson Assembly vs. ColdFusion, etc.
In the case of restriction enzymes, though, they've been around in continuous use since something like the 70s, they're very well characterized, NEB has had a continuous research program where they've been optimized out the wazoo.
Technically speaking they are fundamentally simpler than CRISPR (one component, vs. 2). They also are more generally useful when your genetic manipulation is done outside the cell. So there's a clear tooling difference in CRISPR/RE. Most people who use CRISPR, will use REs in the process of make the DNA piece they're putting in alongside the CRISPR. You'd be a fool to use CRISPR in E. Coli or Yeast.
At the end of the day, they are just tools, and tools however technically shiny, don't by themselves deliver better end products.