alt Hacker NewsThat's arguably what String Theory is good for, producing interesting, entertaining, and possibly even useful math. What it seems to fail at is making realistically testable predictions about nature that can't be matched or exceeded by simpler competing theories.
No Theory of Everything is going to make realistically testable predictions. That's a problem of the domain, not the theory. The unification energy between the graviton and quantum field theory is on the order of 10^19 GeV, over a dozen orders of magnitude beyond anything we can generate.
We might get lucky that some ToE would generate low-energy predictions different from GR and QFT, but there's no reason to think that it must.
It's not like there's some great low-energy predictions that we're just ignoring. The difficulty of a beyond-Standard-Model theory is inherent to the domain of the question, and that's going to plague any alternative to String Theory just as much.