>> Oh, and string theory has made predictions that are testable. <<
Actually string theory predicts undetectable dimensions that when they are tested for with new machinery, are simply reworked to be smaller than the new detection methods can detect. The theory always runs just ahead of the experiment to prevent falsification.
I’m not saying that string theory doesn’t have its problems, but there are some predictions that project down to energies accessible now and at LHC. Even if string theory gets it wrong, I expect it will stick around though.