String theory has a deep mathematical framework, and such mathematics, as well as the superior critical thinking and intellect required to learn string theory and thus signaled, are invaluable for all sorts of professions, some of which pay well: computer science, hedge fund/quant trading, philosophy, teaching physics and math, engineering, etc.This is the standard party line which faculty will feed you when trying to recruit you. What you eventually grow up to realize is:
the superior critical thinking and intellect required to learn string theory and thus signaled: What is actually signals is delusional arrogance and/or cluelessness, leading to 5+ years spent specializing on a topic perfectly well known to offer negligible career opportunities. A prospective employer looking at this argument will simply ask: if you are so intellectually superior, why are you not sticking with string theory instead of trying to mix with us lowbrows?
invaluable for all sorts of professions, some of which pay well: computer science: There is very little overlap between the mathematics used in theoretical physics and computer science, beyond what you will already have picked up as an undergrad.
hedge fund/quant trading: No. There was a time when physicists in general were considered good quant material. These days there is a lot more competition, and employers have wised up. Look at job offers for such positions, and you will see that those which still mention physics as an adequate educational background will now tend to say experimental physics.
The reason is, again, negligible overlap between the mathematics used by HEP theorists and experimenters. The latter, to the extent that they are not building hardware (in which case they are essentially electrical and/or mechanical engineers) are effectively statisticians doing big data work. If you spent years programming computer clusters to look for statistical anomalies in petabytes of LHC data, you can make a good case for having transferable skills useful in finance. If you spent years using Mathematica to explore differential geometry in 11 dimensions, not so much.
philosophy: A booming industry?
teaching physics and math: This is the real one. Of all the people I've known who went for a Ph.D. in theoretical physics, most ended up teaching math & physics in high school. This is rarely seen as a good outcome. They spent their youth subject to high stress and uncertainty in return for very little money, usually doing not what they may have dreamed of doing but what was considered likely to lead to publication within the time frame of a Ph.D. program and, of course, consistent with their advisor's interests. And for that, they were rewarded with a low-paying job requiring intense interpersonal interactions, pretty much the opposite of the ideal for the kind of personality which is attracted to highly theoretical endeavors. Those who are not miserable have learned to cope by focusing on other aspects of their life than work.
engineering, etc.: I can honestly say that I have never known a Ph.D. in theoretical physics who went on to become an engineer. Experimental physicists, sure, lots; they are really engineers with an unusual specialization. But not theorists. I guess you can find them if you look hard enough, but again: the overlap is negligible. If you've been doing string theory, the "switch" essentially amounts to throwing away what you learned in grad school and falling back on your undergraduate core curriculum.