Because the market rates for Bob who just finished a boot camp and Cindy who has years of coding experience are very different numbers. (I also expect the market clearing price for Bob and Dave, who recently graduated Stanford CS but has no professional experience, are also different.)
Concrete example: the going rate for 3-5 years of Rails experience in the Valley is creeping above $140,000 right now. I'm willing to bet you can get many bootcampers to agree to work for $80k. If you're willing to put them on a team with experienced Rails programmers and mentor them a bit, that's a heck of a lot more bang for you buck for e.g. just wiring up actions and views. You probably won't have them rearchitect a payment processor's queuing systems by themselves over the weekend while live transactions are flying over the wire but, hey, that's not a requirement for most Rails work (by count or by weight).
Does this exploit Bob? Well, if we assume that prior to the coding camp Bob was a generic young liberal arts graduate, Bob's employment prospects in 2014 are a) not that great and b) likely offer a salary in the $30k a year range. $10k plus a few months of work plus opportunity cost strikes me as "More than possibly worth it" to secure an $80k a year job and the career trajectory that programmers have as opposed to the one that e.g. baristas or office managers have.
In theory a good bootcamp should accelerate your learning and give you a strong multiple over what you'd learn fighting your way through online resources. In practice, someone who does the latter is a much stronger signal to me as a hiring manager. In a way it reminds me of my experiences trying to work with outsourced teams in India 10 years ago where it was obvious that many of the programmers were there were simply clock-punchers encouraged to enter the field by their parents and saw programming more as a rote exercise of finding the right recipe rather than as a craft to be mastered through deep mental engagement.