Not saying all the schools are good. (some look slightly sketchy to me)
"Why would a company hire you, who just finished a boot camp, as opposed to someone who have years of coding experience and knowledge?"
An odd premise, its like asking why a company would hire someone straight out of college.
Companies hire from devcamps because the good ones are taught by people who work in the industry and know what they would want a colleague to know, furthermore they ask the potential employers what they are looking for and teach that in a hands-on environment.
I am still not convinced that this is anything more than a scam. Unless you are exceptionally brilliant at learning new language, no matter how much you push yourself, you will be barely good enough to even know the rudimentary concept of programming in 2-3 months.
To me the best way to learning programming is to make fuckton worth of crude programs and then make insanely silly and sometimes good mistakes in the process and then try to figure out how to fix those mistakes by yourself (or at least with some help).
2-3 months is barely enough time to understand _WELL_ rudimentary concepts of programming let alone learn from your mistakes.
If you already have prior programming knowledge and know at least one programming language, joining this camp is silly and it doesn't prove anything one way or another. If you already know one programming language, learning a new language should be fairly easier than for someone who have never programmed in his life.
I have to disagree here. If you taught yourself enough for what you needed at the time and never knew any other programmers then you face a huge impediment to learning. Every time you hit a wall where you know how to describe the problem to a human but not to google or stack overflow you can spend hours (notice I didn't say waste) hunting down what is the proper way to describe the problem. While this can be extremely educational in the broad sense it can be slow in terms of actual progress towards being a better programmer since the overhead associated with troubleshooting/debugging is multiplied when it is most costly.
Also, if you're spending 12+ hours a day writing code on average (in my program we did) for 12 hours a week, then you're putting in a hair over a thousand hours of coding time. Assuming a 12 week semester, 2 semesters a year, and 4 years this comes out to about 10 hours per week of actual coding. This clearly isn't the best student in a 4 year program's work schedule, but it's not unrealistic for an undergrad. It takes time to get situated with a given problem and into writing code, but if you're essentially on a 12 week coding binge where you take breaks to sleep and eat then that context switching overhead is reduced. The productive hours of a 12 week bootcamp may seriously outpace the hypothetical student who spends that raw time over 4 years.
As someone in this group, I'd have to disagree with you.
I work in a life sci research setting, and am usually the only person with any programming knowledge or experience whatsoever. In research, "it works," is generally the only quality control placed on bespoke software. I'm capable of writing working programs to perform simple tasks & do on a regular basis, but I could not begin to write a large web application using best industry practices. I've also never spoken at length with other programmers regarding the problems I'm facing, and as such, I would have a significant degree of difficulty discussing conceptual frameworks with a team.
Having spoken to a few people who attended these schools, it seems that they have a thing or two to teach someone like me.
Perhaps it's not worth $10k, but I think that's a separate argument.
But if these camps are letting just about anyone in I would have a problem with that. You have to have a solid foundation in the basic of CS at least.
I'd love to find out who's hiring these people, and how satisfied they are with the results. The two conditions where I can imagine hiring somebody with ~3 months total coding experience are A) when it's mainly a filter for ability to learn, and I expect to invest a lot of time getting them to the level of a production coder and then supervising their work, or B) a situation where output quality doesn't really matter.
I'm hoping all those people end up in situations like A. But there are definitely things that fit B. Back in Bubble 1.0 there were people selling themselves as an "HTML programmer". And there were consulting shops billing total idiots  out at $1500/day plus expenses. Your developers don't have to be very good if your sales guys are great and you have at least one solid technical person to put out fires and impress clients.
 this kind of idiot: http://www.codinghorror.com/blog/2010/02/the-nonprogramming-...