I have to wonder what sort of code your average bootcamp graduate is really capable of producing. I'd say that you could get really, really good at a single concrete stack and toolset in that time (e.g. osx, bash, Linode, postgres, rails, chrome, sublime, chrome dev tools, maybe nginx, wireshark, redis).
But as soon as you swap out even one item, the pain will start.
Now do it with Windows, power shell, Azure, mysql, php, firefox, notepad++, firebug, IIS, mongo. Now do it with Linux, zsh, Digital Ocean Meteor, Safari, vi, django, apache. Now do it with Linux, bash, AWS, Java, Spring, Oracle, Eclipse, Glassfish. Now do it with RedHat in Docker on AWS running Postgres, Redis, Nginx, Java and Node, over WebSockets over any browser.
Now you will know what you are doing. It will take about 10 years.
Of course, within ten years half of the listed products will be gone, and replaced with the new hotness. But there are definitely a lot of low-level and high-level concepts you pick up along the way, and what I tell everyone who is learning is that the MORE stuff you try, the better you will be at learning MORE stuff. Every piece you learn about, you'll find another piece later that mirrors, borrows concepts, or shares ancestry, and you'll pick it up twice as fast as the last time.
I'd go even further, that if you learn how to do it with lots of different tools in different combinations, then you get a feel for what is important and what is not, what is fashion and what is important, and what software applications are, in general. Most of all, you can begin to make the critical distinction between "application" and "infrastructure".
But this is a very difficult path, one made slippery with tears.
You think people with cs degrees deal with stack swaps better than boot campers? I went to a top 10 cs program and I don't think so. Both populations push out equally ineffectual grads, only passionate programmers can really be fluid, something that no program can instill
Partly quoting myself from another wall-of-text discussion forum where code academies were discussed.[1]
Some developers need an introduction and/or mentoring. However, once past the absolute fundamentals (e.g. basic control structures and logic), mentoring becomes more strategic (less about technical details), and improving as a coder is a matter of how adroitly one can understand more complex patterns and absorb novel formalisms (frameworks, languages, hardware, integrations, etc.)
Growing beyond the level of individual contributor is working with individuals and teams to coordinate the development of more complex software systems. These human/machine systems are very complex and highly dynamic, to the point that there are no accurate models to describe how they work.
EVERYTHING is custom and cutting edge. (I'm thinking here of some of what Paul Ford said in his "What is Code?"[2], specifically section 5.6 "Off the Shelf")
This means that the career paths of developers cannot be predicted by simply taking into account educational backgrounds. Models that try to predict developer careers paths from educational backgrounds are not even close to being wrong because they model the reality of software development so poorly.
There will be lots of room for grunt coders who start off in code academies, sure. But some of those coders will start entry level and, after much more learning, go on to help develop world-altering highly-regarded software. Some coders trained at the most elite institutions will never contribute even one code block to a useful project of any kind. (I'm exaggerating, but you get my point.)
TL;DR: Code academies are simply another means by which humans will become part of the software industry and the status and stature of developers so initiated will be manifold.
> Code academies are simply another means by which humans will become part of the software industry and the status and stature of developers so initiated will be manifold.
Or to put it another way: dev quality is independent of education. I agree with that.
The point I was making was more about what it actually takes to become a really good programmer. Against the 10 or 20 years of experience it takes, both college and a boot camp look roughly the same: approximately irrelevant (although studying classic computer science, operating systems, algorithms and data-structures, are all quite relevant)
But as soon as you swap out even one item, the pain will start.
Now do it with Windows, power shell, Azure, mysql, php, firefox, notepad++, firebug, IIS, mongo. Now do it with Linux, zsh, Digital Ocean Meteor, Safari, vi, django, apache. Now do it with Linux, bash, AWS, Java, Spring, Oracle, Eclipse, Glassfish. Now do it with RedHat in Docker on AWS running Postgres, Redis, Nginx, Java and Node, over WebSockets over any browser.
Now you will know what you are doing. It will take about 10 years.