Um... I disagree. I worked for a large company that does just that: hire liberal arts majors with high gpas and teach them to program. It works, well, horribly. Ask any former employee about the code quality at epic systems.
So honestly, I don't think your approach is too bad.
You mean in the sense that both approaches work horribly?
Your way gets you 10 programmers that do a really good job. Now find the other 69,990 programmers. (I like to think I get a lot of work done, but I can't get 7,000 people's work done.)
(If it was my company, would I have a bunch of unrelated development departments with no common standards on 5 continents? No. But it's not my company, so I don't get to decide, I can only observe.)
All I am saying is that using open source as a way to find talented individuals who love what they do and are really good is one solid way I have found to find new talent. If that can only get me X number of devs a year than that is X number I don't have to hunt through resumes for.
I agree with the blogpost that colleges aren't doing a good job of transitioning students to the challenges of startup/modern tech workforce. But, it seems to me that the solution would be a number of big employers like Google and others who are more able to take the risks should hire up the most qualified and teach them those skills. Simply, if the undergraduate education isn't good enough, someone's gotta teach em.
From what I've seen, you've got a lot of developers commanding 60-80k salaries. From a startup perspective, I don't really understand why this is the only acceptable path.
It just seems smarter to locate near some technical colleges in cities that have a lower cost of living, and pay entry level developers a salary between 35-45k to start. As they improve their skills, you up their salary... assuming that improvement comes with an increase in revenues/profits.
At that level, they hire for aptitude, not experience. The $60-80K that Google pays for beginning hires is on the anticipation that after training, they'll be worth at least that much, and Google can retain them long enough for them to make it back. It's fairly common knowledge that a new Google engineer won't be productive for 6-8 months at least.
The risk that a startup runs by paying $35-45K is that they'll get people with lower aptitude who'll never turn into experienced engineers. That's a pretty big financial risk to take, particularly for a startup with much less runway than Google.
Actual developers cost much more than that.
You can start here: http://arstechnica.com/civis/viewtopic.php?f=25&t=223259...
or glassdoor
or virtually any former dev there
With all the OSS projects going on, I can’t believe anyone serious about development wouldn’t just use that for experience.