[intellegacy]

I've been looking into these schools as of late (including the one in SF - forgot the name)

And I've been trying to calculate how good of an investment these programs are. They help with job placement, which is valuable. Then again, the market is so hot for developers now that anyone who self-studies for 6 months could find a job fairly easily on his own.

As for the instruction itself, there are so many resources online (free, I might add) that it seems almost wasteful to drop 20k to learn when you could learn for free, assuming one possesses sufficient self-motivation and knows where to look.

Finally, there is the possibility of a potential huge opportunity cost one can't overlook:

the option of finding a paying jr developer job where one learns on-site. Instead of -20k you get 30k (based on 60k/year working for 1/2 year) and I assume end up roughly even in coding ability. That's $50,000 extra in your pocket, with demonstrable real-world experience rather than "school" experience.

[dbecker]

All of this just begs the question of how good these guys are.

Yes, you can find a job on your own. But not all jobs are the same. Maybe they'll find you a better job.

Yes, you can learn on your own. But I don't think there's any question that a good teacher can help you learn faster.

gSchool is new. So it's too early to find hard evidence either way about the quality of jobs you'll get, etc. But if they give you the right teacher, education can be a good investment.

[steveklabnik]

gSchool is run by Jumpstart Lab (my employer), who also did Hungry Academy for LivingSocial a while back. I wasn't directly involved in teaching HA, but I did do evaluations of the students from time to time, and they were all competent developers.

What they weren't was super quick to resolve errors. That can only really come from experience. They'd be able to figure out if something went wrong, but whereas I (#50 on the Rails contributors list) would know instantly what was wrong, it'd take them some time. That's to be expected from new devs, though, no way around that but time.

[thisiswei]

steve,will you be teaching this time?

[steveklabnik]

We're figuring it out. At the very least, I will be swinging by from time to time to teach periodically, but I will probably not be there full-time.

[lorangb]

The guys at gSchool believe strongly that truly learning everything online is really hard. The instructors are world class - IMO that will make all the difference.

[intellegacy]

I've never heard of gSchool. Maybe I'm out of the loop, but what evidence is there that the instructors are world class?

Also, there are people who have learned everything online. They are out there - thousands- heck- tens of thousands of them on the web. Granted, maybe gSchool will help educate some people who can't learn themselves. To those people the 20k will be worth it.

[jthree]

Nothing will speak more clearly than the results, so I can understand some skepticism.

I'm the lead instructor for gSchool, Jeff Casimir. I've been teaching Ruby since 2005 and started Jumpstart Lab in 2009. I have both more classroom experience (total) and more hours spent teaching these technologies than anyone in the world.

I wish there were dozens or hundreds of people learning these skills by themselves on the web, but it's just not happening fast enough. The talent shortage is the #1 problem facing most small software businesses, and there's no fix in sight. This is just our little contribution.

[thisiswei]

agree!!!

[steveklabnik]

gSchool is run by Jumpstart Lab (my employer), and we did Hungry Academy for LivingSocial a few months ago. Our client list includes Nasa, Boeing, Sony, BlueBox, LivingSocial, Accenture, and a host of other companies.

I personally am #50 on the Rails Contributors list all-time, so I know a thing or two about Rails. ;).

> To those people the 20k will be worth it.

While there are many people who are able to learn online, there are many that are not. Also, our schedule is just a tad more intense than most autodidacts go for. And the money isn't worth it just for money: it's for cred. I once had a funded startup, and the fact that we went through a YC-like program opened many doors, simply due to social proof. Those who teach themselves have to gain this proof some other way.

[jfarmer]

Caveat lector: I help run Dev Bootcamp (http://devbootcamp.com), which is probably "the one in SF" whose name you forgot. :)

I have two thoughts.

First, there's more to learning how to be a software developer than downloading a bunch of information into your head. Most software projects fail for human reasons, not technical reasons. Giving feedback effectively, receiving feedback non-defensively, articulating your ideas to other people of varying backgrounds and skill, convincing teammates to follow your plan, inspiring your co-workers with your clarity of thought and action, etc. are all real, hard-to-learn skills that are critical to being a great software developer which. One would also be hard-pressed to learn on them on their own since each is, by definition, an interpersonal skill.

As someone who has interviewed and hired dozens of engineers, my experience is that "junior developers," as deep as their knowledge of data structures and algorithms might be, are lacking in exactly this dimension. It's the #1 reason people are afraid to hire them and dismiss as "too junior."

As a case study, Hipmunk (http://hipmunk.com), a Python shop, hired a summer student from Dev Bootcamp. He was a great developer, but they obviously didn't hire him for his mad Python skills since DBC teaches Ruby and Rails. He had no significant prior programming experience when he entered DBC.

I know, I know: anecdotal, etc. etc.

Second, when developers talk about self-learning, I think there's a huge selection bias. The way people are "meant" to learn computer science, software engineering, etc. selects for autodidacts. There are so many more dimensions to how people learn effectively.

This is coming from someone who taught himself C using K&R in high school.

The two things you said are categorically false for many people: they just don't possess the internal mental models necessary to learn how to program on their own, and don't know how to find answers. They might not even know what they're seeing is an answer, unless they've formed the question correctly.

When friends ask me what programming language they should learn, I tell them they should learn whatever programming language more of their friends know.

Social support, similarly motivated peers, role models, and even just the knowledge that they have access to expertise that can get them unstuck when they're tired and frustrated are huge factors in getting people to learn. They will be more fearless, more intrepid, and work their way up to a place where they have the confidence and skill to both learn on their own and on the job.

If you're curious about the skill level of some of the students graduating from programs like gSchool and Dev Bootcamp, here are some examples from the fall:

* http://0xfffc.tumblr.com/post/35751220092/6502-assembler-and...

* http://www.grocery48.com/

* https://gist.github.com/aafabcfadafb4563fd3e

* http://flavorite.tumblr.com/post/35185577034/behind-the-scen...

Shameless plug

Dev Bootcamp's fall cohort graduates next week, and December 7th is our hiring day. If you're interested in attending, sign up here: http://bit.ly/hn_dbc

[xiaoma]

>"Social support, similarly motivated peers, role models, and even just the knowledge that they have access to expertise that can get them unstuck when they're tired and frustrated are huge factors in getting people to learn. They will be more fearless, more intrepid, and work their way up to a place where they have the confidence and skill to both learn on their own and on the job."

Yes! That's what makes this kind of thing worthwhile!