I've a question: how is Lambda School different from the other many other coding bootcamp programs (apart from the unique model of paying the fee only upon finding a job with adequate salary)?
Incentives are hard in this space. Not sure how for-profit companies who get paid by students can ever be exactly incentivized to focus on honest evaluation and strong fundamentals.
Maybe the FAANG companies should start/fund a bootcamp, but then you'd have to be selective, and you'd end up with something totally different... :/
But yes you’re correct.
This reminds me of some recent comments about the iPhone XS performance for the Speedometer 2.0 JavaScript benchmark [1]. The comment was something like "Apple just optimized their processor to beat benchmarks." In reality, the benchmark uses the TodoMVC project to run through dozens of real-world JavaScript frameworks. So even if Apple was just optimizing for some benchmarks, they've also improved the performance of real-world applications that use React and Ember.js.
Lambda School might be teaching some "buzzwordy" programming languages and frameworks, and the students might learn just enough to pass technical coding interviews and get a job, and they might have just enough experience to start contributing code and working on features that bring in more paying customers (or raise more VC funding), and they can ramp up quickly enough to not get fired. And those customers/VCs give money to the company, so the company can pay the employee, who gives some of the money to Lambda. This seems like a good "benchmark" to optimize for.
[1] https://twitter.com/dhh/status/1043277162676072449?lang=en
Lambda itself may be fine, I can't judge it specifically (and appreciate Austen engaging :)).
That said, I agree that if they come in with "enough experience to start contributing code [etc]", that's a good benchmark!
But for the school to be successful, they don't really _need_ to go that far, which is my point about incentives: They just need them to look good enough on paper to get in the door, and know just enough that someone will take a chance on some percent of them, and rely on the fact that firing is hard -- in this model, their 'incentive' is to just churn out as many grads as possible, with the nicest resumes possible.
After writing that, though, I think I'm being more cynical than needed, prob based on some sub-par bootcamp grad interview experiences. I hope Lambda is really committed to good education, and I wish them the best, it's a worthy ambition and a hard problem.
It seems that would be a very bad idea long-term.
Bootcams can focus on what’s “hot” right now without regard to job placement though the good ones do.
The reality is the alignment of incentives causes everything else to be different. There are a few concrete examples, but it really flows through everything we do.
1. Length: your average code school needs to pump you in and out of the same physical location and still have margins on what they charge upfront, so the standard is 12 weeks. That’s generally 8 weeks of “instruction” time and 4 weeks of “project” time. The fact that they can get anyone at all to semi-employable with 8 weeks of instruction is a miracle, but most employers agree it’s not enough, and frankly most bootcamp grads look pretty weak. Lambda School is 30 weeks full-time plus 4 weeks of required precourse work, so our instruction time is at least 3x that of most code schools. That, of course, lets us cheat relative to most other schools.
2, Curriculum. What you’ll usually hear separates us are that we teach CS fundamentals, write code in Python and C not just JS, and that kind of thing. But what is harder to explain is how much time and effort we put into instructional design. We’re entirely online and free upfront, so if we suck you close your laptop in week 3 and walk away, and we’re required by contract to forgive the entire income share agreement. We have some of the best instructional designers in the world on staff, and if an instructor isn’t performing they’re sadly let go. It’s that simple.
3. Mastery-based progression, and Bloom’s 2 sigma problem (https://en.m.wikipedia.org/wiki/Bloom%27s_2_sigma_problem). Perhaps the most remarkable study in the history of pedagogy was Bloom’s 2 Sigma Problem. I’m not the instructional designer so I’ll let you read about it yourselves, but basically we practice small group and one-on-one instruction with a mastery component. Specifically, every student is placed in a group of 8 students with one PM (what we call our TAs), and at the end of every week we have what we call a “sprint challenge.” That tests all of the cumulative knowledge you should have gained that week, and you must pass all of them to move on to the next level and graduate. So what if you don’t? You simply repeat the week in a new group until you’ve mastered the concept. Now every educator on planet earth knows that’s a superior way to learn, but it’s very expensive manage. We spend the money because we need to be able to confidently say, “Every Lambda grad can do all of these things” to every employer if we’re going to win long-term.
4. Very robust career services. Lambda School doesn’t stop once you graduate; in fact that’s probably the most important part to a student’s success. We have a program called “Lambda Next” that is still structured and rigorous, but helps you in writing code that signals what you now know, sourcing and applying for jobs, interviewing, and negotiating a salary/benefits package. We’ve had a remarkable number of students earn back the entire price they’ll pay Lambda in negotiation alone.
So these are just a few examples, but really it comes down to the fact that our DNA is structured differently than most code bootcamps as a result of our business model.