Lambda School is SV big bet on reinventing education but students say is a cult | Hacker News Reader

Lambda School is SV big bet on reinventing education but students say is a cult

businessinsider.com

109 pointslhuser12365 comments

>> "then locks them into an intensive program that ultimately leaves them on the hook to pay back thousands of dollars in their future wages."

So like any other student loan? Lambda school is offering an alternative to help those who don't have funds or can't qualify otherwise. Nobody is forced into it but of course it's not free.

>> "when a white male student wore a Mexican sombrero to a presentation in front of the class."

What? That's not racist nor even a problem in the slightest. People who bring up these kind of perceived faults will usually find offense in anything they can and are better removed from the program for the sake of everyone else who's actually there to learn.

Complaints like these make it clear this is a targeted hit piece with anecdotes rather than a hard-hitting investigation.

Spot on. If anything, it makes me much more supportive of what is Lambda doing.

Shortage of good SWE will always be there, so I’m pretty happy it’s Lambda making money of this with actual results rather than leaving it expensive universities who don’t deliver.

> According to the Council on Integrity in Results Reporting, 60.9% of Lambda School students were employed 90 days after graduation, going up to 85.9% within 180 days. Graduates earn a median annual base salary of $60,000, according to that same study — although a Lambda School spokesperson puts that figure at $70,000, and notes that many graduates are in rural areas where average pay is lower.

This is one of the only paragraphs of the article with data rather than anecdotes. The article only focuses on students with negative experiences, but if this data is correct, it looks like Lambda School is objectively successful, and I bet only a small fraction of its students regret attending.

Unclear if these figures account for students that dropped out...I’m guessing no. Per the ISA terms you owe 40% of the ISA after 5 weeks of the 36 week course (<14% complete).

My issue is the incentive that would create to enroll students you in admissions know would likely not complete the program, and then still collect 40% of the value of the ISA.

It becomes a Survivor-esque learning environment where a significant number of students might be dropping out and you’re left with a much smaller number of students who do graduate.

Graduation rate would be a good figure to publish on their site. I searched for it out of curiosity, and according to the CEO they have ~85% graduation rate.

https://twitter.com/Austen/status/1140636581679624192

This figure is not audited, and Lambda school only recently switched to 9 months from 6 months, so saying that it’s 85% for a “9 month program” is misleading if most of their past cohorts were on a schedule much shorter than that.

That said, .15 + (.85 * .15) = 0.2775, the proportion of the student body who either didn't graduate or didn't find a job within 180 days of graduating. So there is a significant fraction who might regret attending the program. But it still compares well with universities.

Which isn't the right point of comparison. How does it compare with University CS departments?

Usually a big component to these programs are weekly exams to make sure you’re on track. I’ve looked into some of the local ones and usually if you fail a few you will be dismissed from the program.

85% means little if Lambda has a low bar to dismissing students that are struggling and not keeping up with the material.

At least they don't have to pay for it if they can't find a job.

> My issue is the incentive that would create to enroll students you in admissions know would likely not complete the program, and then still collect 40% of the value of the ISA.

They only collect money if the student is employed in a tech field, which means they're more incentivized to get you into one of those.

If someone dropped out after completing 14% of a program and later did another program or self-study to land a job without Lambda’s help why should Lambda get 40% of the ISA value?

If this scenario never happens to dropouts (you’re assuming they have no other path or desire to work in tech after dropping out) then Lambda would either change the percent owed to something proportional to time spent or increase the number of weeks you spend at Lambda before owing anything.

A simple answer could be opportunity costs. Lambda is oversubscribed and places are limited, so by accepting a place and then dropping out, Lambda is stuck with nearly all of the same costs as if you had completed the course, which they could easily have filled with someone else.

I don’t think they really have limited capacity. There’s minimal additional cost to adding a student participating in a massive Zoom call.

As someone currently seven weeks in to one of their classes, this isn't entirely accurate - the full class I'm in is about 50 students plus a dozen assorted staff. At that size, there's room for a fair amount of student-lecturer participation (comparable to 300+ level college classes), and I am at least familiar with most of my classmates. We also break off into 7- or 8-person groups in the afternoons, which are led by paid team leads (prior students), which would also somewhat limit capacity.

Thus far, I've seen that most of the people who dropped out did so in the first couple weeks - IIRC, there have been two people who left since week four, but I don't know if they dropped out or got "flexed" (pushed back to the class that started a month after mine).

That being said, I do agree that 40% after the first month is steep, though the limiting of the applicability of the ISA to tech jobs helps. I'll be interested to see how it plays out, and if more people drop out over time. I certainly hope it works out well for me, but I'll have to see how the rest of the program plays out.

The ISA only activates if you get a job in that field of study.

I think it would be tough to comment on any of this without having gone through the program myself, though I do wish there was an “undercover” senior software engineer who would sign up for the part time program and do maybe a month or more of the curriculum and write about it.

Knowing nothing else about the program the biggest red flag For me is that you owe 40% of the ISA after only 5 weeks of the 9 month program (e.g. less than 14% complete). That seems really excessive, and would seem like it would drive employees to “accept” students even if they knew there was minimal chance they’d be able to complete the entire program.

I don’t have any opinion about Lambda School, but I believe it’s normal for schools to charge tuition before classes are completed. I believe the drop deadline at my college (after which you can’t be refunded at all) was around halfway through the semester. I think a pro-rated tuition policy would be laudable but atypically consumer-friendly.

Don't the payments only kick in for a decent job? That can be bad if you already have a decent job, but maybe don't sign up if you are starting from there and aren't sure about it.

The ISA is highly restricted though.

Only kicks in if you're employed in a software job making more than 50k/yr.

ISA stops after repayment or ten years whichever happens first.

My favorite part of it is the ISA structure, it aligns the incentives between the school and the students. Universities don't have skin in the game, charge outrageous tuition fees and provide primarily signaling (even in the best cases). If Universities had to switch to an ISA model we'd see which ones were actually valuable - if they'd even be capable of adapting.

Doesn’t matter in my opinion. 40% after 5 weeks is a joke. If the student felt the need to drop out due to any reason so early and then ended up either completing another program or becoming self-taught and then finding a job it is silly that Lambda would still collect up to $12k from the student for very little value.

If Lambda were extremely confident in their program and who they select they would at least move it up to 25% of the curriculum complete, not 14%, or lower the 40% to something actually proportional to the time spent in the program.

Yeah - I think that's a fair criticism.

It sounds like you agree with the ISA concept but are only unhappy with the price. I'm guessing the terms are this way to discourage the less motivated.

If I drop out of college in a given quarter, I only owe tuition for that quarter. This is like dropping out my first quarter in college and owing my university 5 quarters of tuition.

No I don’t really like ISAs and personally would never do one myself.

If you drop out of a two-quarter college in the first quarter, you only owe the first quarter (50%) which is roughly equivalent

A four year university is 12 quarters. There is no such thing as a 2-quarter degree as far as I’m aware. Potentially this is more equivalent to a 2-year associates degree, which would be 6 quarters.

I mean Lambda School is the equivalent of a 2 quarter college program. Everyone understands how fractions work -- the push back is that people don't think the fraction analogy you're using makes sense here.

I teach a 10 week program and students who drop out 4 weeks in are offered an option to return to complete during the next offering but are on the hook for the full program costs. This does not mean I think someone who drops out of college after 1.5 years should have to pay for 4 years. It's not the same situation.

No it's not. Even the CEO has said the amount of hour is equivalent to a 2-year program.

Traditional college courses meet for a few hours a week. This is a 9-month all day program.

40% after 5 weeks is a joke.

This makes sense to me if their capacity to admit students for a particular class is fixed. If someone starts the class and later drops out, they can't be replaced. So once you have started the program, Lambda School is paying for your instructor for the whole class no matter what.

Don't you still have to pay for the whole semester if you drop out of a regular university after 5 weeks?

If I drop out of college in a given quarter, I only owe tuition for that quarter. This is like dropping out my first quarter in college and owing my university 5 quarters of tuition.

It’s an online course. There’s minimal cost to adding one person in a massive Zoom session.

9 months is 3 quarters, so it seems more like they are charging 1.2 quarters if you drop out after 5 weeks.

A four year university is 12 quarters. 40% of the total quarters is about 5 quarters.

I know one current student of Lambda. I kind of disagree with much of what is said in the article. The student and family are very grateful for the opportunity. My only concern is that, in their website they could be more specific as to the fact that you need to already have some of the skills. There’s just not enough time for beginners. To their credit, they give plenty of opportunities and alternatives for the ones that get behind. It just going to take more time than anticipated.

I'm confused by this article's complaint that the instruction was so poor that it requires students to "self-teach", and use materials outside of the coursework to learn.

In my experience, and sufficiently technical topic will require a ton of self-teaching. Should LS be expected to teach computer science in a way that prevents their students from ever having to use Stack Overflow? If that is the case, they are not preparing their students for solving problems in the real world, where there are no clear answers. (except the ones you find on Stack Overflow.)

FWIW I went to college long before Stack Overflow and it was definitely possible for them to teach CS without the Internet. Obviously teaching techniques should adapt but that's not an excuse for lazy teaching.

I am married to a current lambda program student. I also regularly interview and hire developers which I believe gives me a good sense of what the market wants.

Lambda is a good program with meaningful room for improvement. I’m excited for where they’ll be in a few more iterations and believe that they are worth it today.

> However, current and former students say that they've been disappointed that Lambda School's teaching staff isn't diverse, in turn: most of the school's instructors are male or are not from underrepresented groups.

Is this like all the eggs being angry there's no chickens yet? Where do you think qualified teachers are going to come from, if not the existing pool of professionals?

This reminds me a lot of Teach For America, a program I was a part of. It has very similar complaints, and very similar defenses. The truth is, there's a lot of truth to go around. Teach For America is not perfect. It's not for everyone. There's a lot of hype on the outside and Kool-Aid to go around on the inside. It inarguably has done harm to some people.

But for a lot of people, it works as designed. And I know a lot of people who did become great teachers.

My own story is somewhere in the middle [1]. It changed my life immeasurably, even if I never became a great teacher.

My guess is that the same broad spectrum of real stories and real outcomes is true of Lambda School.

[1] https://acjay.com/a-former-teachers-story/, if you care to read about it

"If you're keeping up, you either already had a foundation or you're self-teaching. The actual school is not effective at teaching. People are going outside to get what they need."

That sounds like a lot of technical education.

More than maybe any other discipline, classes have such a wide curve of expertise, they become almost impossible to teach to everyone at once.

On the need for outside study, I guess I've heard similar things about, of all things, tough music schools. You just end up in a race to see who can practice every waking hour.

Maybe there's an analogy there. We talk about computer science in terms of concepts and problem solving, but undersell how much muscle memory is involved. There is a ton of training your brain to reflexively handle certain common problems.

https://outline.com/fSVvSB

Reader view worked for me! (FF/iOS)

I read feedback about any bootcamp or any CS program with the background understanding that there must necessarily be loads of dissatisfied customers, people who can’t handle CS or software development. It is illuminating to poke through the article to the source material.

In my experience (went to a bootcamp 5 years ago, taught at a few since) bootcamps are amazing for people who have the drive and interest, and ideally some prior knowledge. They give you a place to meet like minded people, and really focus on learning/implementing your knowledge. The quality of the education is generally less important than the quality of the people. People who are just expecting to get a quick job and don't actually give a shit about the internet and programming should not be in bootcamps, but often are, and are probably the source of most negative feedback.

The data is consistent with most other bootcamps, though. There is a standard for reporting student outcomes here: https://cirr.org/data

Hack Reactor, now Galvanize, was one of the founding organizations of the CIRR, circa 2014: https://www.hackreactor.com/blog/tag/student+outcomes

I can't fathom why someone would go to a school that takes 17% of your paycheck for 2 years. I went to Western Governors University to get my undergrad, it costed 6k for 2 terms and I breezed through and got a paper that will help me for life, and I'm planning on heading to Georgia Tech for a Masters. There are better options out there, there is no need to go to a shady school such as this, even if they promise 0 tuition up-front. Additionally, as tech companies become more lax on their degree requirements, there is less and less value you gain from having a degree in Silicon Valley versus spending a few months grinding leetcode, which definitely pays off.

On a side note: it took literally 1 minute to hide the modal window and find the div containing the actual content and expand it, and turn the overflow back on. Did they design it this way on purpose? Otherwise, I think they have to find some new UI developers..

For those with lower income, even $6k can be too much of a risk. An ISA, in theory, mitigates some of the short term risk.

That 17% is only if you're making above a certain amount (I believe $50k+), and this makes a lot of sense for a lot of lower income people.

My brother is enrolled in Lambda School, and so far is enjoying it. They do move fairly fast and the curriculum is challenging but they are learning a lot of modern architectures and skills like React and NodeJs, skills that will actually help them get a job after graduating.

I went the traditional CS major in undergrad route and most of what I learned in school was not applicable to my day to day work.

Different things work for different people, nothing wrong with options. And Lambda does have the option of paying upfront and then not having to give them any of your paycheck.

> I went the traditional CS major in undergrad route and most of what I learned in school was not applicable to my day to day work.

I did a B. Eng in Software Engineering, and I found that the most important thing I learned at school wasn't any specific library or language. All of that stuff is extremely fluid and changes year to year. It was the skill of 'learning how to learn'.

Consequently, that's my main criticism of schools that teach to a certain set of 'skills' like React or Node. IMO what you should be learning are fundamentals that can apply anywhere. Data structures, algorithms, etc. Granted, I think learning those probably takes a bit longer.

Your college was teaching you skills too derivatives beyond mine. All of my classes were on CS and math. I didn't have a single class on "how to learn", or the very meta class "learning how to learn".

This is a common argument, but it seems like learning one computer language or library will make it easier to learn another one? Also, being familiar with some concrete examples should make it easier to learn general principles?

You're not going to get four years of work in six months, but I don't really see bootcamps as a hindrance to learning more later, either going back to school or on your own.

It took me almost 6 years to both teach myself enough (part-time) while working full time and moving my career in a direction that enabled me to get hired as a data scientist. The raise I got at my first DS job relative to what I was making was more 17% (and would have been a 100% raise on my first job). If Lambda School could have enabled me to get that job in one year, the payoff is pretty obviously worth it.

but specifically for DS i doubt lambda school or anyone else could land you a job in a year going from 0->100. from what i can tell the market is looking for either advanced degrees or many many years of experience.

I had a liberal arts BA and wrote my first line of _any_ code a year after graduating. When I got my first DS job, I only had 5 years of (non-DS analyst writing SQL) experience.

congrats! that's pretty impressive.

i'm just skeptical this could be achieved at scale.. data _analyst_ is a different story

Even getting a data analyst job in one year would have been worth it.

There are huge advantages to the way Lambda School is structured compared to other American universities.

I've written about this in more depth [1], but while ability is evenly distributed, opportunity is not - and aligning the incentives of the students with the institution is a much better model.

If your choice is risking 100k up front for a college education that doesn't guarantee a good job (and has non-defaultable loans) or paying 30k only if you end up with a software job making more than 50k a year it'll be an easy choice for most people.

[1] http://zalberico.com/essay/2019/04/08/lambda-school.html

What do you mean that ability is evenly distributed?

College education doesn’t cost $100k. In-state tuition at a public university in California is $15k. You can save money by going to a community college and transferring. With $30k you could do two years at community college and transfer UCLA to get your final degree and that doesn’t include financial aid.

There does appear to be a contingent of Lambda students that have had a poor experience. IIRC they were mostly related to a certain set of instructors.

I guess this is the issue of a popular startup - you get everyone talking about you, good and bad. Every startup has the problems listed in this article. It's pretty much the reality of startup life. But Lambda gets the microscope treatment.

In the end, though, it'll make them better for it. It's still interesting to note that this type of criticism will _never_ exist of any education system. Public dialogue here is good!

In fairness, the only reason I know about Lambda is because the founder, some employees, and some of their VCs are really active on Twitter and write incessantly about it in a “too good to be true” way.

Literally a few weeks ago their Chief of Staff wrote “if you don’t believe Lambda School is a $100B company then you don’t understand American History.” Really?

Honestly I don't understand the complaints.. For almost any technical career you will need to do some form of self study.. Also this program is aiming to compete with 4 year comp sci students. So yes you're going to need to grind to make up in <1 year what took someone 4.

However I do think they could do better with diversity.

Thinkful has been doing what Lambda is for the past decade. What makes Lambda different from the rest of bootcamps, absolutely noting.

Not to mention Austen Allred always came off like a snake oil salesmen.

Why does hacker news allow paywalled content?

Yet another PR story from Lambda School spun as news on HN.

Fake news.

The story is obviously negative toward Lambda School. With PR like that who needs hit pieces?

Reply on news.ycombinator.com