Lambda School’s Misleading Promises
nymag.com
nymag.com
Last year I decided to take action and started a free coding group at our local library: https://www.meetup.com/San-Jose-C0D3
I show up before work every day (M-F at 8am) to help students who are learning how to code. So far no students have gotten a job yet, but our group consistently gets 4-8 students who show up promptly at 8am. I answer their questions, give them guidance, and teach them best practices I follow as a software engineer with 10 years of work experience. I ask for nothing in return except the joy of students going "ahhh" when something clicked for them.
Things are still early for us, but my dream is to inspire other software engineers to help create a free and open learning center at their local libraries so people have an alternative to coding bootcamps.
Few questions if you do not mind:
1. How did you broadcast your Meetup to your students?
2. It looks like you have students pull lessons from a repo. Do you find it challenging to manage a handful of students who are all at different places in your lessons?
3. How do you avoid burnout?
4. Do you plan on open sourcing your lessons?
1. I simply created the meetup group. Showed up, people came. It was empty the first week couple of days, I just made it a point to show up. When there were no people, I simply just read hn or did some work. When the first student came, I made sure to be nice so he feels inspired to come back. The library took note, wanted to make it official with their library program and give us a dedicated room / parking even though we only have 4 consistent students. (Libraries love it when tech community volunteers to help teach).
2. I was honest with students that I can't help everyone and my time is limited (though I try). If they helped each other it would make my life easier. They have been good at helping each other.
3. I don't understand burnout, I'm guessing it's a symptom of ambition and expectations (I'm really not sure because I've never understood it. If I experienced it previously I never noticed). I just make sure to have no expectations. All I hold myself accountable for is to show up. If I did something wrong I apologize and move on, there's no point beating myself up over anything. I don't have big ambitions to make millions and millions, I just want physically show up for people who need help.
4. Yes. I want to open source it completely, such that if people wanted to repackage it into their own bootcamp they have the freedom to do so (not sure what license that would be, but I'm not at that point yet). Right now I just document everything in notion: https://www.notion.so/garagescript/Table-of-Contents-a83980f...
This is such a good mindset and, knowing myself and those around me, a rare one.
I am really into what you are building here - are you looking for volunteers to help?
Basically... if you are a seasoned engineer, just going through what we are working with and being curious about it and understanding our approach is really helpful.
What wouldn't be helpful is to show up (in person or online), try to implement a bunch of new changes without understanding the current student journey, and then don't follow through. The intention is good, but I think the students end up more confused.
It is a lot of work but so rewarding and I feel it has made me a better developer. I myself have learned so much from doing it. I have met so many amazing kids & parents too and it's made me a much bigger part of my community.
I think mainstream bootcamps value getting a job more than actually becoming a good engineer, which is something I disagree with. This also motivates me to show up at the library every day.
If I find some great engineers in the library and bring them into my team directly, will they do good work? If they prove to do good engineering, would it be a sustainable model to hire directly from our volunteer program?
If this works out, we could replicate this model in other public libraries as well.
The comments above are my moonshot ideas, I'm currently experimenting to see if it can be possible.
You can reach me via email: library at zheng.network
Other posts here are suggesting the choice is either a commercial school like Lambda or Harvard, so it's also worth noting that there is a whole range of options for folks.
Obviously, there are public and community colleges that teach programming. Many of these have a long history of working with non-traditional students who have jobs or other commitments. Some are remote-only or remote-friendly. In my state (GA), tuition for an accredited public community college is about $6,000 over 2 years for a programming diploma. That's $3,000 per year, and there's no income share required. (I'm not sure if people can do it faster for less money.)
There are also a set of nonprofits and foundations that offer free programming programs to certain populations. I believe NYC has something along these lines.
These are very much not the same. For profit schools have an incentive to get more students through while maintaining an acceptable level of quality because that way they make more money. Non-profits have much less motivation to make more money. If they get more applications than they want to fill their prestige goes up as does the quality of their student body. So you see the top US universities having more or less the same size student bodies as when there were 200m people living in the US.
Non-profits optimise for a pleasant work environment for faculty, for profits for maximum numbers students.
Some do, some don't. Some Non-profits optimize for maximize impact to their vision.
> for profits for maximum numbers students.
Some do, some don't. Some for-profits optimize a small number of wealthy clients.
Some of the regular students that come into the library lives off non-profit assistance and one is homeless. They speak highly of local nonprofits providing them food / shelter. I've volunteered at a few shelters that were impactful (don't know them by name unfortunately, I'm too privileged). Staff there definitely didn't have the comforts I have at my current work place.
I spent a chunk of my childhood growing up in Beverly Hills around some peers whose parents run profitable businesses. They have always valued few wealthy clients over many average clients.
Again, take my experience with a grain of salt, I'm not interested enough to prove this point.
Good on you for responding and explaining your reasoning.
I'm not a beginner coder, but I haven't worked on projects in a team environment before, so I signed up on c0d3.com hoping to gain some experience.
Any pointers?
If you think it is helpful, you can breeze through the lessons and exercises so you know / understand how other students use the curriculum, then work with the more senior students as a team on c0d3.com and build features to make it better. Proudly put all your contributions (features, mentoring work, authoring work - if you contributed to the course material) on your resume.
One thing I suggest is focus on code reviews for your students. I think that’s the best way to tech programming. They’ve already taken a stab at something, so writing comments on how it can be done better and why that way is better is much more useful than a lecture about something abstract.
It's my opinion that some companies should never be done for profit, particularly Political Tech but also other social tech including portions of ed tech.
These industries should be left for FI/RE folks who are rewarded with various social credits, like awards, tax credits, and peer validation.
This idea doesn't pass a common sense test to me - I'm sure bootcamps can be profitable, but the people running them are used to building things that scale. Bootcamps definitely don't scale. If these tech people were looking to get-rich-quick they surely wouldn't be running a school, of all things, even a profitable one.
There’s no requirement they be part of the “tech scene” ... whatever that is.
The companies they're trying to get people hired into.
There are definitely people who can do a challenging Master's degree with minimal help and feedback, just books, marked essay assignments and a final written exam. This is a model that's as old as the University of London. TAs and structure will increase the proportion of those who start who actually finish.
What about running a school obviously doesn't scale?
They said bootcamps don't scale, not that education or schools as a whole doesn't scale. And your solution of raising the entry requirements so only people who can teach themselves can attend is proving their point. Someone who can teach themself isn't going to pay to go to a bootcamp, but people who want to switch careers and can't figure it out without help, will.
I mentioned two solutions, raising entry standards and providing greater support. There are people who can teach themselves. There are very few people who can teach themselves as fast and as well in a self-directed environment as they can learn with a real curriculum and some feedback on their work. If you are confident you can teach yourself in two years or reach the same level of employability after nine months of Lambda School the latter looks pretty good even at $30K once you get a job paying over $50K.
20-30x students per cohort who paid ~$20K upfront for a 12 week program. 20% signing fee (based on 1st year comp) from employer on placement.
We definitely were not paying top of market as some of these students ended up at Uber and Facebook. That said the all in 1st year cost between base + signing bonus + equity wasn't much short of $200K. So:
30 * $20K + 28 * $200K * 20% = $1.72M/cohort
As for outgoings, all of the mentors were volunteers. As were most of the instructors. The content is mostly a one-time sunk cost to produce and is redelivered across cohorts. The largest overhead would have been a building lease. The biggest constraint on growth is how large you can make a cohort or how many cohorts you run (either multiple per year, or opening new locations).
Really felt like a bit of a racket that had found what was almost an arbitrage: between the inability of Bay Area companies to find local talent, the huge costs and risk associated trying to relocate people via H1B, and the desire for people to re-skill at any cost because tech jobs/salaries were distorting everything else in their city.
Sure it's not a $1B outcome. It's a pretty profitable and repeatable business, and especially given the limited downside risk (mostly carried by the students, who've already paid).
As additional evidence for this, the number of 1B+ exits in this space can be counted on one hand. General Assembly, for example, was acquired for ~$400M and it was one of the largest players. There was a player from Utah whose name eludes me now that had a, I think, ~$2B exit but that might be the only one.
Are you thinking of Instructure? They're based in Salt Lake City, and Marketwatch gives their current market cap as $1.86B
It's true that Lambda is incredibly disorganised and the build weeks etc are chaotic. Equally true that they don't do a good enough job of ensuring we have something to show for ourselves on our portfolio.
It's also true that their admission standards are seemingly incredibly lax. About 40% of my cohort struggle to code at a fundamental level - I don't mean that harshly, it's Lambda's fault
With that said, I've really enjoyed my time at Lambda overall and it saddens me to see it fail like this. The atmosphere and internal culture that they cultivated is second to none and I have enjoyed my time there a lot.
As with many people at Lambda, I joined them at a difficult time of my life, when I was suffering from pretty severe depression. I knew I loved coding but barely spent any time doing it and struggled with impostor syndrome, etc.
While at Lambda I benefitted hugely from the daily structure and discipline, and from having a community of people in the same position as me. I've made some great friends, and met some very smart and talented people.
What pains me is the embarrassment of appearing like some clueless fool who got caught up in some get-rich-quick scheme. I love programming, and I just wanted a structured curriculum to train as a professional.
Don't let this stop you. The world needs more good engineers, and if you practice your craft you will always find a home. There are plenty of industry professionals who now look a bit silly for their choice of company (Uber, WeWork) but ultimately it's all just a job and if you have the raw skills you can find a new gig.
Been there. I think this feeling comes mostly from the fact that this initiative has had a lot of attention, and of course, a lot of public criticism from people that weren't really able to give qualified opinion.
If you are having a good time there, if you are perceiving value, learning new things that makes sense and seem useful, just ignore those bad opinions and try not to think how others are looking at you.
I myself had a course on investing given by a moderately famous youtuber, and the environment around it was full of sarcasm and debauchery. I somehow managed to ignore it, took the course to the bone and now, less than 12 months after finishing, had a more than 3-fold return on the money I spent on it.
If you are learning more than you would have otherwise, or if you get a decent job at the end of this joining does not make you a clueless fool. Other people who don't know you or what you have done might think so but they don't matter. What matters is what you've learned, and that you'll get a job out of this in the end. Haters gonna hate.
Expecting 10 people who don't have any experience programming to cooperate on a project without any support or oversight is just asking for student failures.
Of course they'll say that the students have supports through their PMs, EMs or TLs (depending on the mood, they change the role title), but they're never available and miss meetings constantly. Also, they've reogranized the curriculum multiple times during the tenure of my friends, and don't wait till the next batch like a sane school would.
I feel really bad for the excellent teachers they brought on board. They ended up with a lot more than they bargained for.
Half or more of the program is composed of the crappy group projects.
The people who succeed after lambda school is in spite of the program, not because of it.
If you want the stability and reliability of Harvard, don't get your education from a scrappy upstart.
Having said that, it still seems like a bad deal even for students who get jobs. If the curriculum is not up to par, it's likely that successful students would have also been successful with self-teaching or less-expensive online courses. There's also the opportunity cost of being out of work for 9 months.
It seems unfair to treat your students as guinea pigs for iterating a startup - especially when you're already often targeting desperate people.
> Of course they'll say that the students have supports through their PMs, EMs or TLs (depending on the mood, they change the role title), but they're never available and miss meetings constantly.
This statement is full of hyperbole, however it gets at what I feel is Lambda's biggest problem: Team Lead quality and consistency.
The problem is a misalignment of incentives (sound familiar?). TLs are current students who are further ahead in the curriculum. The schedule makes it effectively impossible to have a job and TL, in addition to being a student, so I'm guessing most of those who apply for a TL position do so largely for the paycheck.
I don't mean to say that all TLs are bad or don't care about the success of their teams. In fact my experience has been mostly positive. But ultimately students are not at Lambda to be TLs. They are at Lambda to get high-paying tech jobs.
From the point of the business, I can see the trade off. Running a cheap shop is essential for the kind of deal they offer: A time-capped value-capped full-risk sharing agreement.
Lambda's model could have much, much higher quality if there were a less restraining cap on the ISA, but that will affect sign-ups and might increase the trouble of the people that fall-off the track.
I dont know what is the best positioning of price-quality, but I'm sure many people would prefer to pay more through ISA to have a higher success rate and stronger guidance.
That said, Lambda is really just 3 years old, they need to figure out a million things.
Not sure when you graduated - I'm sure it was better when there were more resources and less students - but today that's not the case.
The idea of bootcamps is fantastic, they allow people that may have not been exposed to computer science to get up to speed on technology and transition careers.
However, there are several problems. First and foremost, while 3-6-9 months are great when you are going from zero knowledge, the challenge is that isn't enough time to really be a junior developer unless you have prior experience. So you will need to continue to augment your education after graduation to ensure that you get a well paying job.
Most people attending bootcamps are doing so after college and later in life, which means even though some bootcamps are cheaper than attending a semester or a full year of college they are still quite expensive because the students are not "students" in that they are usually adults and need to figure out how to pay for school, attend classes, while effectively receiving zero income.
And applying for loans is much more complex because this isn't the same as taking out student debt for college.
The second issue is that all of these bootcamps say that they are for beginners with zero knowledge, but looking over their curriculums that simply isn't the case. Software engineering has gotten much more complex over the past decade. When I first started fumbling around with it myself I could just write some PHP or Perl code, and get up and running quickly. Today you need to know about github, javascript libraries, frameworks, and the typical "Hello World" application isn't a direct route.
Most of these schools don't realize that their first students are already exposed to these concepts, but later students aren't so they don't really adjust their curriculum.
With lambda in particular it also is a bit confusing because the school is online. Which should mean that you are able to provide the service for a lower cost, but they are charging the same amount as in person physically attended schools.
There are a lot of deceptive practices in the industry, lambda isn't alone, such as taking recent grads and giving them low paying jobs as TA (teacher assistants) so that they can provide help to students at a lower cost while also allowing the school to claim that they have higher placement.
The curriculums are really dependent on a per school basis, but I have seen a bunch of stuff that simply doesn't make sense and makes it more challenging for students.
One school had students do a group project, which wouldn't count as part of their final for the first "mod" of the school. However they continued to teach things that you need for the final which would be a personal project. If you got assigned to a bad team you would be working much slower and not able to keep up and then your final which has nothing to do with a group determines whether you proceed or repeat the course (they charge you to repeat). Also it's important to note that 50% of the students didn't pass the first mod, which means you have a 50% chance of being on a slow team that would hamper your learning. Their advice was that it is important to learn to pair, and I agree, but when you get to your first job you are pairing with people that have experience, not where your partner has a 50% chance of failing out.
I firmly believe that bootcamps and providing secondary education choices are essential and if done correctly can really begin to combat the monopoly that colleges hold over education, but it's a challenging mission.
With education there are student loans that you can take out and I think that is essential to get this going in the US because it is simply impossible for most people to not have any income and still pay for schooling for even 6 months, much less a longer period of time.
The other challenge is that you really need to have 3 terms. Beginner, intermediate, advanced. Each student can then apply based on skill set to determine where they place and students can move from one to the next, with each section being 4-5 months. If you did 15 months of education you would be much better off than what the school provides. It's great to get from zero to one in terms of knowledge, but students are still left far from having skills that are immediately beneficial to employers.
There's definitely more work that needs to be done.
As for Lambda itself, when you look at how other coding bootcamps have fared financially it doesn't paint a rosy picture. It's a challenging space to operate and the VC style returns simply aren't there. If you want to offer an online only education that is fantastic but you have places like codecademy that do that and also do not charge you $30k for the privilege of basically accessing information that is online for free.
The challenge with coding is that it really is much easier to get going when you have someone you can ask questions from, so helping to improve that aspect of it while providing it online at a low cost is really the challenge.
With the belief that education is all about laying solid foundation for life-long learning, for job or not, I don't really see any need for coding camps. I'm not denying there are success stories, but I don't see coding camps make statistical sense.
Basically it's unplanned and so much harder to able to commit two years.
Certainly do-able, but challenging when you are thinking of it from a consumer perspective. Spend two years working towards a career shift or get it done in 6 months.
Maybe community colleges can do a better job of marketing themselves.
But ultimately I think the fact that computer science isn't a requirement in all education is criminal. We study "Math" and "English" in school. "Computer Science" is the equivalent of math 100 years ago, it needs to be a mandated requirement.
Would mind expanding on that?
You also can't live in the modern world without language and the ability to communicate.
To me math is a language. Different than our verbal languages, but it is still a language none the less and essential.
Programming to me is also a language. And it is as essential today as math was originally. There was a long period of history where lower economic classes were prevented from learning and educating themselves and the fact that we have education that is subsidized by the government to various levels in all major countries is something that we take for granted, but it wasn't a right that our ancestors had.
The way schools force all kids to learn math, I feel is how schools today should force all kids to learn programming and computer science.
1. No or low admissions bar. Some students very simply SHOULD NOT be admitted to what is intended to be a 3 month course. They either don’t have the learning ability that is required to absorb all of the content in 3 months or the drive.
2. Curriculum pacing - pretty much every human learns different concepts at different pace. A lot of code schools are just trying to assembly-line their course & put people in one end and shove them out the other end in 3 months. If these boot camps offered remediation at a unit level, students on the whole would go much further much faster.
3. Curriculum focus - some people want to be data scientists, others want to write backend code, others want to be iOS developers, others front end. A lot of bootcamps really have curriculum hitting very broad strokes in that 3 month period: big-o, SOLID, html, css, npm/yarn/etc, command line interface, JavaScript, react, elixir, etc. etc. is too much to cram into 3 months IMO. This is a consequence of students hearing this buzzword soup from friends and internet & bootcamps not owning responsibility to serve students.
4. Outcome reporting is hazy at best - CIRR is IMO the best framework for reporting student numbers & should be used as the gold standard. However it should be enforced strictly, if a member doesn’t report CIRR numbers within 3 or 6 months of a cohort finishing they should be penalized in some way by being removed or fined or something. Students need good & recent data to make good decisions & it’s just not their these days.
There are more problems but those are the biggest IMO. There is still a huge opportunity in helping people learn to code but I think it looks more like learncodethehardway.com than Lambdaschool.com
It is certainly not meaningless! Selling an ISA means that Lambda no longer has any financial interest in its outcome. Borrowing against an ISA is completely different; if the ISA doesn't pay out then Lambda goes bankrupt, which is precisely the incentive alignment they claim to have.
If investors in Lambda ISAs are not getting good returns, the value of buying a Lambda ISA will go down. If Lambda can't sell ISAs for a good price, they won't get the operating income they need or will have to shoulder the ISA themselves.
Still, selling the ISA just feels shady relative to the marketing that touts ISAs as aligning incentives.
Just like if you take out a mortgage on a house and fail to pay the loan back, then the bank owns the house.
In this case since the ISA is used as collateral, the company that originated the loan now owns a lien on the ISA, effectively giving it ownership.
Similar to when you lease a car, there is a company that provides the finances for the lease, and has a lien on the car, which means you do not own it. Otherwise you could lease a car for $250/mo, then sell it the next day for $30k, but you can't because there is a lien on the title.
So in this case, while the loan is outstanding, the originating loan company effectively owns the asset used for collateral. Ownership means you have 100% control over the asset, and in this case, Lambda has given away 100% control over the ISA.
It also means that they are not aligned anymore, since they have received financial compensation for the ISA up front, they can default on their repayment of the loan as it doesn't matter because the collateral aren't shares in their company, but just the ISA itself.
So if I take out a mortgage, and fail to pay for it, the bank gets the house, I am not still obligated to pay the remaining loan amount. Sure there are a few other items that occur in that process, but ultimately the loan is forgiven, of course with some credit penalties and future ability to take out loans. But I am not responsible to continue repaying the loan.
If the loan is secured by the ISA the same thing applies here. Sure there could be other stipulations and without reviewing the contracts there is no way to know, but stands to reason that the more likely situation is that they get the ISA and then can use aggressive tactics to go after the students for collections such as garnishing wages and also potentially charging them an interest for failure to repay. Though the original ISA believe has no interest, that's not to say that late penalties or other fees can't be added in later potentially.
To bring up the car loan example you mentioned earlier, if you decide after a year that you don't want your car anymore you can't just drive it to your bank and drop it off instead of paying off the rest of the loan. If you stop paying it they can repossess it but you will still owe whatever is left on your loan balance after they auction it.
I think most homeowners would disagree that when they bought the house, it was actually sold to the bank.
You also make reference to some internal documents. Perhaps sharing them will help further the story?
It's hard to argue that there's clearly a major risk involved in attending a program like this, but at the same time it seems to be working for some individuals. Would these high-achieving individuals be able to teach themselves? Possibly. But I'm not sure that going back to the traditional college route is the right choice either. And how can we better help those who lack the aptitude or enjoyment of programming and find out too late into these programs?
Was this at $10k/ISA? And was it half of all ISAs or all of half of the ISAs?
Also, CoderPad was life changing when we started using it at Airbnb. It made phone interviews so much more effective, thanks for creating it.
"The income share agreement has no interest. It's a flat percentage that goes away once you've reached the $30k payment cap, you've made 24 payments, or after 60 months of deferred payments (even if you haven't paid us anything)."
But you wrote in the article:
"Students with no safety nets experience real financial pain from the nine-month hiatus from work, in addition to the looming dread of possibly having to pay Lambda $30K one day."
Overall, I think this was an insightful article, but cutting that out seems misleading, even underhanded.
But if the editor cut those 14 words out, it wasn't to make the article more concise, it was to make the audience more outraged and more likely to share that article with others -- thereby increase the number of page views.
Not that I would expect the author to admit that. The author can't speak to the intent of his editor, obviously. And even if he could, I don't think his editor would be too pleased with him if he did that.
Lambda School is <5 years old. Pretty disingenuous man.
Is that too difficult to answer for you?
It’s a pretty horrible analogy - demeaning to cancer patients aside from anything else, and seems to seek to tarnish a seemingly well-intentioned business by linking it to something awful, without any basis.
I don’t suggest the company is beyond criticism or without a need to improve; it’s still a fairly new company after all.
This class of things includes pretty much anything that's really important in the world - including economic systems, government policies, human behaviours, and, yes, companies, including ones that offer deferred payment for education that can lead to higher incomes and better standards of living for the customer.
You're clearly intelligent enough to know this but are poisoning the debate by making a comparison with something that is horrible but not at all comparable.
It is clear that I not only didn't get my point across but that people don't even believe that the point existed, and I miscommunicated so badly that you're not even willing to believe that I was ever arguing in good faith. Fuck this.
People need to learn things. If funding education through an ISV is the best way to do that, then they should do that, even if there are problems with it. What the article really seemed to be implying was that they shouldn't do it that way, that there was something wrong with it, which justified an extremely negative view of the company, and leaves people with no option to learn to code without large upfront costs or loans to pay back if they are unsuccessful.
Describing the very best option in purely negative terms is misleading.
If you are telling people to stop doing a thing with a x% chance of working, but the other options have a lower chance of working, you are effectively saying don't do the thing at all.
Also how much time did you spend on all of the research for this piece?
As a totally outsider (not even from the US), Lambda School seemed too good to be true, in a good way, and I said so here in HN. 9 months of serious commitment, no upfront payment, ISA, all of it seemed good to me.
These days, I don't think the same way. Lambda School and it's founder seem more worried about being a billion dollar company than doing the right thing for its students. Personally, that's strategically stupid and counter-productive. They won't be a billion dollar company because they don't care about their students. But it looks like the direction they took. This mistake is not inherent to VC ventured companies or Silicon Valley, so I think the ones to blame are Lambda leaders, not the system.
So, I still recommend freeCodeCamp.org as the best thing that exists for people that want to learn to code and get a job. I wish some of these effective philanthropy multi millionaire would give them generous money to pursue their mission.
It changed my life.
Media outlets have incentives to either paint you as the second coming of Christ or as Satan. It appears Lambda, for a while, actually succeeded at convincing journalists they were the former.
After a while, people get bored with that though. The incentives that drive clicks flip. Suddenly Lambda is now Satan. Burn it down! Downvote all sympathizers!
Here’s the reality: all models for education can work for certain people in certain instances. Lambda is definitely the best choice for some people. But no single company is going to solve something like “education” or “healthcare” because they are political institutions tied to the power dynamics that determine how society is arranged. You cannot brute force this without gaining influence over government itself.
This is not as simple as disrupting where people buy their shampoo or where they see ads.
If you stated that you have an 86% placement record and in reality it is 50%, that is a pretty large discrepancy. If the original 86% placement was from the first 70-ish students and now you are over 2500 students, that seems a bit fraudulent.
If placement rates aren't critical to you getting students, then you can say it is 50% publicly and see if that affects your enrollment numbers or not. Otherwise, it would stand to reason that stating a higher placement rate gets you more students.
Also, this isn't run a a non-profit organization, its a for-profit enterprise. So they have a financial incentive to get more students because that equates to more value for them.
This is my point. Journalists are picking a narrative first, then seeking out facts to justify that position.
When they thought lambda was going to fix education they were more than happy to report the 86% number without any research. Now that lambda is “evil” they look for the lowest number they can find.
I’m sure the truth is somewhere in the middle.
Also if you were an institution interested in disclosing information fairly you would simply list all cohorts and let the consumer decide.
But they aren't doing that and instead claim a single number.
By the way, having had recent experiences with a number of bootcamps through friends I would say that the 50% number anecdotally is accurate. That's also not taking in to account the drop out rate. If you look at people starting the bootcamp to placement it would be even lower.
If you have 465,000 customers and you want to round that up to 500k - ok, not exactly true, but whether you have 465k or 500k isn't going to change my decision about using your product.
Telling me you place 80% of your students and actually placing 50% is a big deal. 80% means I have a 4 out of 5 chance of being successful. 50% means I have a 1 out of 2 chance of being successful. Those odds are very different and certainly part of the marketing push to get people to sign up.
What this means is that bullshitting works until it doesn't.
I've seen people have success in startups using similar tactics throughout my career, and one realization I've had about this is that sometimes perception matters more than the actual numbers. It's especially true when it comes to investing and VC fueled businesses where success tends to follow the funding (i.e., if you get enough money you can buy your way to success).
Eventually, however, everything comes out in the wash. At some point lying about the numbers won't work anymore, and maybe for Allred this is an example of that. As a founder, if you don't eventually deliver on what you've promised, it will all unwind and you'll be left with nothing.
We moderate HN less, not more, when YC or YC-funded startups are in a story. That's literally the first principle of HN moderation, in the sense that it's the first thing PG told me when he was training me to do this in 2012. I didn't even have time to grab a chair before he blurted it out.
It's natural for people to question this, so such questions will always come up. We're scrupulous about following this rule for two reasons. The first is that we need to be able to answer the questions in good conscience; the second is that moderating any other way would be dumb. The community's trust is the most valuable asset HN has. In fact, if you think about it, it's the only asset HN has.
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
[0] https://lambdaschool.com/the-commons/announcing-our-new-isa-...
[1] https://lambdaschool.com/the-commons/building-better-data-tr...
On a more meta level, it feels like we’ve barely touched the margins of how humans learn best. That’s always struck me as odd.
Like nobody has quite cracked the nut for how to do it. Not with programming, not with language, not with cooking. It’s always so much effort. As if the lessons people have learned are forgotten as soon as a few years pass.
We know a little bit. Spaced repetition helps. Not doing anything after studying intensely helps (as in literally just doing nothing after studying improves retention).
Going for long walks helps. So does transforming concepts into something you can see (I suspect because we have a lot of “machinery” for visual computations).
Engaging more of your muscles during an activity (writing as opposed to typing) helps too.
Engaging extra senses (like smell) seems to help.
But that’s about it. There’s no common modality, no library of how humans learn, retain, and transfer knowledge.
It’s been like what? 200,000 years since we’ve been on this planet.
We’ve barely touched the surface of what we’re capable of doing because we keep forgetting.
I think it’s naive to think there one way that will work for everyone. Attempts to find such a solution inevitably fail because it’s not something that can be generalized and then scaled. The more mechanized and focused the process, the more outliers get left behind.
It's very well studied. I'd start by reading a cognitive science textbook and go from there. I have Cognition by Reisberg[0], it was well-written and well-structured.
Alternatively, an entertaining and informative popular account is Make It Stick by Brown, Roediger and McDaniel[1].
I do agree that the research findings are greatly underapplied, though.
[0] https://www.amazon.com/Cognition-Exploring-Science-Mind-Seve...
[1] https://www.amazon.com/Make-Stick-Science-Successful-Learnin...
Thank you.
When someone shows how to install i.e. kubernetes they show you the happy path. Once something goes wrong you can't debug the problem. The alternative is to simply try installing kubernetes yourself. You will probably make some major mistakes but you will also gain enough experience to deal with all the edge cases when they do occur.
In my 30+ year carreer in IT the most beneficial insights came from brilliant people I had the pleasure to work with and who were willing to share.
In a way that's self-evident. Really smart people with a true clue about what they are doing are not wallowing in angst that their trainees (for lack of a better word) would ever endanger their position and they're often happy to share.
That does not discount the trial and error approach, which could work quite brilliantly when the documentation was good. (being something of a dinosaur I'd happily refer to anything DEC in its heyday).
But there is that list of people, which I can't really thank enough for instilling some of the invaluable knowledge, which defined my career.
He got starry-eyed working in growth at LendUp and wanted to build his own rocket ship. I don't think he ever put the students above running a trendy fast-growing startup.
There are real beefs Lambda will have to sort out on how it runs things, but running with VC money isn't one of them. Also, weird to have that opinion in HN, made by an incubator that sells startups to VCs.
I think the bundling and selling of student CDOs works for a company in its early stages, since the costs for recovering unpaid tuition can range from high to being declared a total loss. Some of the anecdotes make it seem like the company doesn't receive enough feedback from employers in order to strengthen parts of their curriculum that make interviewees fail at some stages during the interview process. Creating a complete curriculum takes more than bundling together Leetcode and Hackerrank problems, where the actual teaching reflects that bootcamps and a CS education aren't in direct parity with one another. [1]
Regarding the finances, student loan debt through an established institution can't be removed through filing for bankruptcy. This allows access to federal loans and underwriters that give better rates and agreements for all the business parties invloved. I've been denied refinancing my loans from multiple companies for a combination of those effects including not graduating and having too low of an income. The business seems like a good idea but too fixated on higher short-term valuations in order to attract VC capital; rather than fixing its underlying structural problems.
[0] https://news.ycombinator.com/item?id=15011033
[1] https://hackernoon.com/bootcamps-vs-college-11dd76a4d127
Overall I think they're trying.
The problem with the PM/TL program is the people who are normally the best at it tend to be the best at finding jobs. The TLs that stick around forever tend to be ones that aren't the best or the worst who are normally replaced when someone complains. Instead it's the okay people who are just cruising.
One thing that seems to be overlooked about Lambda is the fact that they are remote. An applicant from the middle of nowhere who lives hours from the nearest city has as much chance of getting in as anyone. It can be harder to find a job there and they won't relocate. Remote work is much harder to get in as a junior dev.
Overall though, I wouldn't be in the position I'm currently in without them though. Maybe I just made it in before the line, maybe they're just undergoing growing pains as expand to a profitable size.
Lambda makes money through some financial engineering : Ivy league schools are essentially hedge funds with their multibillion-dollar endowments
Lambda's curriculum is lacking : Most CS students are taught theory, but learn very few of the skills they'll actually need on the job ( e.g. application deployment, source code management, etc.)
Lambda inflates its job placement stats : This is never defensible, but also highly prevalent in both non-profit and for-profit higher education institutions.
I commend Austen for at least attempting to better align students' and educators' financial incentives.
- There are a relatively fixed number of high skill jobs.
- More people with the same skills/portfolio reduces the value of that skill/portfolio.
- While people can learn new skills, people have different learning rates and integrating that knowledge deeply often takes much longer than teachers/bootcamps want to believe.
- Most edu info exists for free online, what school/bootcamps uniquely provide is feedback. Quality feedback is HARD to scale.
- How much does pre-application screening find people who don't really need that much help vs trying to help every applicant and getting unmotivated people in your bootcamp. I have seen very few examples of a bootcamp fundamentally changing someones motivation/work ethic.
ISAs will be the big winner in the ed tech space and providers of education will likely come and go unless someone cracks all the issues above. Perhaps the current distribution of students to universities (having plurality) is actually more suited to serving this function than a typical tech world power law distribution where one or two companies would provide 80% of the edu.
What is it about programming that makes people believe that they can learn it in a hilariously short amount of time? Sure, we programmers may be harboring some imposter syndrome and secretly believe that programming is super easy. Sure, programming doesn't require lots of math. But it's still hard! It's still a craft that requires problem solving ability, lots of semi-arcane knowledge and a detail oriented mind.
I also helped a person go from zero knowledge of programming to pumping out android games in a month of self study. So I know that smart, motivated people can get pretty far pretty fast. In regards to your point, I don't think this person knows UML or ever learned big O notation, but this person was able to gain the skills required to do what they wanted, plus some pro-programmer friends and stackoverflow to fill in the "semi-arcane knowledge" when required.
> front-end developers
I suspect this is part of the issue. Plenty of people believe that front-end is easier or lesser than back-end, and therefore can be learned quickly. Lots and lots of terrible front-end code has taught me that this is not true. Good front-end developers are really hard to find and subpar ones can lead to horrible user interfaces and direct impact to the bottom line. Even if we're talking the most minimal of front-end stacks, i.e. HTML, CSS and JS, which, I'm not even sure people hire for anyways, there's a lot of subtle issues with accessibility, responsive UI, writing halfway decent JS, etc. If we add on the various libraries (a sign of a good developer is also knowing when to use these libs and when to avoid them), then the amount of knowledge required is far far more than 9 months can provide.
In my own 20 years of programming experience I've kept improving significantly every year. There are so many different areas to cover (hardware internals, compilers, languages, tooling, databases, networking, graphics, user interaction, i18n, cryptography, compression, etc) and in each area there are so many different competing ideas. Each of them, when deeply understood, will provide insights in other areas. I don't see my learning hitting significantly diminishing returns before death by old age.
I'm curious, at what year of your programming history did you find yourself not progressing in fundamental software thinking and how long has the plateau lasted?
> So where does that 86 percent figure come from? Lambda has reported graduate-outcome statistics at the Council on Integrity in Results Reporting (CIRR), a voluntary trade organization of coding boot camps whose purpose is to ensure that participating schools publish truthful information about student outcomes. Allred has often used this report to defend his company online. But where other boot camps have multiple reports spanning many student cohorts, Lambda has only reported statistics for its first 71 graduates — 86 percent of who, the school claims, found jobs. Sheree Speakman, the CEO of CIRR, told me that Lambda has not undergone the standard independent auditing for the sole report it has submitted, and that her communications to Lambda School regarding further reporting and auditing have gone unanswered.
I wonder whether the quality of education at coding bootcamps has gone downhill. From what I know of the Web Development Immersive program at General Assembly in 2014, or the software engineer program at Hack Reactor in 2015, they both were very decent.
And the problems mentioned in the article are straightforward to fix:
(a) Be honest and transparent about the job placement rate for each batch
(b) Don't sell the ISA or borrow against it
(c) Improve the quality of teaching and accept a lower growth rate
The problem is that switchup uses the affiliate links of the bootcamps, so it has a financial incentive to sell the bootcamps it features to its viewers. That's how it makes its money.
Now Amazon does the same thing, you'll say. And you'd be right, but with Amazon, it has a reputation to maintain. If Amazon reviews are unreliable, people will stop shopping at Amazon.
With Switchup, it's different. Most people will only buy one coding bootcamp in their lifetime. So withchup doesn't have the same incentive to make its reviews super accurate. It mostly has the incentive to sell as many coding bootcamps as it can.
https://leif.org/api/products/5b5b8bd0e59b743f9a086ed9/pdf
It says that you agree that you "are entering into this Agreement in good faith and with the intention to pay us" and will "make reasonable and good faith efforts to seek employment" as long as you are not paying them.
I don't see any specific provisions describing how that would be enforced and I bet in practice it is not a major issue for the reason I mentioned above: there isn't much reason to go through with Lambda School if you don't actually want a job in tech, so it probably doesn't happen much.
The Lambda model is set up to operate at scale -- and teaching people who don't want to get a career in tech or change their mind is baked into their operational model.
"lmao I woke my boy up" [1]
Sounds like he's keeping score of something, which isn't a quality I necessarily want from my "journalists".
30k? Seriously? my trade tech certification was only $2900 for two years. Is Lambda School a 4 year program or a college of some sort? Is room and board of some kind included? I mean I get that STEM pays a lot...but the cost here seems a little steep.
There are only 4 STEM related courses on the site. my trade-tech school had nearly 20 by the time I graduated.
I would argue those other aspects of higher education are far more critical than the bootcamp model and vision.
And eventually, it boils down to a waste of time for some subset of students, which is not helpful to our higher education interests.
Coding bootcamps are not all bad, though, it's just that I think if you are going to do one then 1) you really need a solid financial cushion (either parental or own savings) and 2) you should really do a top rated in person one. I can't imagine doing this through a Zoom for 9 months in my apartment alone. The value of having social support you get from instructors and being around other students in the same boat as you can't be understated.
I'd like to see ten-year outcomes before claiming this. That said, it's also hard to control for different student backgrounds.
are they collecting 17% of your paycheck?
There are plenty of self-studied developers who have full-time jobs on HN. I've met a vast number of devs in real life who didn't go to college. It's not incredibly rare. It costs almost nothing but time to self-study and build a portfolio.
The other option would be go thru university. And there are many universities that offer a BS in CS for a reasonable price such as University of Florida. You can apply for a FAFSA, and depending on your financial needs, you could get it all paid for in grants. It will likely take 3 years but you get a legit degree by an accredited University, not some University of Phoenix equivalent scam. Some of these online BSCS students even get accepted to reputable PhD programs, good luck doing that from Lambda School because it's not happening.
At least in the context of something like Lambda school, that does not teach computer science, as I understand it.
if you decide you want an accredited university degree, then you can get FAFSA to fund community college, then transfer an online university such as Florida
I think the negative picture painted by the author masks that the company is placing 50% of graduates which mostly paid nothing for the education they received. Lambda is not high quality education, but its super cheap and higher quality than just reading a Knuth Book, or Skenna's algo book which has no completion rate.
Either way, you are not becoming a true machine learning researcher via self-study nor a bootcamp. You will need at least a master's in CS/Math/Stats to land those roles. A bootcamp is neither the lowest cost to a softwaredev career nor is it the a career effective maximizing choice.
If government agencies really had their acts together, then these courses, as well as many others would be free, as part of 'extended learning' programs.
For all the crazy amount of grant money being thrown around by so many governments, and given the need for certain disciplines, it's odd that this doesn't happen.
The housing bubble occurred, in part, because there was optimism that acquiring a certain tangible asset - a house - was a sure path to prosperity for the buyer. And the buyers of mortgage-backed securities had decades of data showing that these were solid investments.
What about intangible assets like knowing how to code? Is this what Lambda School is really doing?
And like the housing bubble, once the originator can sell the debt, they have some incentives to inflate expectations on both sides.
Perhaps Lambda School’s ISA-selling isn’t vulnerable to this flaw. Maybe the ISA-buyers are doing due diligence. But you would have thought the same was true in 2008 too.
Like everyone with a Twitter account I’ve interacted with Austen. And I had a mildly positive impression, or at least I thought he deserved more credit. Not sure about that any more. I would like to see his response.
https://www.theverge.com/2020/2/11/21131848/lambda-school-co...