Lambda School leaked documents show poor performance over the last two years • Hacker News Board

Re: Lambda School leaked documents show poor performance over the last two years

#61

Post by virtuous_signal » Mon, Oct 25, 2021, 5:59 PM UTC

whoisjuan wrote: post #19

In their eagerness to disrupt traditional education, all these programs forgot that the one single thing that makes traditional university education succesful is teaching people with aptitude and drive. To get accepted and graduate from a reputable university you have to get good results in standardized tests and work your self through multiple courses for at least 3 years. That by itself doesn't mean anything but th…

> To get accepted and graduate from a reputable university you have to get good results in standardized tests and work your self through multiple courses for at least 3 years. That by itself doesn't mean anything but there's a level of rigor and self-drive required to succesfully traverse this experience.

It's worth noting that Lamda school actually does use an industry-standard "IQ test" [0], although it seems the cutoff is quite low (23 out of 50 whereas the median score according to Criteria Corp is 24).

[0] https://lambdaschool.com/the-commons/how-to-ace-the-criteria...

Re: Lambda School leaked documents show poor performance over the last two years

#62

Post by mywittyname » Mon, Oct 25, 2021, 5:59 PM UTC

Gortal278 wrote: post #31

I have always been surprised by how many students of lambda and equivalent graduate to immediately teach at these bootcamps. It just seems gross.

I get the impression that the business model for these bootcamps is largely the modeled after the one used in startup funding, where the bulk of the money will be made on the top 5-10% of the class.

There's also likely some element of grabbing mind-share while they work to automate the process (plus, gathering data on who make profitable students). Eliminating the instructor will allow N to grow much higher than 150 per course and reduce costs dramatically.

Re: Lambda School leaked documents show poor performance over the last two years

#63

Post by clpm4j » Mon, Oct 25, 2021, 6:00 PM UTC

notJim wrote: post #27

sam0x17 wrote: post #24

It isn't just Lambda School. I mentor for [insert large coding bootcamp here] and I will say things have been going badly across the board for more than 50% of students. They recently had to cut a lot of staff, and completion rates seem to hover around only 50-60% when you take into account all the students they remove from this calculation when they withdraw for "personal reasons". The employment numbers I don't hav…

> two of our best junior devs (at Arist YC S20) are Lambda School and Ironclad grads I just want to echo this, I have also worked with some really really sharp engineers coming out of these bootcamps. They were also people changing careers. I would really hate for the questionable aspects of these bootcamps to reflect on all of their grads, because some of the grads are fantastic.

From what I've seen, I agree about some of the grads being fantastic - the types who could have been extremely successful in a traditional college CS path if they had chosen. However, these are few and far between, the exception to the rule. Most (again, from what I can tell) really struggle, and don't have what it takes to succeed in this field (be it lack of genuine interest, focus, curiosity in technology - whatever). Add to this the fact that every company has had the "only hire the best; hiring is the most important thing in the world" mantra beat into their heads, and suddenly all of these would-be developers are far from employable. I routinely see and hear of entry level dev jobs (at small mediocre tech startups) getting over a thousand applications per role, and they reject all of them.

Re: Lambda School leaked documents show poor performance over the last two years

#64

Post by boringg » Mon, Oct 25, 2021, 6:01 PM UTC

throwaway789257 wrote: post #55

I used to work as a technical recruiter and made several hires out of coding bootcamps (though never from Lambda, which didn't exist at the time). One rule of thumb that we drew after many trials was that, if a candidate coming out of a coding bootcamp did not have a math or science background prior to that bootcamp, they probably would not pass our interview process. That is, people with a certain intellectual found…

I think this is a far assessment. There is a certain level of mathematical aptitude that is essential for being able to be successful at programming. Maybe no in the strict sense of math but in understanding how different structures and concepts are able to relate to each other. However in the strict sense of math, there is a requirement there.

I also posit that as bootcamps grew - the quality control and quality of hires(students) decreased with the continued pressure of growing for their investors. Leads to bad outcomes.

It is not dissimilar to when a company blitzscales and the quality of incoming hires diminishes the further down the graph you go. I realize its a bit harsh of a comparison but I do tend to find the earlier hires at solid companies typically have something the later hires don't (though many of the later hires are quite good at their specific roles).

Re: Lambda School leaked documents show poor performance over the last two years

#65

Post by tjpnz » Mon, Oct 25, 2021, 6:03 PM UTC

ctvo wrote: post #38

My partner went through a bootcamp, and there was an interview process after the application process. In my partner's case, they hold an advance degree in a STEM field from an Ivy League. A few of their peers had similar backgrounds, and that group didn't have trouble finishing. I wonder if the push to scale out these programs meant lowering standards, since folks with the ability / discipline to finish are few. If s…

I've interviewed a few candidates in a similar situation to your partner. My impression is that the bootcamps had very little to do with their success. I would put it more down to personal drive and possibly the programming STEM grads tend to get exposed to.

Re: Lambda School leaked documents show poor performance over the last two years

#66

Post by sam0x17 » Mon, Oct 25, 2021, 6:04 PM UTC

torben-friis wrote: post #44

sam0x17 wrote: post #35

Earlier quoted context omitted.

One huge mistake I think a lot (but not all of) the bootcamps are making is taking the easy route of simply using JavaScript to teach both frontend and backend. JavaScript is a TERRIBLE first programming language. I usually have to spend 2-3 sessions per mentee explaining the idiosyncrasies of JavaScript, where some of the OOP concepts actually come from, and how things work in other languages to put into context the…

Why not teach them functional concepts (map, filter, reduce, etc are well supported and comfortable) and leave the horrors of Js classes alone?

They run into enough problems with objectness just learning map / filter / reduce. I've seen so many angry / frustrated students trying to use a for each loop to iterate over an Array only to find that although this works in most languages, in js these just loop over the fucking KEYS of the object (which now opens up the discussion about how `obj[1]` is not a key but `obj['hey']` is a key and the differences therein.

Re: Lambda School leaked documents show poor performance over the last two years

#67

Post by ggreer » Mon, Oct 25, 2021, 6:04 PM UTC

akanet wrote: post #7

pillowkusis wrote: post #4

This article takes the angle that Lambda's profitability efforts may hurt its students: > the school can still reach profitability by enrolling 2,000 students a month while only placing less than half of its graduates in qualifying jobs. >Class sizes are like 150 students to one instructor — I've only heard that number going up I don't see what the problem is here. For students, the only cost associated with Lambda i…

It depends on what you conceive of the school as - is it an educational institution that has an incentive to improve the lives of each one of its students, incurring variable cost for each? Or is it a fishing expedition to land as many students as possible for as cheaply as possible, and letting the ones who were going to get jobs on their own anyway pay for the rest?

If we judged colleges by the same standards, I doubt they'd come off better. Normal colleges are fishing expeditions to land as many students as possible for as cheaply as possible and burden students with inescapable debt. Students have to pay even if they don't graduate or get a good job. 4-year graduation rates are below 50%. After 6 years, only 62% of college students have gotten a 4-year degree. The rest are stuck with debt that not even bankruptcy can forgive.

I don't know much about Lambda School, but my guess is that it does what colleges do, but faster and with incentives better aligned. College is more about distinguishing people than training them. (If education is actually about training, then why is >60% of the wage benefit of education from the credential rather than the years or credits earned?[1]) Someone with a college degree has shown they are reasonably intelligent and can obey instructions to complete boring tasks over long periods of time. This means they're probably a useful employee. I think Lambda School (and similar outfits) distinguish people more quickly than college and try to train them better.

To give people some idea of how bad education is: I learned computer science at an ABET-accredited college. The vast majority of the career skills I developed were from learning on my own, not from my coursework. Nowhere in my classes was there any discussion of source control. I had to teach my classmates how to use Subversion.

So often we forget that almost all education is terrible. We're only jarred out of status quo bias when a new type of school comes along.

1. https://en.wikipedia.org/wiki/Sheepskin_effect

Re: Lambda School leaked documents show poor performance over the last two years

#68

Post by akanet » Mon, Oct 25, 2021, 6:07 PM UTC

sam0x17 wrote: post #51

sam0x17 wrote: post #24

It isn't just Lambda School. I mentor for [insert large coding bootcamp here] and I will say things have been going badly across the board for more than 50% of students. They recently had to cut a lot of staff, and completion rates seem to hover around only 50-60% when you take into account all the students they remove from this calculation when they withdraw for "personal reasons". The employment numbers I don't hav…

This one particular student sticks with me. He was in his 60s-70s, and had a burgeoning career in IT/Security in the late 90s, but quit his job to take care of his ailing mother around 2001 or so. He took care of her, living off of her reverse mortgage for 20 years, but now she has passed away, and the term of the reverse mortgage ends soon (or I believe by now has ended) so the bank will be re-posessing his house, h…

Thanks for the story. The most brutal thing for me about reporting this story was each individual human horror-show. It's such a serious thing to be responsible for attempting to transform the lives of people who need it so desperately. It is what's most noble about Lambda's mission, to me.

Re: Lambda School leaked documents show poor performance over the last two years

#69

Post by notJim » Mon, Oct 25, 2021, 6:08 PM UTC

sam0x17 wrote: post #35

sam0x17 wrote: post #24

It isn't just Lambda School. I mentor for [insert large coding bootcamp here] and I will say things have been going badly across the board for more than 50% of students. They recently had to cut a lot of staff, and completion rates seem to hover around only 50-60% when you take into account all the students they remove from this calculation when they withdraw for "personal reasons". The employment numbers I don't hav…

One huge mistake I think a lot (but not all of) the bootcamps are making is taking the easy route of simply using JavaScript to teach both frontend and backend. JavaScript is a TERRIBLE first programming language. I usually have to spend 2-3 sessions per mentee explaining the idiosyncrasies of JavaScript, where some of the OOP concepts actually come from, and how things work in other languages to put into context the…

I agree with this. It seems like they should start with python or even java to learn the basics in a relatively straightforward environment, but one that also implements a lot of constructs in a fairly standard way. They also both have well-documented and large standard libraries. I feel like they should spend a bit of time on that, and then switch to JS, which is relevant for so many jobs these days.

Re: Lambda School leaked documents show poor performance over the last two years

#70

Post by bwing » Mon, Oct 25, 2021, 6:09 PM UTC

I was at Lambda when they announced the switch from 9 to 6 months and the elimination of paid team leaders. The feedback was universally negative. In a channel for open student discussion, Austen Allred deleted a poll from Slack because of how lopsided the reaction was. He explained the deletion by saying the poll was "misleading."

Students then got a survey "explaining" why the change was actually a good thing by asking questions such as "Do you understand why companies value mentoring experience?" Not just failing to reveal the truth (these were cost-cutting measures), but not even taking the effort to come up with a convincing lie.

It was destabilizing: Austen's twitter account would read ambitious, hyperoptimistic; meanwhile, drastic changes would be made within the program with vague rationales ("after speaking with hiring managers, we've made these changes..."), and probing further simply got deflections or gaslighting surveys.

There were a ton of good people in the program, and I learned a lot there. But fundamentally there needs to be trust between institution and student when you're asking people to make this level of time and financial commitment. And at no point did I feel like Lambda at its top level prioritized student wellbeing over PR, costs, or metrics to be sold to investors.