Incentives Matter

Lambda School is a nine month remote first CS and programming school that uses income share agreements (ISAs) in place of standard tuition. An income share agreement means that tuition is deferred until after the student has graduated and gotten placed in a high paying job (more than 50k/year). The repayment is interest free, capped to 30k, and expires ten years after the student graduates. The exciting part about this is that it directly aligns the incentives of the school and the students towards the student’s success, measured explicitly by ending up in a high paying job.

I have friends that spent four years and a lot of money on an English degree to end up working a job they could have easily worked prior to the degree - why does this happen? Why are all university degrees four years no matter what the major is? Why do people spend four years on something they have little interest in to do a job they could have done anyway? Something about this is broken. I suspect some of this is just not knowing - 16 year olds told they need to go to college to be successful so they do it and pick something they can manage. What’s worse though is college has become a required signal to get hired into jobs that don’t really require it, but for economically productive work there should be a better way.

The current university system in the US is a mess of incentives. Easy access to non-defaultable federal loans means school’s tuition is able to reach extreme heights and competition for this money leads to spending on sports stadiums, expensive dorms, and other things not critical to the success of the students. Universities also get a lot of their prestige from research and the majority of their effort is typically not spent on undergraduate education. This means professors are not typically rewarded for being (or even expected to be) excellent teachers. Universities care about their students to the small extent that it helps their brand to have successful students do important things, but they get their tuition regardless of the actual outcome of their students (with a slight preference for them passing classes and not failing out).

Lambda School attacks this on multiple fronts with the reduced risk (and overall cost) of their ISAs. While ability and intelligence are evenly distributed, opportunity is not. Taking on massive student loans where the return on investment is not clear is probably a bad idea for most people, but it’s especially high risk for people who don’t come from money. Even for upper middle class families it’s a lot to pay for primarily signal where after four years of CS you probably still won’t be prepared for technical interviews. For those that didn’t go to college earlier in their life the risk is even higher - potentially taking on huge loans without a guarantee that they’ll end up in a better position afterwards and missing any income they were making during that time.

If Lambda School succeeds and a future student has the option between taking out 100k of loans to go to school for four years where success is not guaranteed or going to school for nine months and only paying if they’re employed in a high paying job, the choice will be easy and universities should be worried. How could a nine month program possibly be comparable to a four year degree? A four year computer science degree from a university must have better teachers and be more depth than a nine month course right? Well, probably not. Going back to university incentives, teaching undergraduates is a small part of their focus - a program focused explicitly on teaching and hiring the best instructors should be able to out pace them. Since universities measure success by graduation based on their own grading scheme they are at a disadvantage with how that actually compares to student success in the job market. They tend to prepare students poorly if at all for technical interviews and their knowledge of what is helpful for students to learn tends to diverge from what’s actually being done in industry.

Some schools are able to get around these limitations by being famous and being able to select extremely qualified students that are likely to succeed anywhere - this is why MIT, Stanford, etc. still provide high signal for their cost and why companies tend to focus recruitment from them. This can be overcome though since if Lambda School can fix the education piece they’ll be able to actually do what universities are supposed to do and will provide their own signal. As they succeed there’s also a lot of interesting opportunity for helping students prepare for technical interviews, negotiate salaries (since they’ll have a lot of access to market salary data), and get students referred from their alumni network. They’ll also be able to act as a talent agency and will have a major advantage since instead of just trying to pull from a small pool of already employed people they’ll be training up their own highly capable talent.

Admissions is still a hard problem though and I think probably the biggest risk for Lambda School - you need to have a good model for who is likely to succeed in the program. Currently Lambda School requires prework closely related to what you’ll be studying followed by an interview, but an effective system for this is critical for taking on the financial risk. I’m pretty optimistic this is possible though and I think the current hoops colleges require people to jump through (e.g. SATs) are generally biased towards the upper middle class and people that can afford to prepare for them. Lambda’s focus on the prework being related to the actual course work should be a better indication of success and should include more people capable of doing the work from non-standard backgrounds.

Another potential risk could be that the classes are remote, but I suspect for CS this is actually not an issue. Online courses that focus on being remote (Kaplan) tend to be pretty good and quality tools for this make a big difference. I’m skeptical of there being any benefit to sitting in a lecture hall - when I was there the students that showed up were mostly asleep or playing games on their laptop. It’s important to get this right, but I think it can be done right and the advantage of being able to do this is huge for reducing cost and spreading out the opportunity to everyone capable of doing the work.

There are a lot of people capable of doing the work the market needs that don’t currently have the opportunity to get themselves in a place where they can do it. Colleges are failing these people by being overpriced and bad at actually doing education. The potential opportunity here is shifting how everyone thinks about education - it’s huge and it creates an outcome that’s better for everyone. I really hope they succeed.