Why hasn’t technology disrupted higher education already?
slowboring.com
slowboring.com
Most of the capturable value in education is signal of acquired knowledge and other factors. I'll address the knowledge signaling first.
It has been possible for a century to self-study to the equivalent of a bachelor's degree using textbooks but it is very difficult to verify in a reasonable amount of time whether someone has done this. Thus, people use a degree as a first filter because people have a certain degree of trust that if you have a bachelor's degree, there's a minimum amount of knowledge you probably have. No one in ed-tech has built a tool that gives that confidence. Online universities have no pre-existing weight, and it is trivial to imagine hiring someone to do the work for you (this may now also be possible in university, but until that becomes a widespread notion it won't matter for this problem).
There are then other signals that traditional universities provide. Graduating from Stanford not only implies knowledge, it implies that you are some combination of affluent, disciplined, hard working, and smart due to the exclusivity of the admissions process. Ed-tech cannot presently reproduce this without "climbing the ranks" so to speak, siphoning off students from the same people with a better value proposition (which it does not currently have).
It seems then that the technical challenge is certification/assessment. Here, AI could play a role. If a machine can interactively and dynamically have a “conversation” with a person, that is randomised/unique and responsive to the student, all to assess how much they understand a topic, that’d probably do it. Combine it with a certain degree of oversight and random checking and provisions for review/appeal from “actual experts”, and you’d still have something that should scale pretty well beyond the university monopoly.
An additional factor is the importance of cultural/mainstream perception. There’s a lot of “hacking” or gaming of the university assessment system, especially if you contrast the reality with the lofty and important goals of education and edification. I’d be interested to know how many, having gone through the system, really put that much faith in the quality of the “certificate”. I’ve certainly heard stories of some people/institutions wishing to also look at high school or secondary college results because they’re less likely to be “gamed” (though I have no idea how common at all this is).
A more likely use of AI would be to expand the candidate pool by changing the nature of the jobs that are being hired for. An example is the automation of engineering calculations by high level software such as CAD.
I've gone through the system. I wouldn't hire someone based on their credentials alone. On the other hand, a person who gets through the "weeder" courses in engineering or the physical sciences at the state university, but is outright faking it, would be a black swan event, not worth planning for.
I'm not sure solitary self-learning is the only type of learning consistent with a disrupting higher education. If technology is to drive change here, I'd argue that a plurality of competing education/learning resources or services from which an individual can cater their choices to their own needs is what would be aimed for (or at least likely).
> I wouldn't hire someone based on their credentials alone. On the other hand, a person who gets through the "weeder" courses in engineering or the physical sciences at the state university, but is outright faking it, would be a black swan event, not worth planning for.
I agree. However, to my point about cultural perception, which I believe plays a large role in the state of higher education, what happens when higher education if universally seen as mostly "weeding"? Are the costs, inconveniences, cultural hegemonies/hangups/snobberies desirable and worth it?
Teachers are fully aware that a "weeder" course is a problem, and make a sincere effort to get as many students through as possible. But nonetheless it results in a lot of attrition.
State schools have a really unique potential here. There are many now that will confer an on-campus bachelors diploma for 100% remote degrees. The coursework is equally rigorous, administered by a single professor with small class sizes, and the exams are proctored. Any attempt at cheating can be easily snuffed out in this environment. MOOCs are a dead end in terms of providing this. But a state school is a respectable degree and I really hope they’ll extend this to STEM subjects soon.
IMO, this is an extreme bias in the Ed VC community and explains how they managed to flush so much cash down the toilet. In no particular order:
1. Tech focus. Most SWEs do not need a CS or even SE degree, and even if you do need that knowledge it really can be learned by hacking at things alone in a room. This ranges between "a lot less true" and "hilariously out of touch" in a lot of other fields.
2. The startup founder set skews toward the "independent learner" type.
3. Within VC/founder world, it's often actually true that university was for networking and prestige. This is one of my huge "you are so close to understanding how unreasonably BS your world is" pet peeves. Not everyone graduates from Stanford. The majority of college graduates attended colleges and universities that you've never even heard of. There are several thousand of institutions of higher education in the US. How many can you name? Sometimes I really want to say to EdTech founders/VCs: "just because you are incompetent and survive on your fancy Stanford degree despite knowing nothing doesn't meant that's true of all College degrees." (But not you in particular, of course, and this is just my inner narrative frustration... I don't mean it to be an attack on anyone in particular.)
Most students really do need years of in-person learning that does not scale well.
Most institutions provide very little signalling value.
Most of higher ed is completely misunderstood by a set of Standford grads gambling with more money than most institutions have in their endowment.
And I'm highly suspicious of counter-narratives from anyone who hasn't taught Calc II at the branch campus of a state university system.
American institutions of higher education, particularly the non-elite private ones, have been capturing value from actually teaching people useful things for two centuries. It's hard, it won't make anyone a billionaire, and we should get back to the old model of doing things not-for-profit in an unscalable way and without federal debt subsidies. Where by "capturing value" we mean "updating aging buildings and paying faculty enough to live a normal middle class life", not "supporting a massive administrative apparatus" and certainly not "buying yachts for LPs".
edit: various typos. Typed while in a meeting (about this very topic, as it happens)
Signalling is the most important thing universities provide. The vast majority of jobs that require a degree don't require or care about what you learned. They just want someone who is able to get into university and pass with good enough grades. That's the entire value. Even community colleges provide large amounts of signalling value in that you dedicated a significant portion of your life to obtaining the degree, something many people legitimately are unable to do. Startup education has no signalling power because employers don't know about it and don't believe it is difficult to obtain the degree.
1. if you aren't Stanford, trust is hard to gain, easy to lose, and nearly impossible to gain back once lost. One bad placement raises alarms and more than one poisons the well for an entire degree program. There aren't many regional employers per degree program, and national/global employers are a "reach job" for most students. This is something that non-elite institutions know in a visceral and painful way. They are so careful with the few regionally well-regarded degree programs they manage to build. The folly of VC world was obvious to those faculty, and their students (such as me), from the start.
2. if you aren't Stanford, most of your students probably really do need a couple years (at least) of in-person instruction before they are ready for the labor force. Again, anyone who has taught at one of these institutions just knows! Once every 10 years you get the type who didn't really need the degree to learn the material, and you do your best to elevate them past the institution's typical bar for outcomes. But the other 9 years, 100% of students require a lot of real work before they are ready.
In other words, most institutions cannot afford to signal without that signal being extremely accurate. AND they don't have the type of students who don't really need the education. So they educate students and then signal the value of hose students not just through grades but also especially through things like referrals, letters, and so on. But, again, I wouldn't expect a Stanford or UIUC or Michigan grad to understand this world.
Because it is. I'm not bullshitting. I talk with a lot of our top employers. They care about the actual skills and knowledge students bring. They can find plenty of people with good temperaments to promote into those roles if they only need warm bodies with good time management skills.
> I agree that most students do learn a lot from college, but it’s stuff like “how manage your time” and “how to succeed with minimal oversight” not calculus 2.
This simply isn't true for most students in most regional economies. No one is paying a premium for a piece of paper anymore.
I work in the software industry and serve on committees/boards for several schools with which I have various affiliations.
I speak with all types of employers for all types of students, in my capacity as a member of those committees/boards. Ranging from high school dropouts finishing their high school diploma later in life to masters students. My comments here are mostly about traditional undergraduate students, though.
I’ve always been aware that there are (at least) two tiers of software engineering jobs but I have a hard time articulating the reasons for that. My high school friends seem to do roughly similar work to my FAANG-ish job, but for a fraction of the pay, and it bugs me because I know they are smart and deserve more.
A few thoughts follow.
With respect to software work, I have both personal experience and mentoring experience.
First, a lot of the "type of work at FAANG vs. small shop" is extrapolation and guesswork for me. I have experience at both types of places, but in extremely different role types. I've mostly only been in "normal" roles at low-prestige firms, but I've only been in either pure R&D groups or "elite" product groups at high-prestige firms. So, for me, FAANG-style work looks incredibly different and is completely inaccessible to the average student from my alma mater at basically any point in their career, without substantial additional education. A lot of this personal anecdote is due to comparing my career before and after my (more elite-signalling) PhD. So, it's hard for me to know.
Now, in terms of mentoring, I can say some things.
First, the higher paying firms consistently get our best and brightest students. These students are the ones who only pass through our type of institution once every 5-10 years. Those employees are way more versatile, productive, disciplined, and frankly just intelligent. Are they doing to same work? Sometimes yes, sometimes no. But FAANG/high finance/et al. does seem to be choosing the absolute cream of the crop. At least from us. They are definitely buying more versatility and higher general intelligence if nothing else. Again, FROM US, a small non-elite school. I know institutional bias means that this may not be true in general.
Second, my impression is that a lot of our students would have a really hard time simply onboarding to tooling at the two FAANGs I know the most about. Think "Java programmer who uses IntelliJ and drops a JAR onto a single VM. Knows about but still isn't very comfortable with things like command lines or version control. Relies on senior engineers for anything more complicated than developing on one or two branches with rare merges and deployment to a single VM". That's a bit of an exaggeration but not far off, so you can see how landing at Google or Facebook might be a painful way for those students to start their career.
FWIW, I think the department is very pragmatics-focused. We aren't failing to teach git because we're too busy teaching automata theory, for example. We don't even teach the latter. So, very focused on software development career prep. But the quality of student and dedication to craft just isn't there. Getting students into internships gives a huge improvement but is still an uphill battle for us. Improving internship placement is therefore a major focus of the alumni's strategic efforts over the next five years. I suspect that with a couple years of full-time work and good mentorship all of our students would be ready for a normal SWE role at FAANG, but straight out of undergrad would fail at even the basics of a FAANG software job.
Personally I don’t think my mental “versatility” got too much exercise there, except maybe in rescuing production systems in emergencies. But presumably if you stick enough smart people together someone will eventually invent something novel and profitable, even if the median employee never does.
Given the cost of every minute of downtime at FAANG scale, the ability to rescue a production system a few hours faster than your competition is worth embarrassing multiples of your wage premium against non-FAANG engineers.
Much of the value of higher education as it is structured is the credentialing and the networking. Everyone knows what a bachelors is, a master’s is, a doctorate is. This is important. It means something.
However you are also right that education works in person for a lot of people and there are a lot of values there. Community colleges are an under utilized resource to learn.
Making money by educating people is a bad business. It is expensive and time consuming and the more you try to cut expenses the worse the education gets. Non-profit is the right model for it.
If you really want to make money off the back of education build something useful for colleges and universities. They are big complex organizations with specialized needs where off the shelf solutions don’t cut it and there are enough of them to make a nice business out of if you come up with a solution that is dramatically better than what is out there. A lot of the technology solutions that are used barely work but there are few options.
The only one that isn't public and sparten is not rich amd not elite. All are way more lean than most businesses (<10% of staff are non-teaching, and almost all of those are feeding students or keeping buildings standing/clean/powered... I can name every employee on the Administration floor off the top of my head).
If you really want to make money off the back of education... don't. Build your thing and give it away at cost.
But I do understand why someone who only spent time in places like my PhD school would feel differently and view these places as just another industry. I just don't think they are ready for the fact that 90+% of the market competition is way more lean, way more low-profit, and way more human-capital-intensive than any VC is interested in.
When I talk about making money I am talking about system utilized by colleges like enrollment CRMs, grading software etc. In most cases the only way those get made and are supported is through for-profit businesses.
In some countries private universities even get a bad reputation, with exception of a few renowed ones, as their graduates are seen by the employers as having bought their degree, even if they battled throught it as everyone else, e.g. in Portugal.
A bachelor's is a glorified dropout. A master's is the primary academic degree. A doctorate is more like a title rather than a degree, and if you do it full-time early in your career, it signifies that you can advance from junior to intermediate positions.
Or they could mean something else, depending on which traditions you are used to.
A bachelor's is an extension of high school. You show you can assimilate knowledge at a professional level.
A masters shows you can do extensive research of existing knowledge in a narrow field.
A doctorate shows you can perform novel research and add knowledge to a field.
...but perhaps that is not really applicable to all fields of study. In any case, "glorified dropout" is imho too harsh, as it implies a failure. For most people, the economic tradeoff in getting a masters degree or higher (both in terms of tuition and time lost not working) simply makes it a bad choice.
Master's degree was the primary undergraduate degree. Bachelor's degree existed on paper, but it wasn't recognized in the society. People often got their bachelor's and master's on the same day.
At some point, politicians started promoting bachelor's degrees to get people graduate faster. The academia accepted that, but the job market didn't, at least not initially. Because people were not familiar with bachelor's degrees, they often didn't consider them serious degrees. Some people knew them but took them as a negative sign. The idea was that if a university degree is a sign that you can achieve something, leaving the university after bachelor's is sign that you failed at that.
I left Finland a long time ago, and I'm not familiar with the job market relevance of a bachelor's degree today.
> I'm not familiar with the job market relevance of a bachelor's degree today.
It's OK to say nothing if you have no information to add
A bachelor's degree from the UK is not at all like one from the US, and the differences between PhDs are even bigger. Many European countries consider master's the primary undergraduate degree. A master's may be a research degree, a taught academic degree with/without a thesis, or a professional degree. Law/medical degrees may be graduate or undergraduate degrees. A bachelor's degree may or may not be a prerequisite for a master's degree, and a master's degree may or may not be a prerequisite for a doctorate. There may be intermediate degrees between a bachelor's and a master's and between a master's and a doctorate. Not all research doctorates are PhDs. An MD is not necessarily a doctorate, and the doctorate in medicine could be an MD, a PhD, or something else. Having a PhD in medicine may imply that you are not an MD.
Whatever your experiences from the academia may be, the academia is not universally like that.
No. Read again what you wrote. It wasn't a qualified statement about different values in different places, ie "in this culture X is more like Y" it was a general claim with no background:
"A bachelor's is a glorified dropout. A master's is the primary academic degree"
I don't think this is true.
Most universities' signaling value is from the fact that they are accredited universities. They draw their signaling power for the institution of the university, maintained in part by the department of education.
While I haven't checked recently if this survived, in the 2010s several universities offered their courses online and offered certificate of completion for $50-$300 per course. These kinds of programs fizzled because even if students did learn the material the certificates did not sufficiently signal to employers that the student had learned it. A friend of mine from one of these universities did an internal study showing that of those students who completed the equivalent of a bachelor's degree almost none found employment in their field. This result should be surprising if we think low tier universities have no signaling power.
On a more conciliatory note; I generally agree VC ed-tech has been a boondoggle for similar reasons - a combination of misunderstanding education and realizing that no other part of education is monetizable enough to produce the kind of returns VCs want (barring certain extra-ordinary tech innovations that have not been forthcoming e.g. strong AI)
To me, it's not at all surprising that "got the [equivalent of a] degree but didn't have the normal level of interaction with faculty" results in a vastly different outcome.
You're assuming that the credential itself is the signal. For regional economies in the US who hire primarily from regional universities, the credential often isn't the primary signal. They'll ask for a reference and actually talk with a faculty member listed.
Colleges and especially individual faculty have ties to major regional employers. Those employers understand that output is more variable (in large part because input in more variable). Other employers don't know anything about the institution at all and are likely to bin the app for any role that isn't eg low-pay or pure commission.
So the institution maintains relationships with alumni and others making hiring decisions at their "strategic" new grad employers, mostly regional small and medium sized companies. Frank assessments from trusted faculty play an important role in securing first placements. This is particularly true for (low-)paid internships, which are often the primary feeder for new grad positions.
If I know that the institution has mixed output and my phone call to the professor listed as a reference is basically "they completed some online modules but we didn't interact much", then I will probably pass.
Universities have pre-existing weight and "old money" recognition, but there's nothing inherently more "reliable" about them than online. A lot of people cheat in college, a lot of people take courses with inflated grades or courses which teach old concepts, and then have to learn everything again when they get their first job. It's absolutely true that if someone has a degree from a reputable college they very likely know their field, but if someone has an degree from a "reputable" online college then they also very likely know their field.
It's not fair that in order to prove "acquired knowledge" you need to pay over $10K per semester and travel to an expensive college. It would be nice if we start to see people from online universities succeed in the workforce and build great things, then online degrees would likely be accepted more.
If your claim is that this style of studying isn't preparing students for the real world where the problems cannot be easily predicted then I definitely agree. But cramming isn't cheating, students are just trying to succeed within the system.
With the exception of a small percent of the population this isn’t true. Most people don’t have the ability, patience, wherewithal to self study a topic in depth. HN comments are littered with examples of educated people in one area making foolish comments in other areas with those people thinking they are actually making insightful comments. Without training most peoples’ capacity to understand why what they think they know is wrong when it is pointed out to them is quite low. The classroom is much better at transmitting how to think about a subject than reading a book or even watching a video lecture.
https://www.brookings.edu/blog/brown-center-chalkboard/2021/...
Parent is talking about textbooks, but it doesn't have to be limited to that and you'd still have the "tells us in three sentences how educated you are" problem.
I also think it's an interesting "problem", if you'd try to evaluate how knowledgeable someone would be after listening to economic audiobooks for a decade for instance.
I assume it is more likely than not that this person is a crank with superficial knowledge and is dangerous when given a stack of cash or access to policy levers. Even for highly literate people, the crucible of feedback is very important. That is something that's often missed by people who don't need as much external feedback to learn. Literacy isn't enough.
Is that how you meant it?
I suspect, in either case, that this is a great Rorschach test of something or other.
But then, maybe that's also the point. I also wonder: how many people listen to endless audiobooks on macroeconomics or fiscal policy? How many people listen to endless audiobooks on tax law or corporate accounting?
The CS equivalent might be the number of people who consume endless quantities of data science fluff content on today's-ChatGPT-stable-diffusion-DRL-etc. vs the number of people who spend the same amount of time watching videos on CUDA programming, data cleaning, how to design annotation projects, and so on.
I don't disagree, but didn't feel the nuance added meaningful information to my post.
Most of these people also are unlikely to complete an online course - ed tech has a hard time reproducing the "social pressure" aspect of education. It's worth noting that there's also a larger population of people who've achieved a bachelor's level of competence through study + work e.g. my experience is that it's not that rare for certain types of technicians to study to the level of an Electrical Engineering degree and no real effort is made to recognize that expertise.
I misread what you wrote here. I thought your point was that the value of education was merely as a signal for knowledge rather than in the acquisition of knowledge. Rereading your post I see what you were getting at is the capturable value from an ed tech/ed business perspective. Sorry about the misreading.
Textbooks and/or (current) edtech are not going to be effective there.
But in hard scientific topics (mathematics, algorithms, engineering) there are facts and ideas that are taught, and it's equally easy or hard to learn from prose as it is from listening. In my experience you can be very effective self-studying as long as you already have basic background about the field and what's the point. I skipped 2-3 years of math in school by simply reading the textbook over the summer, doing the exercises, and just asking to skip to the next class. You do this a few times and then you realize that in-person classes are also just them reading the textbook to you. I don't regret it whatsoever.
I'd argue the difference is how mathematized a field is, probably along the same lines as SMBC:
a(b + c) = ab + ac
It’s easy for me to convince someone that because of this property we get the following:
3(x + 2) = 3x + 6
It’s much harder to convince someone that
3x + 6 = 3(x + 2)
It’s hard to convince someone that factoring is just using the distributive property reading right to left instead of left to right. It’s much harder to convince them of this just from reading it in a book. The nuance will be lost. They need examples that one goes over in a classroom to see this and a book can’t contain enough examples without being too many pages or without boring the reader with what appears to them to be minutia. They can’t grasp the nuance from reading. They grasp it from doing problems and being guided on the problems at the time they do them.
Books and videos can not replace a teacher. Your experience is not the norm and you should not use it as a guide for what is realistically possible for most people.
I agree with the point of your example, but maybe not the choice of example.
For your example, Gallian (which imo is the standard intro algebra book) includes this property in the definition of a ring: "Property 6. a(b + c) = ab + ac and (b + c)a = ba + bc". (Probably in order to disclaim the confusion you're talking about.) Then he goes on to derive 6-7 other properties of groups that will (of course) be useful in the exercises.
I definitely found it's really important when self-studying to pick the right textbook. Definitely you have to be an experienced educator to write a good book that anticipates most of these things. But I agree that it's impossible for any textbook to anticipate every place someone reading it might get stuck. Let's say for a given student it happens 5-6 times in a really high quality textbook like Gallian or Rudin PoMA. They have 3 options (1) ask on mathematics.stackexchange.com and probably get an answer because it's a great community, (2) try to figure it out themselves, or (3) pay $600 and invest 3h/wk to take an algebra class that might cover the first 1/2 of the textbook in 4 months.
I think where we disagree is what are the constraints of the tradeoff between attending a class vs. reading the book. My experience has been that, if the subject is interesting enough, reading a good quality textbook is cheaper and 2-3x faster than taking a course, but at certain points it can be much more challenging. I think where it depends on which student is how much more challenging, i.e., will they be able to dig themselves out of those holes in a few minutes or a few hours. I was personally in the middle of these two extremes, but I still thought it was worth it to self study after taking 3-4 classes in the math department, then I skipped a bunch (7 semesters) of analysis/algebra/topology and came back and took 2-3 grad courses, where I didn't really understand the main goal of the subjects until I took the classes. And then I went into CS industry and never used any of it again. But I don't regret the experience; it was legitimately interesting to learn about math.
I think you have to be legitimately interested in a subject to self study it successfully. But that's true about studying serious math in general though: you have to be unrelentingly into it, or else you're just really misguided and shouldn't be there, given the high competition, poor odds for any future in math, and no practical use for any of the theory.
That's my experience with it and why I made the argument that I made.
I contend you are an outlier in the extreme. I think very few people can read a math book and get anything out of it unless they have had training. For one thing, in lower level math we lie to students all the time. We just can’t be honest about things because students at that level can’t appreciate the insane number of nuances. For example, ask a first year calculus student what “dx” means and their answer will be complete nonsense but they don’t know that. We can’t tell students what “dx” really is because it is quite beyond their level of understanding.
Thanks for the discussion. I appreciate your perspective.
The two properties you quoted are about the fact the distributivity works whether you're multiplying on the left or on the right. That's one possible left-right confusion, but I would argue most weak students believe it holds even when it doesn't, so in a sense they're too permissive in their reasoning.
However, syzarian's example is about a different left-right confusion: whether you can read an equality both forwards and backwards. I've seen this confusion in students: they will readily believe that you can distribute a factor over a sum (going forward), but be very skeptical about the act of factoring out (going backward), even though it's justified by the same equation. In this case, the students aren't permissive enough in their reasoning.
This is the kind of misconception (that equality has a "direction") that's much easier to suss out with an in-person interaction, whether it's with a teacher or other students.
I obviously agree that a human is extremely effective at explaining that "equals" is symmetric and not the same as "implies". I'm just arguing that the task also can be done in prose to similar, if inferior, effect.
I did agree however that because math has so many gotchas like this, then no textbook will ever disclaim all potential sources of confusion like the example by GGP. And so our discussion boils down to the tradeoff between the price and time investment of taking a class vs. the increased difficulty of self-studying. And how different people assign value differently to each side of the tradeoff, which I claim is the root of our disagreement.
this has been doing rounds on twitter how people do not have grasp of basic maths and it points out a good argument for people who have been taught "good basics"... like i recently asked a doctor who says they are bad at maths and they just blurted out 3. i asked why and they responded, "uh.... bodmas?"
its such a small thing really, like in your primary and secondary education, of all the classes you attend, this is just a tiny tiny topic that must've been taught but because you were taught its importance and use it as a building block early on, it just feels like second nature.
like the tables of 2-9. pretty hard for kids to remember but once they do and they are taught to HOW to use it, it really helps them
Whereas ring theory is very much specific to algebra and you're not going to use it programming, doing chemistry, or whatever else outside math. Finding motivation to learn pure math is a bit harder.
GP's example illustrates a good point. I agree with their point.
Up to that point I too thought I could just learn by reading the textbook. University (assuming a good one, mine wasn't world famous but has a good reputation nationally & in Europe) level maths is completely different level than high school level math, and having access to the professors makes a big difference (although I was often the only student coming by at their office hours if it wasn't before a test, so I guess not everyone agrees with that).
You can absolutely do this for university math using textbooks if you're motivated enough. Which I was.
Sorry for the colloquialism. "School" means university in the U.S. tech industry, but apparently not in Europe.
Edit: Reply to child comment before I log off. Haha I agree. The really valuable things I learned from math professors had nothing to do with the curriculum though. It was more like stories about all the people they saw going crazy trying to get tenure, ultimately failing and then seeking revenge against those who denied them; how they all go to conferences sometimes and drink very heavily afterwards in the nearest bar; and just little colloquialisms. Math academia was a really interesting subculture of very interesting people.
My experience was that pretty quickly professors (and TAs) provided insight and guidance I couldn’t (or at least didn’t) get from just reading the textbook. But also that this increases the further you go in your education- in a 200 student auditorium class for linear algebra or calculus the professor isn’t going to be giving you any individual attention but in my school even by the 3rd-4th semester classes were already often small (15-20 students) and we had actual conversations with the professors about the material (and beyond).
And the idea that in, say, computer science, that there is precisely one correct approach to a given problem is also completely bogus once you've moved beyond trivial questions.
Yup. Plus, it's best suited for certain types of topics, and even then it's tough.
My college had one intro-level Logic course in the Philosophy dept that was entirely self-study. Get the books, study, take a series of tests (limit 1/day), each one as many times as you liked until you passed and moved to the next test. Each test bumped you 1/3 grade (e.g., from C to C+), and you could in theory pass the whole course with an A+ in 21 days into the term, especially by studying ahead of the term.
Reality was that hardly anyone actually ever did that, and this was a competitive Ivy League school. I took the course, prepped ahead, but still not to what I'd hoped, and it took me much longer than the 21 days to get the A+. It was actually a very good course and method of learning, but perhaps one of the things I learned best is that self-paced learning, especially in a rigorous sense, is far harder than it seems before we try it.
in other words - education and what it measures (grades) may or may not correlate with a successful person.
is there a way to measure or predict if a person is or will be successful? And this isn't just for hiring, it is also part of a good feedback loop for helping people get an education.
The best predictor of being able to apply knowledge is better mastery of that knowledge and the context in which you are applying it, e.g. to be a creative mathematician you must know some things relating to the problems you're solving deeply.
> is there a way to measure or predict if a person is or will be successful
Most companies use the following signals
1. A track record of success
2. Intelligence
3. Consentiousness - the willingness to do required work even if you don't want to
4. Stress tolerance
This is somewhat unsatisfying. One would hope you'd also try to detect attitudes or ideologies that blind you to entire solution spaces or being a massive jerk and alienating people around you.
You survived the 3-5 year ordeal and got your BSc or MSc, this implies you can apply some knowledge in some way.
> in other words - education and what it measures (grades) may or may not correlate with a successful person.
The signal universities provide is not perfect, or even very strongly correlated with anyone's ability to excel in a specific domain. But it is somewhat positively correlated, which means using it as employment filter will weed out more bad than good candidates, which makes hiring cheaper and less risky. In other words: however bad the correlation is, it's already good enough to be useful.
> is there a way to measure or predict if a person is or will be successful?
I worry this is an anti-inductive problem: i.e. if you can measure the right factors and predict if a person will be successful accurately enough, you can make people successful (look at it as closed-loop control problem) - but it's most likely cheaper to just make them seem successful instead. The measure becomes a target, and so eventually will cease to become a good measure.
(In a sense, BSc and MSc titles are suffering from just that. They're no longer anywhere as strong a signal than they were 50 years ago.)
> And this isn't just for hiring, it is also part of a good feedback loop for helping people get an education.
Unfortunately, the two goals - teaching for economic prospects, and educating for a rich and successful life in a more general sense - are at odds with each other.
I challenge this. In fact I've wondered why people don't devise a test that capitalizes on this.
Some edtech company should make a test - let's call it "the Genius test" - that's so hard it would challenge Terence Tao. I mean, why not? You have the whole world of possible test takers, anyone with access to a computer. If you make the incentives good enough, you can attract a million people to take your test.
So this is your class population; now weed them out.
Include International Math Olympiad questions on there. Partial differential equations. Algorithms. Quantum mechanics questions. Every hard question you can imagine should be on this test.
Then make the reward something like a full ride to an Ivy, or a strong state school, or a group of them.
You'll probably only have a hundred people with decent scores on it; and it will be so immediately obvious that they cheated if you call them up and ask them to explain a single answer, most fraudsters won't try.
Seems like a good model to me. I'm surprise it hasn't been tried.
No thanks, anyone capable of solving even the easiest of your tests will know they're better off playing the actual lottery, and spending the rest of the time on any kind of job or social activity. You'd have to offer something truly out of this world, like guaranteed UBI for the winner and their whole family, two generations back and two generation ahead, for it to be worth the risk.
But then, with incentives that strong, you'll probably have only fraudsters getting close to the finish line. Unfortunately, the deck is already stacked against higher education here: as long as your long-term economic prospects hinge directly on your educational performance, the most motivated and effective crowd will be cheaters. If you weed out people based on performance, you'll end up selecting for cheaters, and against honest students.
It has? You are describing every centralised college admissions exam, from the Chinese NCEE ("gaokao") to the Indian JEE-A or the French concours for admission to the better universities...
That is not the point of a liberal higher education, though.
Historically (for much longer than a century) people were tutored k-12 in their own homes, with their families and neighbors, and not in these government-run public schools. You could learn whatever you wanted from books at home, and a governess and key instructors.
Still, you were sent off to college at a certain age, to be finished as either a man or a woman at a place like Harvard or a women’s college. The purpose was not principally to teach you math or anthropology that you couldn’t otherwise learn yourself (though you would learn some math and anthropology). There was another purpose
If your family was wealthy enough for you to be economically inactive as a youth, and to pay for said governess and instructors.
Higher education was for the rich, too. As you said, a different purpose, mostly social (definitely social for the wealthy young women).
All those little, often church-supported, colleges and land-grant state universities and teachers' colleges that sprang up all over the US in the late 19th and early 20th centuries? That was a revolution: higher education for the non-rich, and mostly for practical purposes.
But college education does have value for most people beyond signalling. Whether it's worth $80K-$100K (depending on your state) I think is a harder question, but I definitely think it's >>$0.
I think the reason for this is hindsight bias, we don't realize how dumb we were before college vs. after college. Maybe going to an expensive day care where you are spoonfed super basic math and sociology or whatever is not valuable. But what I do think was valuable for me when I graduated 5-6 years ago from a state U.S. school was
(1) learning to talk with other people in a professional way,
(2) dedicating yourself for 4 years to studying one thing and becoming generally knowledgeable and mature about that subject, so that very little can surprise you, and
(3) learning self control and independence, and also how to drink/use drugs without embarrassing yourself. Valuable professional skill especially in tech.
I don't think (2) can be accomplished unless you take the initiative to go beyond the basic requirements for the undergrad curriculum. At least in math/CS, which is what I studied. But in industry, that maturity comes in handy especially as you're given loosely scoped tasks and you have to determine what's possible or impossible and then convince various manager types from foreign countries who are pushing back and sometimes threatening you. And as your peers whine and are weeded out, you see more and more how valuable that maturity is.
Most people don't even believe that becoming more educated makes you "smarter". "Smart" is this CPU clock speed you naturally have and then "education" is running the cpu through various benchmarks.
School is just standardization of the benchmarks so it can't be cheap and freely available. We don't want everyone to join the guild. As a society we want joining the guild to be as prohibitively expensive as possible while pretending we wish the guild could be open and free. All a racket and bullshit for 99% of the population.
Importantly, though, the topic here is higher education, not education. Higher education refers to a type of business. And the business of higher education isn't in the business of education at all, but rather in the business of socialization. The value proposition is exposure to other, generally like-minded, people in which to engage with and form bonds with, particularly with respect to mating.
While at some level something like this very forum could disrupt that there is something missing. Tech hasn't solved for the "human touch" and until then there will be value in stepping out into the real world. And so long as there is value in existing in the real world, people will pay for that.
Colleges and college degrees are first and foremost a social signal. There is no tech substitute to being part of an elite closed group and having a name-brand degree hanging on your wall. It's like asking why can't tech make Louis Vuitton bags more accessible to the masses.
On the other hand technology has absolutely disrupted learning. Information and skills are more accessible today than any point in history. Fewer disciplines need formal structured education and credentials. Ultimately technology won't "fix" the higher education system, but simply make it less relevant in the practical world.
I wonder why one would want to "disrupt* [XYZ] in general, and education in particular. Disruption seems to me to almost always being a net negative for everyone except the upstarts who displaced the incumbents by literally breaking their market - particularly, the customers end up holding the short end of the stick.
> Ultimately technology won't "fix" the higher education system, but simply make it less relevant in the practical world.
That is how technology works, which is why I don't like the whole "you can't fix a social problem with technology" line. You absolutely can. It's usually the only way you can fix it. And most of the time, the fix doesn't address the problem directly, just makes it go away by restructuring incentives that lead to it in the first place.
Thus, my belief is that the answer to "why hasn't technology disrupted [XYZ]" is that the technology the asker has in mind barks up the wrong tree. The [XYZ] in question isn't a fixed artifact you can sculpt, but an equilibrium of economic and social incentives. It's dynamically maintained, so pushing directly against it means you'll be pushed back at. You can make a small dent somewhere, then you run out of money or will, and status quo resumes.
EDIT:
Or, if you have access to billions of dollars of VC money, you can try and do it the hard way, to detonate a metaphorical nuke in a valley to reshape it completely. Maybe you'll succeed, but more likely, you'll just starve every established company/institution to death, and then still run out of money, leaving just a completely broken [XYZ]. See my previous point about disruption not being all that positive. It's what's currently happening in food delivery and private transportation, as restaurants and taxi companies are being outcompeted by unsustainable corporations operating at a loss.
There's plenty of "customers" (students) who wind up holding the short end of the current stick, too. People who, say, wind up $50,000 in debt, and still didn't manage to get a degree. Or people who wind up $50,000 in debt and graduate, but with a degree that isn't worth that. Or even people who get a worthwhile degree, but still wind up $50,000 in debt.
> That is how technology works, which is why I don't like the whole "you can't fix a social problem with technology" line. You absolutely can. It's usually the only way you can fix it.
Yes and no. You can maybe stop your gullible grandmother from getting so many scam phone calls with SHAKEN/STIR. You can't stop her next-door neighbor from pulling her into a multi-level marketing deal, though. SHAKEN/STIR may stop (most) scam phone calls, but it doesn't fix the gullibility at all.
So outside US, and a few other countries with similar approach to high level education, there isn't nothing to disrupt.
We've figured out how to increase physical goods and enable efficient use of scarce resources that can't be infinitely replicated incredibly well, despite a huge range of different emotions, motivations, and conflicts about how those physical resources should be allocated. I'd argue that's mostly because signals related to physical goods are relatively clear, direct, and measurable.
Things like "knowledge" and "skill training" aren't as clear, direct, and measurable, and when you walk up the hierarchy of abstraction, it's a lot more fuzzy and difficult to measure what you're actually getting as a consumer, what you're delivering as a producer, and what the source of problems are, especially if there's a major time delay and variability for payoff. That's the case with modern education, and also applies to "investment payoff in quality and effectiveness of health care", "investment payoff in quality and effectiveness of living space", "investment payoff in quality and effectiveness of public outreach", etc. It makes proper negotiation very difficult and leaves a lot of room for confusion and abuse.
One possible solution to this, which seems impossible/idk how this would work in the modern era, is to try to return to a system where a whole town would work more like a single unit so as to try to reduce some of the sprawl and complexity. Earlier in America there'd be a town known for making shoes, a town known for making lightbulbs, a town known for making steel, etc. There'd be a rallying, measurable "physical thing" the town could help use to calibrate whether it's healthy/allocating resources correctly. If the town isn't working as a cohesive unit, the central "thing", like lightbulbs, would get measurably worse. High school education is bad? Fewer competent lightbulb workers, fewer lightbulbs produced, clear signal. Healthcare isn't working? More sick days, lower moral, fewer lightbulbs produced, clear signal. Everyone in the town would have a better map of the intricacies of what's working and not working in that smaller system than they could possibly have in a global/gargantuan system, and they'd have a relatively objective measuring stick to determine whether they're working effectively as a unit.
As an Engineer, I have had to explain this to countless people... a mechanic is not a mechanical engineer. An electrician, is not an electrical engineer, and so on and so on.
His other anecdote of lecturer is a straw man also, as no one has done that for centuries. The idea of structured education (college) is to force you to learn ideas you would not normally learn, and have a guide to help you understand (the professor), and have those ideas organized in a pattern that aids the greater idea (e.g. taking maths: calculus, differential equations and linear algebra are not only necessary for much of engineering, but they help you to think in ways of solving problems that engineers come across often). This is one of the key problems with self-teaching, they end up not having the vast maths background to tackle complex mathematical problems, and then end up simplifying and making mistakes. In engineering, these mistakes can be deadly. In software, these mistakes can now be deadly).
I asked the electrician wiring my house to not put the phone wires through the same holes as the A/C wires. He asked why. I said because induction will put a 60 cycle hum in the phone wires. He had no idea what that was. I fired him and did the phone wires myself.
When I had a house generator installed (long story) the electrician wanted to run the 12V battery charge line through the same conduit as the A/C. I said wouldn't inductive coupling put 120V on the battery charger? He had no idea what I was talking about. I was unable to explain it to him, and just worked with him to change the design.
Electricians know the code, the rules, can figure out the right wire size from a chart, and how to do things without getting electrocuted. Step outside of that, and they have no idea.
(The electrician tried to run cat3 or some other cheap garbage wire. He also tried to run RG59 coax, which I ripped out and replaced with RG6 quad shielded wire.)
> recent decade
This was in 1999. The wiring has survived several upgrades in digital equipment and bandwidth without change.
> which it probably does
Then it evidently is a problem, as otherwise the filter wouldn't be necessary. Besides, my phone was built in the 1950s.
P.S. I'm not an EE. I'm an ME.
Sounds like you're a tech person! I have a problem with my printer, do you think you could help?
Most people go for at least two or more of these reasons. Even if a school is not ideal in a lot of ways, there are no alternatives that provide even a fraction of these benefits.
The only way that I could see higher education disrupted is by the government. If the government created a non-military service where young adults could work for the government, get trained in a career, while fraternizing with other young adults and getting paid a low but reasonable salary, perhaps that could divert people from university. Of course this would probably be mostly students who come from low-income backgrounds, as this service would not fulfill the requisite status and network building requirements that more privileged students seek. Which, now that I think about it, is also why a lot of low-income students attend college. So perhaps this would not work as well.
[1]: Continuation is generous. Often it's a necessary supplement to a watered down, propagandist, taught-to-the-test high school education.
[2]: Especially young adults who are queer, or who do not follow the political or religious beliefs of their parents.
However, with anything that remotely smells like it's close to the technology field? Any clarification you might get, is probably 10 years out of date. At best.
I ask one of the TAs (same year as me) - "Why the fuck are we paying so much for this education. My high school teachers put in 10x more effort." He says - "You have it all wrong - it's not about the education, it's about the credential and the networking that happens."
15 years later, and I finally realize he was ABSOLUTELY FRICKIN RIGHT.
Based on the followings I noticed:
1. Some alumni are reluctant to reveal where they went to school, instead opting to say things like, "I went to school up state NY" or "one of the ivies". We used to talk about why this was the case and concluded that it's because though Cornellians are proud, they still have inferiority complex against people who went to Yale, Harvard, and Columbia. And Stanford if you're from California ;-)
2. Your story about large biochem course and how it's run. I had a roommate who was a biochem major (though in the 90s) and lamented how Cornell had these 'weed-out' courses deliberately designed to winnow out the smartest. ('Orgo' i.e. organic chemistry being on of them)
3. Use of lingos like "biochem" or "TAs". I'm sure other ivies use these, but when I heard these two works, I had a flashback to my undergrad days.
Btw, off topic, I recently visited and stayed at the Statler hotel, booked a room overlooking the Tower road and relived the old campus days. I'm not the one to regret much, but I spent too much time studying and didn't take advantage of the beautiful Ithaca while I was there. Better late than never ;-)
Every time this conversation comes up it seems so surreal: paying for branding? Signaling status? Not for learning? Not my experience.
Separately, the accreditation scheme upholds a high barrier to entry for disruptors. Specifically, a school must be accredited (gatekept by the US Ed Dept) for matriculating students to be eligible for government financial aid.
Accreditation doesn't prevent schools from adopting disruptive technologies.
There is a LOT of paperwork and hassle involved. Also years of time.
It's enough of a hassle that some new institutions have found it worth their while to buy out a bankrupt school that already has accreditation rather than starting the process from scratch.
Tell me you've never looked into tried to get a school accredited without telling me. /s
Accreditation is a multi-year effort, and institutions do not necessarily receive accreditation on their first attempt.
> Accreditation doesn't prevent schools from adopting disruptive technologies.
Disruptive technologies don't matter if the market doesn't believe your degree is worth the paper it's printed on. If your school is not accredited, students may find it difficult to transfer credits or apply to graduate programs. Additionally, students cannot use federal aid—grants or loans—for non-accredited programs.
> Is the accreditation scheme really a high barrier to entry?
Yes. Yes, it is.
Make School failed to get accreditation. I doubt Make School is worse than the worst accredited university.
Why is this an advantage?
Actually, most historians tend to trace the oldest university back to the late 1100s (or maybe up to a hundred years older), with the University of Bologna being considered the first university. This is especially true if you consider university to be less "place of learning" and more an intentional community of students and their instructors.
I predict that very soon we will have AI tutors like ChatGPT, and then entire courses will be automated in technology. We already have "100% tech" online schools with one teacher per 100+ students or no teachers at all; but with AI tutors, I expect these schools will become much more effective, much more popular, and ultimately much more prevalent. I don't believe that the traditional school system will be replaced or that any public schools will become online-only: in-person learning is still necessary for many students, as are factors like free lunch and mandatory socialization, and AI will never solve those issues. But I do believe that fully-online learning, which is already rapidly growing, will start to grow even more rapidly.
And this isn't necessarily a bad thing either. Frankly, many people's public education system sucks because of lack of resources, and I don't see any alternative way that it can improve, because of lack of resources. For the students, a good AI tutor can be much more helpful than a teacher (and much more patient). For the teachers, it may be harder to get a "traditional" general-elective teaching job, but maybe instead they can teach small focus groups in special classes, where the teaching is much more fun and interesting, but also harder to automate. The state of education in modern society is awful and this has serious consequences (see: recent elections); a low-cost, effective teaching solution would be a huge benefit to global society.
The next version of it however...
I learned way more from discussing the topic with my classmate during breaks than I ever did from the course material itself. Learning a fact is easy, deeply understanding it is not - and discussing it with your peers is the best way to discover each others' flaws in understanding.
In addition to this, people asking questions is a crucial part of the lecture. Sometimes a question is because someone was asleep during the previous lecture, but most of the times it was because a concept was not explained clearly. And in a few rare but valuable cases someone asked a question, and the lecturer's eyes lit up as someone had rediscovered an open research problem - or made a critical link between two seemingly unrelated topics.
Textbooks and videos can teach you a lot, but they do not allow for interaction. If you wish to simply get a basic understanding of a well-known topic, that might be enough. But if you want to truly master it and perhaps even contribute to advancing it yourself, it simply isn't enough.
Guess how well that turned out?
You can watch MIT lectures, go to your local library, read blog posts, listen to podcasts, read open source. It's all free. But it won't give you that scarce resource of brand affiliation on your resume no matter how hard you try.
Whatever you learn is only one of the many components of the total package.
Would be interesting to see MOOCs on daily stuff: cooking, gardening, cleaning & organizing, style, hospitality.
Or MOOCs around hobbies: amateur radio, photography, sketching, watercolor, dance, music. Here there’s potential for the tech to meet users halfway.
There are some exceptions for hobbies with a well delimited body of knowledge (eg birdwatching)
While you can certainly get it for free, having a cohort and a badge that perhaps could serve as a prerequisite could benefit all students and all teachers/schools.
Education in itself doesn't need disrupting. If anything, it needs stability.
I don't need to live with others to know who I am: I can introspect in my room and maybe use the money saved in rent to pursue something that really makes me grow as a person, music for me. The only thing that I learbed living with others is that living alone is one of the greatest luxuries.
What the future of education looks like is a lot more fractured than we can see in the past. We might think of "higher education" as just a college/university, but in reality there are already alternative forms of higher education and those will continue to be added to.
Sure MooC's seemed like scams, but some MooC's allow a huge scale of knowledge transfer. And some Cohort Based Courses as well, now. I like how Maven talks about courses as "transformational" and not just "knowledge transfer".
I'm looking forward to more and more types of higher education in the near future. Retreats, Online-Only, Hybrid, etc.
You're right, there are certain types of things that you can learn quite a lot about on the internet - this is especially true if there is no prerequisites. It's my favourite thing about the internet, and I'll never stop being grateful for it. I'm a self-taught hack, but I have a very intelligent/fortunate cousin that went to Berkeley (granted for business), and my interpretation is that it's a contacts/prestige thing. The information is mostly available elsewhere.
I'm curious to see where this discussion goes.
https://citejournal.org/volume-2/issue-3-02/seminal-articles...
Alas, I came away disappointed. Instead of aiming for a quantum leap, the author only intends to automate the existing processes.
IMHO, the only thing that is going to disrupt higher education in the US is a change in the federal student loan system. Otherwise, the money is going to keep flowing and those institutions will have little motivation to adapt to survive
edit:
Good article! Thanks for sharing
Although that sounds large, it's not in corporate terms. Charging the customer 2.3 to 3 times an employee's salary seems about standard. And universities are notorious for casual using work to an underpaid "precariate".
That suggests there isn't a whole lot of margin for technology to capture in the first place.
After which, you find that degree-awarding is a regulated activity in most countries, offered by public institutions subsidised by government loan and contribution schemes.
Tech has had a massive influence on higher education in the last ten years, but it seems to go via universities buying more systems and services rather than "University of the Unregulated Start Up" suddenly sweeping the public-controlled market for education.
A variety of fun reasons. First: it _has_ changed over time. I had recorded online lectures, grade books, course registration, etc. 20 years ago. The pandemic has sort of been a punctuated evolution for instructors, and IDK if we have any sense yet of how much will stick in post-pandemic education.
Second: the cost of education is sort of a feature. Biologists and economists use the concept of "signals" -- actions you might take to indicate your fitness as a mate, or as an employee. Getting into a highly selective school is then very valuable, and the return on education relies in part on school selectivity. If thats the case, we shouldn't expect the price to go down much over time at the elite schools; they control access to a valuable commodity that only goes up as population grows. A technology that allows a school to teach 10x the students would be disastrous for signaling.
Third: the cost of education at many (all?) state schools has been subsidized, and over the past 100y has been an a downward ratchet. Every time a recession devastates a state budget, the unis take another budget cut, but only rarely see subsequent increases. Often these cuts come at the same time the uni itself is showing record enrollments. Faced with a loss of funding and strong demand, school presidents take the action economists predict and raise prices. Which means, even as students are sent ever increasing tuition bills, the actual cost to deliver instruction is not so high, and in many cases going down. And since the return is so high on education, student take the offer they are given.
Fourth: the opportunity cost of faculty delivered education is high. You want a Nobel laureate to teach you intro level physics? Well, every course they pick up is a grant they aren't applying for. Better make it worth the department's time -- every research dollar funds more grad students and PhDs on those important "similar schools" metrics.
There's nothing equivalent to saying you graduated from an ivy school or such.
Online education is making a dent. Some prestigious universities have an online curriculum and are awarding diplomas that can't be distinguished from a on-campus diplomas but they are the minority by far.
The way I see it. Online education should be a game changer in 3rd world countries. Yet, it doesn't seem to be. It looks like it really does come down to motivation and your ability to push your way through the classes.
And YouTube is far better than bored and underpaid postdoc!
There should be a filter on HN to hide blog posts that are just technobabble musings with no actionable or definitive information. If I wanted my time wasted I'd troll op-eds in political rags. At least that crap is generic enough that I could have an argument in a bar about it. (or maybe I'm missing the point, and people in SF actually do talk about this shit in bars?)
Here's how it will happen: digitalize knowlege, I mean really digital, like how LEAN math cover nearly all undergrad math subjects in mathlib.
Make learning courses like MUD or multi-ending visual novels so learners can hae adaptive route of learning. The higher education won't be replaced until we hae complete single-source-of-truth in digital.
With a tech close enough to AGI, it might be possible to have an even better 1:1 mentoring with infinite scale.
There are clearly unknowns and probably unknowns unknowns on this path, but the potential positive effects are too important to pass.
My hunch (and it's only a hunch) is that there's something essential about in-person learning that technology, to date, hasn't captured... and that's why it hasn't disrupted higher ed.
Maybe you're the exception, but when I meet people who've learned quantum mechanics from books, they've soaked up all the cool weird stuff, but can't write an equation.
Disclosure: I don't know how I would have learned physics without classroom instruction. I was viscerally interested in it, and in math, but didn't have the self discipline to do the work without some guidance.
I was the only non physics student in the course and I remember sitting in the first lecture as the lecturer spoke a million miles an hour about Hamiltonians, Wave Functions and Fermi-Dirac Statistics. Almost everything he spoke about sounded like a completely different language to me and everyone else was nodding along to everything he was saying. I felt so lost. I somehow scraped a pass but in the process had to teach myself the prior assumed physics knowledge in order to do so. I would not recommend.
I think this absolutely misses the mark on the two most critical factors. Number 1: *exposure to and interaction with a community of peers engaged in personal improvement in the same manner*
I can read Aristotle and Mills and Kant on my own in my living room; but this is not the same as being part of a book club (or a college course) reading these texts and discussing them. Finding positions, explaining them to your peers, perhaps defending those viewpoints, listening to others do the same, challenging them on their viewpoints, etc. These are examples of a secondary layer of context and understanding of a subject that is inherently not replicable in isolated learning.
I have a friend who deadlifts over 750 lbs. Do you think he goes to Planet Fitness and hops on the elliptical for 20 minutes and then goes home? No. He goes to a gym filled with similarly powerful, driven lifters. He discusses training loads, regimens, diet, etc. with his peers. The knowledge transfer between skilled practitioners is critical to master of a craft past a certain point. Sure, almost anyone who lifts for 3 years with even moderate dedication can eventually break 3 plates. But 5 plates? 6? 7? 8?! No, that requires something more: a petri dish for cultivating improvement; a community of similarly motivated individuals.
Number 2: A subject matter expert who can react and interact with you and your peers. Having a SME to highlight what is critical, to deprioritize what is 'cruft', to dispel or correct misunderstandings/misconceptions, to push the depth of the learning/training to a more advanced level; this is a massive advantage that in-person education has over impersonal, technologically-distributed courses.
Sure, some things are easily teachabale over YouTube. One of my favorite examples: Simple mechanical work. I've used YouTube multiple times to supplement my existing knowledge & owner's manual. It's helped me flush brake lines, install oil coolers, replace a water pump, and align headlights. Work like this, that is to say, work that doesn't require a high level of expertise, is easily transferable via YouTube. This is akin to taking notes during a lecture. More complicated work, like... say, dropping an engine, replacing the crankshaft, and remounting the engine is a task that requires a higher skill/knowledge base to perform correctly. I would not trust someone to do that with purely internet-based instruction.
Lastly, I would like to say that I believe scale necessarily diminishes the quality of what is being produced. You cannot teach 4000 students online as well as you can teach 400 in person; and likewise a 400 person lecture is necessarily a lower-quality educational product than a 40 person lecture & follow up socratic seminar.
because in every other aspect tech disrupts culture and higher education continuously...
Maybe we should start reconceptualizing technology's role in modern society.
My wife and I started a microschool (in Mumbai) because we couldn’t find any for our kids that were both good and did not require a multi-hour trek to get to daily. Taking up the responsibility of educating others’ kids meant we needed to articulate and clarify very foundational concepts. What exactly is a good school? What makes it so? As parents, we began our search with the “how”. We knew that traditional standardized age-segmented, test-driven schooling isn’t great. There is tons of literature on that. I also happen to have a PhD in neuroscience and have written a book about the mind, so I also know how little we know about how we learn. From whatever little we do, though, we know this current industrialized form of schooling is bad. And yet it’s been bureaucratized and dispensed with dogmatic confidence. It speaks to our resilience that we turn out okay despite its inanity. There are alternative methods that have been around for quite some time (Waldorf/Steiner schools celebrated their centenary recently). We’ve adopted one of the alternative methods.
However, the practice of actually engaging with twenty kids for hours everyday quickly made us realize that it is not enough to change or ask the how. A set of telescoping questions quickly reveal themselves: How? What? Why? To ask how we must educate, we must ask what is education, and why is it required?
Education is learning to play the game of life well. We must learn it because the game changes and we are not born with the tools and wisdom it takes to play it. We acquire them over our lifetime, learning about the game, about ourselves. We are learning even into our middle and late ages. It’s not just numeracy, literacy, or computational skills. It is mental wellbeing. It is understanding how we fit into the fabric of society and how we can navigate it successfully and profitably, and hopefully sustainably. Do you need school or college for that? No, but it helps to have a place where you can share notes and have experienced guides who are not merely transmitting information, but actively learning and exploring. That isn’t happening in most schools or colleges though. We are stuck with the unconscious definition of learning as information transmission and education as the institutional process within which this transmission happens in a measurable, testable, certifiable manner. Systems have been built around this idea and like all systems, these are tough to dismantle.
I think Generative AI’s greatest potential lies in making the transmission of information so effortlessly easy and effective (by personalizing this transmission to suit each individual’s needs) that we’ll come to realize there is more to education, learning, and the game of life than devoting over twenty years of our lives to getting graded and branded within these transmission of information tests and systems that empower them.
If not possible, is HN the right place to put it? Sites with paywalls that can't be bypassed are not okay on HN [0] cc @dang