I was a Udemy instructor for ~10 years selling tech courses but focused more on delivering courses through my own site for the last ~5-6 years.
Something never felt right with how Udemy promoted courses. I used to have a top selling course there, selling thousands of copies a month and now it gets basically no sales but it's still one of the highest rated courses in that niche on their platform. It's just no longer ranked or promoted by Udemy, for years.
I have no evidence of this but my personal opinion is their ranking is probably not fully automated and they have special offers and deals with certain instructors and if you're not a part of this club, oh well.
Again, it's all speculation but I can only go by what my numbers are. They were small scale life changing and now nothing but the quality of the courses I produced didn't change. It doesn't make sense. Of course it could be one big coincidence too, but this has been tracked and analyzed over years.
Modern society totally devalues anything considered even slightly old. I used to notice it as a real lack of intergenerational knowledge transfer, but it's gotten so bad that it seems like more and more people react with "how do you know so much?" and "why would you do that?" over very basic knowledge that isn't even that old. For all the reading the average person claims to do, they sure don't seem to know very much outside of a 10-year window unless they happen to have studied history in college or whatever.
But I don't necessarily blame said people, at least in the proximal sense. The technological industrial complex continuously refines its understanding of the desire for novelty that's always been there and seeks to exploit it; and they've gotten unreasonably good at that. It doesn't matter if your intellectual property is just as relevant as ever, perhaps more so, if there's some hip new alternative. Udemy and of course social media sites know this, and I think there's a feedback loop that goes beyond mere exploitation of the human psyche, but in the actual training of the human psyche to have blindness towards the past.
The only answer right now, besides hosting your own courses (with hookers and blackjack), might be to periodically recreate your online presence from scratch in order to exploit the algorithm back. If your courses on Udemy aren't seeing the traffic they deserve, close your account, and create a new one... assuming that's feasible and they don't check too hard. With the current state of AI, this may just be a cat and mouse game that can't be sustained.
I also started teaching on Udemy in 2019 and even though the number of students was high, I quickly noticed that income was low and most enrolled students did not even start the courses they purchased (let alone complete them). I also decided to invest time and money in my own website/school and that was probably the best decision I've ever made.
Also, I'm not sure most people know that Udemy was never profitable up until 2025. Before going public, Udemy had never been profitable despite good revenue growth. As of mid-2021 (around its IPO filing), the company had accumulated significant losses (hundreds of millions of dollars) and explicitly noted it had not generated a profit in its SEC filing. After its October 2021 IPO, Udemy continued to report net losses most quarters and years, even as revenue grew. Losses persisted through 2023 and into 2024. Finally, in 2025 they saw profits for the first time since its IPO.
> Speculating, but perhaps it needs to be updated once in a while? Last modified might be a (dumb) factor.
It's a fair point. I have over time, such as updating libraries which produced new zip files and also modified lessons. It didn't move the needle for rankings, but it did update the timestamp.
Udemy never sold knowledge. They are selling that feeling of a new beginning, of a better future, of FINALLY doing (buying!!!) that course you always wanted to finish. 50% of buyers don't even finish the first video.
E-learning can be like Steam to some people. You buy the course and then it sits there. You get a dopamine hit when you buy, and you can finish the course later. Sometime.
Some people need structure. But mostly structure is a way of dragging along those who aren't soaking up learning already, who aren't naturally seeking the next problem and breaking it apart. Not everyone does this, and so structure helps as a forcing function.
There are some subjects where you need academic and theoretical grounding. Or expensive equipment. For everything else, it's best to get hands on and just throw yourself at the subject. There's nothing really stopping a motivated person.
As someone who had to drop out of school in the 2008 crisis (family trouble), I owe a good chunk of my learning to the first era of online teaching.
Those courses that were basically “we’re a top university and we let someone record the class from the back” were a literal life changer. Honestly, that was all I wanted.
Everything that came after has been substantially worse. Work is gamified, teachers spend more time building an audience than creating the product… it’s all horribly tainted by profit.
If we went back to recording lectures by the worlds best and putting it online for free with attached books and exercises, we could improve the world a lot.
> Those courses that were basically “we’re a top university and we let someone record the class from the back” were a literal life changer. Honestly, that was all I wanted.
Does anyone remember what happened to UC Berkeley? They had a lot of their courses recorded and uploaded to youtube; an absolute joy. Then, some [beep] sued them for not making the recording accessible enough, i.e. not providing captions alongside the recordings. And they had to take down all their published courses! Because if someone cannot make use of those courses, then no-one can! Such a shame! Especially considering how these days, captions can be generated automatically for anyone who really needs them.
Almost my entire education has been OJT (On the Job Training). I'm a high school dropout, with a G.E.D. I've spent my entire life, looking up the noses of folks that just assume they are better than me. Gets a bit grating, but the plus side is, is that I have to prove myself, over, and over.
Besides that annoyance, it's been excellent. Directing my own learning has been amazing. Having to prove myself, over and over, and over again, has taught me to deliver results, because no one is willing to front me anything, or give me the benefit of the doubt. Delivery is my "at rest" state, and that kind of thing is hard to teach (Play A Boy Named Sue, by Johnny Cash).
What you talk about works well for people like me (and you, from the sound of it), but a lot of folks need more structure. A lot of institutions also need that paper. There are many doors that are closed to people like us.
My first formal school was a fly-by-night tech school, created to milk the GI Bill, after Vietnam. The school has long since, fallen to dust, but it was exactly what I needed, at the time. It taught me structure, troubleshooting, and problem-solving. When I left, I was ready to immediately jump into the deep end.
I like the idea of vocation-oriented post-K12 schooling, including things like union apprenticeships.
The problem is that, in the US, these aren't really supported by "The Establishment," so we tend to get rather dodgy outfits (like the one I attended).
I have heard that German University is highly vocation-oriented. I've been impressed by many of the Germans with whom I've worked. I feel that they are extremely results-driven. That may be because of the particular company that I worked for, and the types of engineers that our field attracts, though.
Tertiary education in Germany is pretty much a continuum. On one end, you have "Berufsakademien" that offer bachelor's degrees with integrated vocational training – you need to find the employer yourselves, they pay your fees and some salary that you can live off (more or less). On the other end, there's Ivory Tower Academia where no one cares about what happens outside of the classroom. And there is everything in between those two poles.
> If we went back to recording lectures by the worlds best and putting it online for free with attached books and exercises, we could improve the world a lot.
I learn well this way. You learn well this way. However, the big revelation from the early experiments with online courses and MOOCs is that most people don’t.
Fundamentals of math, history, physics, and other core topics aren’t changing except maybe for some context on current applications (e.g. how math applies to machine learning, how historical context relates to current events). Those same online course recordings you watched are still valid. There is some room for improvement with new recordings with new gear and better audio, but it’s marginal.
Once those courses are recorded and released, we don’t need to keep doing it every year over and over again. The material is out there, it’s just not popular to self-learn at a self-directed pace.
You need forced exercises, you need grading, you need something of a schedule. Not to say people can't do it. But, especially for difficult material, even a lot of motivated people won't.
I mean I remember what undergrad (and grad school) was like and I'm pretty sure doing that independently and optionally would be tough.
Just look at success for adults learning the things adults commonly take up: languages or musical instruments. There is some good results, but mostly you find people who can say "Hello, how are you" and no more; or they can play some simple songs but nothing major. It takes hours and hours of practice / study time to learn anything hard. You can find a lot of community ed classes that will bring you to about that level in many subjects, but it is hard to find anyone willing to put in the hours needed to learn something to more than a surface level.
There are a lot of activities that you can get the basics of pretty quickly given some natural abilities/talents.
But most adults won't have the time or inclination to send hundreds of hours (and probably money) on often rather boring exercises to reach the next level of an activity like playing an instrument.
> If we went back to recording lectures by the worlds best and putting it online for free with attached books and exercises, we could improve the world a lot.
Maybe improving the world is a good idea regardless of whether it's a viable model?
Anyway, in this case, the cost to the university is quite low and there's no real loss of income, as the real value the vast majority of people pay for is clearly in the status that comes from being there in person and getting the diploma.
Both have garbage content at this point - Coursera was great when they launched, top quality material and university-level instruction. Now it's just bottom of the barrel scraps.
YT has tons of quality instruction - hell nowadays I just ask an LLM to make me a course for whatever I wanna learn.
I tried that out in my field of expertise, to calibrate my expectations. ChatGPT invented multiple references to non-existent but plausibly-titled papers written by me.
I think of that when asking questions about areas I don’t know.
That was about 18mo ago, so maybe this kind of hallucination is under control these days.
I try asking LLMs about a topic I know about once every few months for the past few years, and every time I have tried I got back complete nonsense presented as factual. When pressed, the LLM will happily generate 3 entirely contradictory nonsense answers to the same factual question.
Even basic questions about moderately obscure topics (that could easily be answered off the top of their head by a typical graduate student in the field or figured out from books with ten minutes of effort by a novice with good information search skills) get garbage answers from LLMs.
My impression is that LLMs are great at quickly generating plausibly human-generated writing, something like video game character dialog, personalized spam or disinformation, or high school English papers. But you shouldn't use them to learn factual information.
Instead, learn to navigate the academic literature and search in corpora of scanned books.
I would use an agent (Codex) for this task: use the Pro model in ChatGPT for deep research and to assemble the information and citations, then have Codex systematically go through the citations with a task list to web search and verify or correct each. Codex can be used like a test suite.
My biggest issue with Udemy courses is that it's not easy to vet the instructor. User ratings are unreliable since beginners aren't really in a position to evaluate a teacher's expertise.
If Udemy's pitch were “Learn X as Taught by Notable People in the Field,” I would have signed up in a heartbeat.
Masterclass is really a scam in my opinion. Who needs them to teach about generic stuff, when what we need is how to be good like them, that’s why we pay for the course
Notable people tend to teach on their own sites, or at least more specialized sites rather than generic sites like Udemy. Udemy would need to pay them instead.
I guess it depends on what you ask an LLM to teach you. For certain subjects, I've found them to be a pain in the ass to get right.
For instance, I was hoping that I could use GPT to help me learn to fly a B737-800. This is actually less challenging than people think... if you just want to get in the air and skip all proper procedure and safety checks! If you want to fly a commercial plane like a real pilot, there is a ton of procedure and instruments to understand. There is actually quite a bit of material on this available online via flight crew operations manuals, as well as an old (but still relevant) manual straight from Boeing. So why rely on GPT? It's a bit hard to explain without rambling, but those manuals are designed for pilots with a lot of prior knowledge, not some goofball with X-Plane and a joystick. It would be nice to distill that information down for someone who just wants an idiot's guide to preflight procedure, setting the flight computer, taxiing, taking off, and performing an ILS landing.
Sadly, it turned out I really had to hold the LLM's hand along the way, even when I provided it two PDFs of everything it needed to know, because it would skip many steps and get them out of order, or not be able to correctly specify where a particular instrument or switch was located. It was almost a waste of time, and I actually still have more to do because it's that inefficient.
That said, I still think LLMs can be unreasonably good for learning about very specific subjects so long as you don't blindly believe it. I kinda hate how I have to say that, but I see people all the time believing anything Grok says. :facepalm: GPT has been a big help in learning things about finance, chemistry, and electronics. Not sure I would assume it could create a full blown course, but who knows. I bet it'd be pretty solid at coming up with exam questions.
Don't ask them to teach you, ask them to make a self-study syllabus/roadmap with online references. It's likely that it ingested the work of others in exactly this scenario, so it shouldn't confabulate as easily.
You should have a mental model about how the world works and the fundamental rules of the context where you're operating. Even though you might not know something, you eventually develop an intuition of what makes sense and what doesn't. And yes, that applies even to "university lectures" since a lot of professors make mistakes/are wrong plenty of times.
Taking an LLM's output at face value would be dumb, yes. But it would be equally dumb to take only what's written on a book at face value, or a YouTube video, or anyone you listen to. You have to dig in, you have to do the homework.
LLMs make it much easier for you to do this homework. Sure, they still make mistakes, but they get you 90% of the way in minutes(!) and almost for free.
I don't think it's (necessarily) equally dumb. Maybe if comparing LLM output to a book chosen at random. But I would feel much safer taking a passage from Knuth at face value than a comparable LLM passage on algorithms.
Nah they have definitely reduced massively. I suspect that's just because as models get more powerful their answers are just more likely to be true rather than hallucinations.
I don't think anyone has found any new techniques to prevent them. But maybe we don't need that anyway if models just get so good that they naturally don't hallucinate much.
I don't know much about Coursera, but Udemy has always been quite bad since I remember.
Most drawing/painting courses are taught from people who are juniors at best. The quality is laughable compared to what you can get for free from Marco Bucci/Sinix/Proko channels. And honestly, even those high-quality videos won't teach you how to draw anyway.
That being said, I didn't realize how bad Udemy art courses were when I got started. I think that's a life lesson for me especially in the era of LLM.
It’s remarkable to me that a major new competitor in online distributed learning hasn’t already happened, considering the obvious LLM application.
But this press release makes me sad. At one point both of these companies had big visions for how online learning should happen. To read the announcement, it sounds like they’re being held hostage by a management consultant. There is so much gobbledigook and so little clarity about how to help people learn.
Yeah ditto. I don't know when it happened, but the Coursera courses I tried at first (around 2012 I think?) were very high quality -- I thought it was clearly a competitor to traditional brick and mortar.
Then a few years later, checked it out and there were thousands of courses, many clearly without as much thought or effort.
I am not as familiar with the other online schools that focus on quality (like WGU). I am surprised they have not eaten traditional schools lunches, since the actual quality of instruction is often very variable (I am a former professor, for the most part profs have little oversight in how they run classes). Market for lemons maybe?
Another aspect I am surprised at is that the big companies have not just started their own schools. UT-Dallas where I was at for a few years was basically started to help train up folks for Texas Instruments. (RAND Pardee school is kind-of an exemplar, although that is not focused on software engineering.)
I debate sometimes I shouldn't bother with hiring seniors and just train up everyone. If you have 10k software engineers does it not make sense to just have that level of training internally?
The meaningful competitor wrt. raw educational content is freely available OpenCourseware made available under CC licenses, which prevent any after-the-fact rugpull. Of course online learning also has a big service-provision and perhaps certification component, which is where specialty platforms like Coursera and Udemy may have a real advantage.
The platforms lost because they enshittified and everyone left because of it, not because YouTube existed (it already did when they started). Compare the ancient "legacy" stuff that Coursera had to the stuff it has today. Little wonder nobody actually wants what they're selling.
As another comment here said:
> Those courses that were basically “we’re a top university and we let someone record the class from the back” were a literal life changer. Honestly, that was all I wanted.
The moment they stopped doing that, everything went to shit and this is the natural end result.
> It’s remarkable to me that a major new competitor in online distributed learning hasn’t already happened, considering the obvious LLM application.
I think it would be hard to make it work, without devolving into 50% slop. As in, it would still require very substantial continuous effort by dedicated experts, to provide a high-quality offering.
There was a set of three "legacy" courses on something called Saylor Academy back in the day by an instructor named Kenneth Manning. Statics, Dynamics, and Mechanics of Materials. They were all basically just filmed classroom lectures, but the filming was done well. Great instructor.
I recently bought a course on the Spring Boot from codewithmosh. Despite spring boot being a dry subject, it was probably the best intro backend course I’d ever taken!
Coursera courses used to be good when I still had time to do courses, while udemy was very trashy low effort for my tastes. I am surprised Coursera became as bad as everyone says, I kind of refuse to believe it. But I don't have any spare time right now to study stuff
edit: omg I just looked at coursera and it's so bad!
it's all "AI this" "AI that"
who uses all that stuff? who wants that? the whole site looks so sad now. the OGs are still there but there is so much crap around it
Udemy figured out that selling to enterprise is way more profitable than individuals. Coursera figured out that University/Company brand is more valuable than Joe's Ultimate Course.
But in the last couple years both have been horribly run. Hopefully the AI threat lights a fire. I suspect a well designed course with some context engineering can become far better than ChatGPT by itself.
I think that window is closing pretty fast. Models can currently construct pretty good learning material by themselves. I setup a project using claude code as the agent that researches and constructs learning material and lessons.
The primary limitation right now is "time".. it takes time to do all the research, so it kind of has to be an async process.
Big move. This feels less like a typical merger and more like a bet that AI-driven skills platforms need real scale to matter. Curious to see whether this actually improves learning quality and outcomes—or just creates a larger marketplace with the same challenges.
Interesting development. I had assumed a private equity company was behind this, but it seems like a deal brokered between two public companies, maybe struggling to show growth.
Something tells me the outcome will likely be the same -- years of trying to get competing systems to get aligned or absorbed, attrition of all the most important people who are ready to move on and do more interesting work, and ultimately a poorer experience for the customer.
I bought quite a few courses at udemy, none at coursera though, but I ended up not taking them, instead I used youtube to get some video, and LLM to get the text context these days. Youtube is the true gem, if it spins out of google it could take on netflix at least. In short, google might be undervalued a lot just because of youtube, for entertainment and education purposes.
I get so much decision fatigue when choosing a course series on YouTube. On every technical topic, there are like 15 people making courses anywhere from 10 minutes to 10 hours.
For a basic crash course in Python, is there anything better than the top rated Udemy course, can YT offer something better ? I really don't mind paying the 12$ it costs on Udemy.
This same week, Egghead (https://egghead.io) started offering $500 lifetime access to everything they ever made or will make. There's definitely some excellent material in their catalog. But the signals sure seem to point toward the decline of centralized human-created coursework.
Egghead.io is not worth it, the courses there have a shelf life even shorter than frontendmasters. Authors mostly use it to dump their wares then never update the course afterwards while breaking API changes litter their backlog making most content, unless it was released in the last 3 months, worthless.
Absolutely not worth it since the courses are on par with random youtube tutorials IMO.
Also really dislike the pattern of some popular frontend frameworks selling basic documentation in the form of "courses."
(meta: genuinely curious why my observation was downvoted. I may take a further karma hit for mentioning it but worth it if it elicits a meaningful response. what gives?)
Remember Conglomerates? It just keeps changing name.
Free competitive markets are not an emergent natural phenomenon, they are a technology of civilised societies, and without governments constantly keeping markets free, we keep reverting back to to robber barons and eventually petty warlord kings, that's the natural low-energy state of humans if you let it go unchecked.
End of an era: video (with broadband Internet penetration) was the best tool we had for 15+ years. But LLMs are now good enough, including in image+infographic generation and factuality (especially when grounding resources are provided... which is where human experts still matter). I think video is now better only for learning physical hands-on skills... and those videos tend to be on YouTube rather than on Udemy or Coursera.
Coursera's model will still survive for a while, given people's desire for branded credentials (university degree credits or company-branded certificates)... until the university bubble bursts too in a 10+ years. Start of trend: https://www.nbcnews.com/politics/politics-news/poll-dramatic...
A bit of a plug: we tried building a consumer business, with a learning experience built atop these LLMs: https://uphop.ai/learn . Still offered for free to consumers, but we're now succeeding much better on B2B ("you either die a consumer business or live long enough to become B2B" was v true for us).
Dead end for what? Quality recordings of real lectures have been amazing for self improvement. I've never gone into it expecting a meaningful certificate, just meaning for me living my life.
You're probably in the small minority at least when it comes to forking over material dollars. Most people spending money, at least beyond trade paperback range, are probably looking for something that has at least a plausible connection to real income.
It's not clear to me that independent tutors are generally getting rich. But there clearly are activities that benefit from individual training. I imagine music is one of those. But that mostly probably falls into the luxury goods category.
"Getting rich" is a substantial movement of the goalposts. The point is that people are spending money outside the trade paperback range on education, either for themselves or for their children (I've seen quite a wide distribution while in the waiting room), that has nothing to do with income.
My alma mater (University of Nottingham UK) has just stopped all music and modern language teaching, which (for a very popular, respected, large campus institution) seems a bad sign for universities generally.
Turns out that an online certificate isn't worth anything when layoffs happen and the market is oversaturated with people who have real degrees. MOOCs have their place, but it's a very narrow set of disciplines.
I wouldn't be so tough on the online certificates. The key value I get out of Coursera is an unbeatable "time to knowledge" and some proof it was me who attended the course through the id verification.
Compare that to traditional in-person education, where you are bound to fixed course dates, long approval timelines etc. Until you get feedback from HR that you are eligible for a course/training, i've probably already completed it via my Coursera complete subscription.
I'm not sure offline certificates mean a whole lot when layoffs matter either.
But MOOCs and other purely online options just didn't result in any meaningful certification especially outside of a connection to established universities. And, given that, people/companies weren't interested in paying significant bucks for them.
It was probably a useful experiment. Just not a very successful one. And once the experiment faltered, schools/professors became less interested in putting money and energy into it.
All the evidence is that most of the students/potential students who weren't already motivated to pursuing independent learning didn't really connect to all this online material.
It's probably not a uni specific issue. I went to a top EU uni, and there absolutely were courses that could've just been an ~~email~~ video. Admitedly not everything was bad, but the quality of education isn't as high as it should be.
both have had questionable content for a while, it's a wonder people are still paying for them. especially given that LLMs exist (and youtube for that matter).
If I were a professor at a decent school, I'd probably look at the landscape of MOOCs and go "Why am I spending any time on this?" It seemed like something new and potentially exciting at one point. I certainly wouldn't today.
I honest dislike the idea of having yet another niche going with less competitors.
Coursera somehow got to be not good. I subscribed to the premium thing and it is not worth for me. I'd rather pay for the courses and do that on my time.
Udemy does that "promotion" nonsense to "encourage" to "buy it now" which I think is lame. It is not like they are steam. They are just cycling through the list and add 1000 bucks to whatever is not there. Also they store your cookies and track your device for that. That's despicable
I wish Coursera to rethink this decision and to rethink on the platform itself.
The reality is that most of these courses exist for the certificate not the course material. People still job hunting for entry level roles just want to pad out their LinkedIn.
You and I may not like it, but the reality is that so much of "education" is about certifications. And MOOC certifications were never worth much. I have taken unrequired training courses over the years and did other self-guided training, but a LOT of online training was never worth much as a formal certification thing and that's where a lot of the money goes. (Look at executive MBA programs vs. online courses.)
Udemy course ratings have always felt absurd to me.
The few times I spent a few bucks, out of curiosity, on some technical courses with near perfect scores, was horrified to find the instructor could barely speak English, audio seemed to have been recorded out an internet cafe in some 3rd world country, and explanations were shallow or confusing.
The surprise was not that a $5 to $25 course was bad. The surprise was the mismatch with the numeric rating, reviews and student testimonies, compared to actual course content. I can only imagine, most reviews are fake and the rating system has issues.
The best courses I wanted to take are split between Coursera, Udemy, and EdX. The first two can give certificates cheaply. A merger could be really helpful if they do it in a month or two. ;)
I’d love to see long term usage data on MOOCs. They had so much promise though I don’t know anyone who uses them post-LLM though it could be I live in a bubble.
It feels more like it was sort of a fad thing and, especially once any certification value essentially fell off the back of the truck (and therefore no one really willing to pay)--much less any real value delivered to people who weren't already autodidacts--it sort of faded away.
From where I was at the time Linkedin Learning (or whatever it was called) was a sometimes vaguely useful company benefit for random stuff but I'm not sure to what degree anyone even tracked who used it.
I think what killed all MOOC learning was that they ALL saw this giant TAM for corporate training and thought.. we have to get into that market.
That is what hollowed out the value.. all the incentives are inverse to building long term value.
Everything becomes check box driven product development to close the next "big deal" and then no development is done to really enhance the core of the system or the core value to the learner. It becauses now it morphs into can we show value to the clients/decision makers/learning admins?
I largely disagree. If you look at the people involved (and what they said at the time), I think there was a legitimate "We can rethink higher education" which obviously didn't happen for a variety of reasons.
It mostly morphed to corporate training and courses for people who already had Masters degrees.
LLMs could be a boost to MOOCs because you can use them as a tutor to help with the material. People tend to have trouble finishing MOOCs, and it can be frustrating to get stuck on a particular aspect without much instructor support. Anything that makes it more interactive could help with both of those. I think LLMs are a great complement to MOOCs.
I use Udemy courses all the time; great for compliance, game engine training, and insightful training of soft skills. Good instructors have insight and comprehensive coverage that questioning LLMs do not have.
All forms of education boil down to people putting in the time to engage their brain with the subject matter. Most organized education is based on coercing, peer pressure, or social pressure to get students in situations where they kind of have to engage e.g. in order to pass exams, or other exercises, or by being forced to listen to a teacher for a few hours in a class room.
Online education is not that different. You basically put in the time watching the videos and doing the homework and tests. The test and certificate become the goal.
Self study whether powered by LLMs or by good old books or whatever source of information, basically relies more on things like curiosity and discipline. Some people do this naturally.
The nice thing about LLMs is that they adapt to your curiosity and that it is easy to dumb down stuff to the point where you can understand things. Lots of people engage with LLMs this way. Some do that to feed their confirmation bias, some do it to satisfy their curiosity. Whatever the motivation, the net result is that you learn.
I think LLMs are still severely underused in education. We romanticize the engaged, wise, teacher that works their ass off to get students to see the light. But the reality is that a lot of teachers have to juggle a lot of not so interested students. Some of them aren't that great at the job to begin with. Burnouts are quite common among teachers. And there are a lot of students that fall through the cracks of the education system. I think there's some room there for creative teachers to lighten their workloads and free up more time to engage with students that need it.
I saw a teacher manually checking a students work on the train a few days ago. Nice red pen. Very old school work. She probably had dozens of such tests to review. I imagine you get quite efficient at it after a few decades. But feeding a pdf to chat GPT probably would generate a very thorough evaluation in seconds given some good criteria. She could probably cut a few hours of her day. There are all sorts of ways to leverage LLMs to help teachers or students here. Also plenty of ways for students to cheat. But there are ways to mitigate that.
watching a few good videos is a great way to FEEL like you're learning. just cuz you watch a 15 hour videoo course on c doesn't mean you can write c any more than watching a 2 hour video course on kung foo means you can kick like bruce lee
The best Elearning platform I've found is mathacademy. no videos. just short texts on how to solve a problem and then a bunch of problems with increasing difficulty. much more efficent if you want to actually learn a skill.
Retention is an issue with education more generally (including the meatspace variety) but spaced-repetition systems (SRS) address it quite well. With online video, you can even prompt an LLM to provide a suggested distillation of the content into Q&A flashcards.
One of the challenges is that few people are genuinely interested in a comprehensive view of a topic. Most of the time, I want just enough to get to the next step and get rid of a problem.
I never wish to learn about Docker. I want to know enough to get my containers running. In a pre-LLM world, I did take a course on Docker. I have learned my last bit of Docker in an LLM world.
Something never felt right with how Udemy promoted courses. I used to have a top selling course there, selling thousands of copies a month and now it gets basically no sales but it's still one of the highest rated courses in that niche on their platform. It's just no longer ranked or promoted by Udemy, for years.
I have no evidence of this but my personal opinion is their ranking is probably not fully automated and they have special offers and deals with certain instructors and if you're not a part of this club, oh well.
Again, it's all speculation but I can only go by what my numbers are. They were small scale life changing and now nothing but the quality of the courses I produced didn't change. It doesn't make sense. Of course it could be one big coincidence too, but this has been tracked and analyzed over years.
But I don't necessarily blame said people, at least in the proximal sense. The technological industrial complex continuously refines its understanding of the desire for novelty that's always been there and seeks to exploit it; and they've gotten unreasonably good at that. It doesn't matter if your intellectual property is just as relevant as ever, perhaps more so, if there's some hip new alternative. Udemy and of course social media sites know this, and I think there's a feedback loop that goes beyond mere exploitation of the human psyche, but in the actual training of the human psyche to have blindness towards the past.
The only answer right now, besides hosting your own courses (with hookers and blackjack), might be to periodically recreate your online presence from scratch in order to exploit the algorithm back. If your courses on Udemy aren't seeing the traffic they deserve, close your account, and create a new one... assuming that's feasible and they don't check too hard. With the current state of AI, this may just be a cat and mouse game that can't be sustained.
It's a fair point. I have over time, such as updating libraries which produced new zip files and also modified lessons. It didn't move the needle for rankings, but it did update the timestamp.
I bought a handful before realizing what was happening. I haven't done it since then and I definitely need to consciously override any temptations.
I'm sure some people are disciplined enough to learn from it, but there's no way that's the norm.
E-learning can be like Steam to some people. You buy the course and then it sits there. You get a dopamine hit when you buy, and you can finish the course later. Sometime.
Some people need structure. But mostly structure is a way of dragging along those who aren't soaking up learning already, who aren't naturally seeking the next problem and breaking it apart. Not everyone does this, and so structure helps as a forcing function.
There are some subjects where you need academic and theoretical grounding. Or expensive equipment. For everything else, it's best to get hands on and just throw yourself at the subject. There's nothing really stopping a motivated person.
Those courses that were basically “we’re a top university and we let someone record the class from the back” were a literal life changer. Honestly, that was all I wanted.
Everything that came after has been substantially worse. Work is gamified, teachers spend more time building an audience than creating the product… it’s all horribly tainted by profit.
If we went back to recording lectures by the worlds best and putting it online for free with attached books and exercises, we could improve the world a lot.
Does anyone remember what happened to UC Berkeley? They had a lot of their courses recorded and uploaded to youtube; an absolute joy. Then, some [beep] sued them for not making the recording accessible enough, i.e. not providing captions alongside the recordings. And they had to take down all their published courses! Because if someone cannot make use of those courses, then no-one can! Such a shame! Especially considering how these days, captions can be generated automatically for anyone who really needs them.
Besides that annoyance, it's been excellent. Directing my own learning has been amazing. Having to prove myself, over and over, and over again, has taught me to deliver results, because no one is willing to front me anything, or give me the benefit of the doubt. Delivery is my "at rest" state, and that kind of thing is hard to teach (Play A Boy Named Sue, by Johnny Cash).
What you talk about works well for people like me (and you, from the sound of it), but a lot of folks need more structure. A lot of institutions also need that paper. There are many doors that are closed to people like us.
My first formal school was a fly-by-night tech school, created to milk the GI Bill, after Vietnam. The school has long since, fallen to dust, but it was exactly what I needed, at the time. It taught me structure, troubleshooting, and problem-solving. When I left, I was ready to immediately jump into the deep end.
I like the idea of vocation-oriented post-K12 schooling, including things like union apprenticeships.
The problem is that, in the US, these aren't really supported by "The Establishment," so we tend to get rather dodgy outfits (like the one I attended).
I have heard that German University is highly vocation-oriented. I've been impressed by many of the Germans with whom I've worked. I feel that they are extremely results-driven. That may be because of the particular company that I worked for, and the types of engineers that our field attracts, though.
It's by Shel Silverstein; Johnny Cash just performed it.
I learn well this way. You learn well this way. However, the big revelation from the early experiments with online courses and MOOCs is that most people don’t.
Fundamentals of math, history, physics, and other core topics aren’t changing except maybe for some context on current applications (e.g. how math applies to machine learning, how historical context relates to current events). Those same online course recordings you watched are still valid. There is some room for improvement with new recordings with new gear and better audio, but it’s marginal.
Once those courses are recorded and released, we don’t need to keep doing it every year over and over again. The material is out there, it’s just not popular to self-learn at a self-directed pace.
I mean I remember what undergrad (and grad school) was like and I'm pretty sure doing that independently and optionally would be tough.
There are a lot of activities that you can get the basics of pretty quickly given some natural abilities/talents.
But most adults won't have the time or inclination to send hundreds of hours (and probably money) on often rather boring exercises to reach the next level of an activity like playing an instrument.
How is that a viable model?
Anyway, in this case, the cost to the university is quite low and there's no real loss of income, as the real value the vast majority of people pay for is clearly in the status that comes from being there in person and getting the diploma.
YT has tons of quality instruction - hell nowadays I just ask an LLM to make me a course for whatever I wanna learn.
I think of that when asking questions about areas I don’t know.
That was about 18mo ago, so maybe this kind of hallucination is under control these days.
Even basic questions about moderately obscure topics (that could easily be answered off the top of their head by a typical graduate student in the field or figured out from books with ten minutes of effort by a novice with good information search skills) get garbage answers from LLMs.
My impression is that LLMs are great at quickly generating plausibly human-generated writing, something like video game character dialog, personalized spam or disinformation, or high school English papers. But you shouldn't use them to learn factual information.
Instead, learn to navigate the academic literature and search in corpora of scanned books.
If Udemy's pitch were “Learn X as Taught by Notable People in the Field,” I would have signed up in a heartbeat.
- 3D Graphics taught by Michael Abrash
- Card Manipulation taught by Jeff McBride
- Pianistic Ergonomics taught by Edna Golandsky
MasterClass already is like this, but the content doesn't go as deep as it could to really teach learners.
For instance, I was hoping that I could use GPT to help me learn to fly a B737-800. This is actually less challenging than people think... if you just want to get in the air and skip all proper procedure and safety checks! If you want to fly a commercial plane like a real pilot, there is a ton of procedure and instruments to understand. There is actually quite a bit of material on this available online via flight crew operations manuals, as well as an old (but still relevant) manual straight from Boeing. So why rely on GPT? It's a bit hard to explain without rambling, but those manuals are designed for pilots with a lot of prior knowledge, not some goofball with X-Plane and a joystick. It would be nice to distill that information down for someone who just wants an idiot's guide to preflight procedure, setting the flight computer, taxiing, taking off, and performing an ILS landing.
Sadly, it turned out I really had to hold the LLM's hand along the way, even when I provided it two PDFs of everything it needed to know, because it would skip many steps and get them out of order, or not be able to correctly specify where a particular instrument or switch was located. It was almost a waste of time, and I actually still have more to do because it's that inefficient.
That said, I still think LLMs can be unreasonably good for learning about very specific subjects so long as you don't blindly believe it. I kinda hate how I have to say that, but I see people all the time believing anything Grok says. :facepalm: GPT has been a big help in learning things about finance, chemistry, and electronics. Not sure I would assume it could create a full blown course, but who knows. I bet it'd be pretty solid at coming up with exam questions.
LLMs are vastly superior to compile and spread knowledge than any other thing preceding them.
You should have a mental model about how the world works and the fundamental rules of the context where you're operating. Even though you might not know something, you eventually develop an intuition of what makes sense and what doesn't. And yes, that applies even to "university lectures" since a lot of professors make mistakes/are wrong plenty of times.
Taking an LLM's output at face value would be dumb, yes. But it would be equally dumb to take only what's written on a book at face value, or a YouTube video, or anyone you listen to. You have to dig in, you have to do the homework.
LLMs make it much easier for you to do this homework. Sure, they still make mistakes, but they get you 90% of the way in minutes(!) and almost for free.
LLMs level the playing field for the other 8 billion people.
Reminds of this article[1] that was featured yesterday and which I think was great!
1: https://news.ycombinator.com/item?id=46254794
I don't think anyone has found any new techniques to prevent them. But maybe we don't need that anyway if models just get so good that they naturally don't hallucinate much.
Most drawing/painting courses are taught from people who are juniors at best. The quality is laughable compared to what you can get for free from Marco Bucci/Sinix/Proko channels. And honestly, even those high-quality videos won't teach you how to draw anyway.
That being said, I didn't realize how bad Udemy art courses were when I got started. I think that's a life lesson for me especially in the era of LLM.
But this press release makes me sad. At one point both of these companies had big visions for how online learning should happen. To read the announcement, it sounds like they’re being held hostage by a management consultant. There is so much gobbledigook and so little clarity about how to help people learn.
These platforms lost because of YouTube…not AI.
Then a few years later, checked it out and there were thousands of courses, many clearly without as much thought or effort.
I am not as familiar with the other online schools that focus on quality (like WGU). I am surprised they have not eaten traditional schools lunches, since the actual quality of instruction is often very variable (I am a former professor, for the most part profs have little oversight in how they run classes). Market for lemons maybe?
Another aspect I am surprised at is that the big companies have not just started their own schools. UT-Dallas where I was at for a few years was basically started to help train up folks for Texas Instruments. (RAND Pardee school is kind-of an exemplar, although that is not focused on software engineering.)
I debate sometimes I shouldn't bother with hiring seniors and just train up everyone. If you have 10k software engineers does it not make sense to just have that level of training internally?
As another comment here said:
> Those courses that were basically “we’re a top university and we let someone record the class from the back” were a literal life changer. Honestly, that was all I wanted.
The moment they stopped doing that, everything went to shit and this is the natural end result.
I think it would be hard to make it work, without devolving into 50% slop. As in, it would still require very substantial continuous effort by dedicated experts, to provide a high-quality offering.
On Coursera, I did Andrew Ng’s machine learning course and Dan Boneh’s cryptography course and both were excellent. Time well spent IMHO.
The next thing I want to take is a WinDbg course. Udemy has one that looks pretty good. I should probably also find a modern assembly language course…
The instructor is really passionnate about what she's talking about, which really makes the subject more interesting than I thought it would be.
edit: omg I just looked at coursera and it's so bad!
it's all "AI this" "AI that"
who uses all that stuff? who wants that? the whole site looks so sad now. the OGs are still there but there is so much crap around it
But in the last couple years both have been horribly run. Hopefully the AI threat lights a fire. I suspect a well designed course with some context engineering can become far better than ChatGPT by itself.
The primary limitation right now is "time".. it takes time to do all the research, so it kind of has to be an async process.
Something tells me the outcome will likely be the same -- years of trying to get competing systems to get aligned or absorbed, attrition of all the most important people who are ready to move on and do more interesting work, and ultimately a poorer experience for the customer.
But what do I know.
This is what you get when you have an educator completely dedicated to a single topic and surpasses all expectations of education.
Playlists ...
Python Programming Beginner Tutorials https://www.youtube.com/playlist?list=PL-osiE80TeTskrapNbzXh...
Python OOP Tutorials - Working with Classes https://www.youtube.com/playlist?list=PL-osiE80TeTsqhIuOqKhw...
Python Tutorials https://www.youtube.com/playlist?list=PL-osiE80TeTt2d9bfVyTi...
Being able to skim, jump around, re-read a paragraph or pause on a single sentence is how understanding actually forms for me.
What’s interesting is that LLMs lean hard into this strength of text, they make it interactive, searchable, and contextual.
To me, most of these platforms have optimized video for engagement. Its essentially "press play and hope it sticks".
Absolutely not worth it since the courses are on par with random youtube tutorials IMO.
Also really dislike the pattern of some popular frontend frameworks selling basic documentation in the form of "courses."
Free competitive markets are not an emergent natural phenomenon, they are a technology of civilised societies, and without governments constantly keeping markets free, we keep reverting back to to robber barons and eventually petty warlord kings, that's the natural low-energy state of humans if you let it go unchecked.
Centuries, really, with only periodic exceptions.
Coursera's model will still survive for a while, given people's desire for branded credentials (university degree credits or company-branded certificates)... until the university bubble bursts too in a 10+ years. Start of trend: https://www.nbcnews.com/politics/politics-news/poll-dramatic...
A bit of a plug: we tried building a consumer business, with a learning experience built atop these LLMs: https://uphop.ai/learn . Still offered for free to consumers, but we're now succeeding much better on B2B ("you either die a consumer business or live long enough to become B2B" was v true for us).
But MOOCs and other purely online options just didn't result in any meaningful certification especially outside of a connection to established universities. And, given that, people/companies weren't interested in paying significant bucks for them.
It was probably a useful experiment. Just not a very successful one. And once the experiment faltered, schools/professors became less interested in putting money and energy into it.
All the evidence is that most of the students/potential students who weren't already motivated to pursuing independent learning didn't really connect to all this online material.
(Might be a problem of that university, still ...)
Coursera somehow got to be not good. I subscribed to the premium thing and it is not worth for me. I'd rather pay for the courses and do that on my time.
Udemy does that "promotion" nonsense to "encourage" to "buy it now" which I think is lame. It is not like they are steam. They are just cycling through the list and add 1000 bucks to whatever is not there. Also they store your cookies and track your device for that. That's despicable
I wish Coursera to rethink this decision and to rethink on the platform itself.
The few times I spent a few bucks, out of curiosity, on some technical courses with near perfect scores, was horrified to find the instructor could barely speak English, audio seemed to have been recorded out an internet cafe in some 3rd world country, and explanations were shallow or confusing.
The surprise was not that a $5 to $25 course was bad. The surprise was the mismatch with the numeric rating, reviews and student testimonies, compared to actual course content. I can only imagine, most reviews are fake and the rating system has issues.
https://news.ycombinator.com/item?id=39612710
I’d love to see long term usage data on MOOCs. They had so much promise though I don’t know anyone who uses them post-LLM though it could be I live in a bubble.
It feels more like it was sort of a fad thing and, especially once any certification value essentially fell off the back of the truck (and therefore no one really willing to pay)--much less any real value delivered to people who weren't already autodidacts--it sort of faded away.
From where I was at the time Linkedin Learning (or whatever it was called) was a sometimes vaguely useful company benefit for random stuff but I'm not sure to what degree anyone even tracked who used it.
That is what hollowed out the value.. all the incentives are inverse to building long term value.
Everything becomes check box driven product development to close the next "big deal" and then no development is done to really enhance the core of the system or the core value to the learner. It becauses now it morphs into can we show value to the clients/decision makers/learning admins?
It mostly morphed to corporate training and courses for people who already had Masters degrees.
Online education is not that different. You basically put in the time watching the videos and doing the homework and tests. The test and certificate become the goal.
Self study whether powered by LLMs or by good old books or whatever source of information, basically relies more on things like curiosity and discipline. Some people do this naturally.
The nice thing about LLMs is that they adapt to your curiosity and that it is easy to dumb down stuff to the point where you can understand things. Lots of people engage with LLMs this way. Some do that to feed their confirmation bias, some do it to satisfy their curiosity. Whatever the motivation, the net result is that you learn.
I think LLMs are still severely underused in education. We romanticize the engaged, wise, teacher that works their ass off to get students to see the light. But the reality is that a lot of teachers have to juggle a lot of not so interested students. Some of them aren't that great at the job to begin with. Burnouts are quite common among teachers. And there are a lot of students that fall through the cracks of the education system. I think there's some room there for creative teachers to lighten their workloads and free up more time to engage with students that need it.
I saw a teacher manually checking a students work on the train a few days ago. Nice red pen. Very old school work. She probably had dozens of such tests to review. I imagine you get quite efficient at it after a few decades. But feeding a pdf to chat GPT probably would generate a very thorough evaluation in seconds given some good criteria. She could probably cut a few hours of her day. There are all sorts of ways to leverage LLMs to help teachers or students here. Also plenty of ways for students to cheat. But there are ways to mitigate that.
The best Elearning platform I've found is mathacademy. no videos. just short texts on how to solve a problem and then a bunch of problems with increasing difficulty. much more efficent if you want to actually learn a skill.
TLDR: you learn by DOING
I guess it’s ok for compliance videos but I’m not sure about retention.
I write this as someone who wants online education succeed.
I never wish to learn about Docker. I want to know enough to get my containers running. In a pre-LLM world, I did take a course on Docker. I have learned my last bit of Docker in an LLM world.