Udacity restructures operations, lays off 20 percent of its workforce
techcrunch.com
techcrunch.com
Offerings right now:
1. If you want academic rigor only, go with edX
2. If you want academic rigor and some more approachable/practical courses, go with Coursera; or if you want to do full MS/MBA/MSA online (they have the widest selection right now comparing to edX/Udacity)
3. If you want unique practical skills in hot areas like self-driving/flying cars, deep (reinforcement) learning etc. without fully committing to rigorous practices, go with Udacity
4. If you need anything else, go with Udemy, Masterclass, School of AI and similar
I mean I can pay ~40-50 euros for Coursera and edX (and lots of them are free).
But Udacity costs thousands of euros - I could attend a brick and mortar school with a great reputation for that price here in Spain.
Furthermore, none of the MOOC providers have cracked the issue of credibility to employers, if I felt an Udacity nanodegree would be taken as seriously as a 'real' masters degree then I'd be willing to pay much more for it.
But when I'm just doing them more for self-fulfillment (as I'm sceptical employers give much credit to Coursera, Udacity etc.) then paying any more than 50 euros doesn't make sense.
And that is how it should be. Because it makes a better society for all. More highly educated people gives more productivity and prosperity for all. That it isn't your parent's wealth that decides what school you can go to decreases inequality and we get less people that feel that they never got a chance in life.
I do think that universities should invest more heavily in this mix, obviously without losing their strong standing in 1.
+. I am huge udemy fan. I love that profits get back to the creator of the content. Usually, I wait until there is a good sale then pick up a security [1] course for $10. The teachers are positive and reply back to me. I think the only way this model could work is if Udemy had a total of 5 employees and took a small fee. VCs need to get out of this business - these aren't unicorns.
Competitors to Udemy really need to do better marketing. The fee doesn't matter.
You can get all that and more if you actually went to university.
And considering it costs hundreds of thousands of dollars to get a degree at Stanford, it's not clear that it's the far superior option.
Most schools will provide students with academic scholarships. My school paid for the first year of my schooling because of my GPA and test scores.
Plus I'm pretty sure getting into Kettering University is a lot easier than Stanford. They didn't even have an admissions essay when I applied.
Udacity has always taken bold bets about the future of education, and I sincerely hope that they survive and thrive, simply because of their willingness to push the envelope.
[1] https://www.udacity.com/course/intro-to-statistics--st101
[2] https://www.udacity.com/course/intro-to-descriptive-statisti...
[3] https://www.udacity.com/course/intro-to-inferential-statisti...
Can anyone recommend a good online course for learning about lambda calculus and type theory?
I have some language design ideas, but I suspect I'm at risk of various type system blunders until I understand those basics better.
Edit: I really need it to be an online course, not just a text book. I find self-study to be particularly difficult for this topic.
I'll be forever grateful to Udacity for that but I'm sure it's difficult for them to meet their business goals. I have friends ask me how I got into web dev and I've tried to point them to Udacity but with all of their focus on Nanodegrees people think that's all they offer and balk at the high price point. Now I usually say to look for a Udemy course on sale for $10 or a course from Wes Bos. Which is a shame because I find the instruction and format of Udacity to be high quality.
1. https://venturebeat.com/2018/11/28/udacity-cuts-125-employee...
2. https://techcrunch.com/2018/09/28/online-education-unicorn-u...
The players in the industry that went after enterprises are doing well, the players that stuck to selling mostly to individual customers are having a rough go of it. Udacity definitely had some big enterprise deals with their specific nanodegrees, but it never looked like they became an enterprise sales company from the outside.
I believe most marketing departments (along with people trying to make a name for themselves) have massively cranked up the amount of free content they pump out in the last few years. It is making it terribly difficult to become big in the education space while selling to the masses.
Enterprises, on the other hand, _really_ want to lower their salary expenditures through training and will happily throw money at a company like Pluralsight to train up new employees. For what its worth, I think they're correct that training can drive down payroll costs. But the company I worked for was acquired by Pluralsight, so there's some bias.
The problem is more the overall devaluing of content prices by consumers generally. Especially absent much in the way of certification value for most people, that places a real ceiling on how much you can charge. I might pay something in the neighborhood of book pricing for an online course (or I may take one of the numerous free alternatives). But I'm unlikely to pay triple digits and I certainly won't pay thousands of dollars.
Keep in mind they had a deal with Georgia Tech's OMSCS, but something went awry there too.
It’s not hard to see why: GT said recently that through 5 years of operation the OMSCS has made $14M for the school. That’s wonderful from their perspective, but that’s a total disaster for a VC-funded unicorn company.
"Last year, Udacity generated about $90 million in revenue.
Even as Udacity slashes costs and headcount, it’s trying to expand its enterprise business, which has had recent success. Udacity now has contracts with 60 enterprise customers, including AT&T and PricewaterhouseCoopers. Airbus and Audi recently signed on, as well."
I don't know what this looks like as a business but I can imagine it's easier to sell a relatively modest number of business subscriptions than it is to beat the bushes for a lot of retail $10-50 courses.
Students, seeking some sort of way to make themselves marketable in the job market, will only obtain education if it results in some sort of valuable degree. Hence the reason people are willing to spend hundreds of thousands to obtain these degrees.
If you really want to disrupt education then you need to be directly figuring out how to change those perceptions. Start with the basics of making educational courses that would be considered the best in the world and then leverage the powerful benefits of an internet connected computer to create the best learning experience for the student. You don't need to reinvent the wheel. The way universities are teaching now is good but it is just not efficient and too costly overall. Take what works from traditional universities and then solve the weaknesses through the benefits of a computer.
When I took udacity courses I just felt that the quality wasn't fully there. They felt more like something that would be a supplemental project you would take if you were studying a traditional course. You still need a way to obtain those key foundation skills before tackling the bigger projects.
It's worth pointing out, that in many parts of Europe Master's degree is either free, or costs less than 7k. Of course, the quality cannot be easily compared.
Do we really want everything locked down and centralised or do we want a portfolio of tools, marketplaces and such adding up to an educational platform.
I'm personally not in CS so I guess that is why my views are slightly different compared to everyone else here. I work in accounting field. In terms of quality online curriculum there is basically nothing. The ones offered on coursera are basically a joke.
There is a lot of potential with interactive media using a computer that never gets applied to these online courses because of the technical limitations of teachers and related support staff. It is hard to make really good online courses.
The best you can get online is one on one video calls, and while not too much worse compared to real face to face conversations, you will be hard pressed to get live video to web scale.
From there, it descends into the absolute dumpster fire that is any sort of real discussion around less technical topics.
If it's good enough for in-class, why is this approach insufficient for online? I actually prefer this exact setup when taking college classes online. I can rewatch portions of the lectures if necessary, and there's not a bunch of people asking disruptive questions.
> There is a lot of potential with interactive media using a computer that never gets applied to these online courses because of the technical limitations of teachers and related support staff
I don't feel much of these types of things have been superior to books.
As it turned out, OCW ended up overshadowed by the various straight to consumer MOOCs.
As an example, take the University of Waterloo (by no means a top school in the world, but definitely up there in Canada). For the math faculty, in 2018 we had 16,000 applicants for undergrad math programs and admitted 1,200. That's a 7.5% acceptance rate. Beyond that, we regularly see upwards of 30% of first years fail one or more of their courses. Many of these fail two courses and are asked to withdraw.
You just don't see that with online courses. Anyone can take them and the high percentage who never complete the course are assumed to have lost interest, likely due to a lack of significant investment in the program. Brick and mortar universities signal serious commitment in time and money on the part of the student, in addition to ability.
Wait, after a single failed course?! That seems... cruel to do to people straight out of high school. I failed linear algebra 1 (sort of by accident). But here in a European country, people fail stuff regularly, only like the top 40% maybe have "clean years" every year, actually some very good students sometimes fail stuff too when I think about it.
You can fail your elective and they won't care, but that's really unusual since people choose easier courses for that.
I am pretty sure (90%) I got dropped from consideration by a FANG company because I don't have a degree.
I agree that the educational platform is really the way to innovate: community based learning, interactive platforms, and social goal setting are all possibilities that could help people leverage internet connectivity to learn. We've yet to see a truly "mobile first" educational platform, which connects experts to the knowledge gaps of the population "on demand". Maybe someday I'll fix a flat tire by taking a quick course, or connecting directly with an expert to walk me through the process!
If you can program, bake or pick locks, there's a good chance you learned online. Online learning has also been supporting the "we don't train employees" norms of the modern knowledge economy. The internet is a huge part of how people learn, it just isn't a big part of education. This is one of the reasons I think there is a lot of potential for major disruptions.
If you try tackling credentials in a college-like way, this risks leaning into old paradigms and not really changing much. You'll never complete with Harvard on prestige, but you can probably compete on everything else (from social aspects to educational ones) with enough determination.
Idk what the answers are for udacity, but overall I think the best approach is meandering discovery rather than point-and-row. Taking steps back is part of that.
The Nanodegree programs are not free, and from the courses that are included it's clear that you would gain valuable skills, but are they recognized as a qualification in industry at large? I'm sure some employers would consider them, but would enough employers?
Regardless, it's a fantastic service that they offer, and I'm grateful for it. Hopefully they can keep the ball rolling and pivot to a place that is profitable enough for them to continue.
I don’t have a solution but I think online education has not been solved. Clearly there is a lot of innovation but nobody is making it.
I completely agree with this.
>I don’t have a solution but I think online education has not been solved. Clearly there is a lot of innovation but nobody is making it.
I don't think the problem is online education, I think problem with these ed-tech companies is that they still are trying compete with the credentialing system of universities without access to all the student loan debt the latter have access to.
Gotta love the choices HS graduates can get funneled into/ reinforced by mainstream opinion today: bootcamps, nanodegrees, or loading up on student loan debt if you can edge out agaisnt the legacy/bribery admits.
I think what people are really trying to solve is a cheaper credentialing system, that get's distorted by some of the incentives involved.
Education is not a one size fits all - you cannot offer the same from MIT vs community college no matter if the course content is the same. The difficulty (proxied by the reputation of the school) of the same courses vary wildly and dramatically, as does workload, peer support, TA support, exams, and even the pace and how the final grade is curved.
By necessity they cannot grade difficultly, since they are trying to get people to pay for courses like they are dessert. By necessity they cannot offer different tiers of difficulty based on entrance exams or other tests, since they would insult their customer base (the intuitive argument being that those in most need of such nanodegrees are inversely the ones who may do worse in general).
I agree with this.
>The difficulty (proxied by the reputation of the school) of the same courses vary wildly and dramatically, as does workload, peer support, TA support, exams, and even the pace and how the final grade is curved.
This doesn't sound like education to me, but a system of schooling composed of these factors with varying weights of assumed importance that most people in a society choose/blindly accept to be the one and only equivalent to an education (and all the ills that come from this assumption).
One issue is that, especially in the context of a standardized online course, credentialing arguably runs somewhat counter to educational value. For example, organizations that offer technical certifications have to struggle with how specific to make their feedback so that someone who fails doesn't just look up the specific things they got wrong on a test and retake it. (Without materially changing their overall mastery of the material.)
>The other elephant in the room is that the rigor of their online courses.
I'm not sure that's universally true given in-depth problem sets, reading, etc. that a student dutifully works through. Of course, many will not and evaluation/feedback becomes more difficult as you move away from lecture formats and material that is amenable to computerized grading and evaluation.
But, as you suggest, most people are not really looking to spend the time they would need to in order to work through an MIT or Stanford curriculum, if indeed they are practically able to do so.
Which gets even more difficult when places like Georgia Tech are offering fully accredited Masters of Science in Comp Sci programs, delivered fully online, for a grand total of $7500 in tuition.
I don't think any sort of 'certificate' or accreditation approach was the right strategy for them. I was probably their prime candidate for paid certs -- I am a lawyer who self-studied CS for 4 years and just recently successfully switched careers to software dev. I took so many online courses -- Coursera, edX, MIT's OpenCourseWare, Stanford's online courses. In every interview I had, the interviewers were impressed by the fact I took such a variety of courses (I listed them on my resume, underneath my "real" education history.)
But not a single one ever even asked if I had obtained any sort of certificates from them. No one cared, not even for a second. The whole story you want to tell when you go this way is, "I got the practical learning and experience I needed; I didn't go for an official degree because I felt the piece of paper wasn't worth the money." The proof (for potential employers) that you succeeded is in your abilities -- specifically, your portfolio of personal projects you create and are ready to present. Taking all those online courses shows initiative, curiosity, passion for the field; a certificate of any sort from some place like Udacity means less than nothing.
Would it have mattered any more if it was accredited by some accrediting body? My intuition says no. Employers feel comfortable relying on traditional degrees because they're, well, traditional. There's the whole cultural significance of them, and of the experience of attending a four year program. But, right or wrong, the value of these degrees comes from the school name. Is it a nationally known school? Or a regionally known one? Is it a school in the city that your interviewer (or their coworkers) attended? But if you come to them with some accredited papers from Udacity, there is going to be a very wide respectability gap that will have to be crossed.
It's not an inspiring approach, but I can't help but feel like a bland subscription model is probably the best they could have done. Maybe they could have become Google's Pluralsight or something. I think there's still definitely room for 'disruption' in the online education space, but I think it's probably going to look and feel a lot more like traditional higher ed than many might have thought 5 years ago -- even to the extent that I think we'll see traditional schools leading the way, because those traditional schools come to the starting block already carrying a lot of cachet. Why pay for a piece of paper if the name at the top means nothing?
I agree (my personal experience was similar as a dropout with regards to getting a onsite software jobs, pretty easy when staying away from Big Tech™), they needed some other way, and tried to fight a battle that even traditional schools are finding themselves needing to step into (at least for CS).
>Would it have mattered any more if it was accredited by some accrediting body? My intuition says no. Employers feel comfortable relying on traditional degrees because they're, well, traditional. But, right or wrong
As someone who is constantly looking for alternatives that will fit for me for the types of problems I want to work on, this is the part that sucks to deal with. I enjoyed working/contracting for research labs in different (neuroimaging and a chemical engineering lab) fields than I had when pursing an undergraduate degree that I was only able to get because of someone I knew who could not only identify my interest, but vouch for me personally. From my experience, in the real world there is space for exceptions, but they have to be hacked around now, as there is no clean path. Maybe there is some entrepreneur out there with the right network/skills who could build an online product that could deliver that.
At this point, I can not be bothered paying to go through traditional academic routes because there is so much focus on optimizing for everything but moving a body of knowledge forward. When I see ed-tech companies try to emulate the existing system, I just have to laugh… at least until I see something different. I'm only 27, but I'd recommend to every future HS grad that wants more than the credential value of a degree and gives such a weight of 0, to just abandon all hope from getting that from universities.
One of the strangest things about the MOOC boom in 2012/13 was the fact that so much of what they were doing already existed.
Taking a pre-existing paid-for product and giving it away for free is not innovation, especially when you can only give it away for free because of external funding.
Survivor College https://www.amazon.com/dp/1593300670/ref=cm_sw_r_cp_api_i_mY...
If I could nitpick this particular point (because I agree with the rest), statistically it's true that the vast majority of online courses aren't finished.
But I think, going by my own use of these courses as well as anecdotes from others, this isn't because of a lack of willpower so much as professional working folk primarily being interested in learning a very specific solution-to-a-problem on-demand (exactly the thing Stack Overflow offers) rather than going through an entire course just because lots of other information happens to be bundled with the key information they're actually looking for.
There are of course many exceptions. I wanted to learn how hardware works so went through the entire Nand 2 Tetris course start to finish. But that was because independent of my "professional working folk" persona, I was just personally interested in the topic. But I've much more often used online lectures to get me over occasional humps in understanding specific nuggets of information and this is the use case I've heard from everyone else I know who uses online courses.
And I can't speak to every Nanodegree, but the projects I worked on in my Nanodegree were 'easy' enough that it wasn't hard for the reviewers to mark. There wasn't a ton of room for creativity. Most importantly, their feedback was pretty mixed in terms of how in-depth they went. Often, they'd just link me to posts from other real-life experts to the point where I felt like I was getting most of the 'juice' from outside of Udacity.
On the one hand, I wouldn't say it was a total waste of money. The production value of the videos was very good, the Jupyter notebooks were invaluable as practice, and the structure was passable.
On the other hand, I felt like they should've focused more on having deeper courses. It tried going into too many topics in such a short amount of time. To add, the amount of lecture material was far too small compared to a typical university course, which I was okay with since some university's posted related lecture material online.
In retrospect, it was a fine course as an introduction, but as a paid course for a Nanodegree I wouldn't have recommended it to others. I think people would be better off spending that money on books which go far deeper into whatever material they are interested in, alongside watching university video lectures on YouTube, reading personal posts by practicing experts and research papers.
At least, the React one was. When I asked questions about the errors I was getting, Richard Kalehoff —- the literal face of the course! — answered and clearly did not know what was going on. He told me things that flatly contradicted the documentation and couldn’t do much more than to say “try copying the example”.
To my shame, I paid $800 to learn from people who weren’t much farther ahead on the material, and in some aspects were behind me.
Edit: earlier comment: https://news.ycombinator.com/item?id=18596450#18601298
Link to the course where Kalehoff’s face is on the landing page: https://www.udacity.com/course/react-nanodegree--nd019
You can still take the classes on your own but have no "proof" (It's arguable that certification has any value for your career, but I did like the "I unlocked an achievement!" feeling..)
Also, most of the courses are comparable on other MOOC sites like Udemy and Coursera. Udemy is best you can get in $10 and it helps a lot.
I recently took AI course on Udemy for just $7 in sale and It is hands on course and I am really satisfied with it and the content. Even if it is not good I have not lost too much money. The value I got from that course far better than I would have got from udacity 100K ND courses.
Their Nano Degrees does not provide value in the industry and you can not expect a job after it. After spending 2term fee of college I expect something from the course but it way too expensive.
It's all hype you can do actually really good and get all the stuff and knowledge online. At the end for any job your knowledge will count.
An unemployed PHP web developer struggled for 2 years to get a job. She had taken time off to have children. In that time PHP had acquired pariah status.
I suggested that she do a Udacity "front end developer" course. A thousand dollars and 3 months later, she finished the course, found it valuable and obtained a job. She was promoted to project manager/leader a few months later.
Must have been a horrible experience: you put in all that efforts to learn PHP and soon you get a job that requires a different skill set. Poor woman.
It was a big vision and most startups fail, so maybe it's just they couldn't figure it out -- happens to most of us.
When you're competing with free material like the David Silver lectures you need to produce something really awesome to justify a $1k cost. I think the potential was there, and the partnership with Unity helped, but overall it felt disorganized, outsourced, and of insufficient quality (after the first section which was awesome).
Hopefully Thrun brings a renewed focus on course quality as well.
Hopefully things will improve
For example, in their Robotics Software Engineering degree, one of the topics they cover is Mapping and SLAM. That topic alone is a full university course and a specialization, i.e. Cyrill Stachniss' lectures on YouTube.
I think online courses try too hard to squeeze too much material in a short time, whereby students only acquire a superficial understanding of a topic.
Knowledge always costs some money, for cheap things you should be ware of what really missing.
Gotta hit ESC just after it loads.
While Udacity helped kick-start many similar online self-learning business models, but ultimately I believe they failed in their scaling effort.
Their nanodegree programs often feel a little overpriced. I remember their mission was to democratize online education, but at those prices I'm not so sure. I tried it once and their support really left a lot to be desired.
It's sad to see that they weren't able to figure things out.
Beside VmWare and Akamai, I haven't seen any company founded by university faculty that have survived and made impact (even though VMware was tiny compared to what it is today when EMC acquired them)
Bose, Broadcom, Qualcomm, Financial Engines, Genentech, MathWorks, Boston Dynamics, Pixar (I think it counts), Arista Networks, MIPS, Rambus, Duolingo, Sun Microsystems. Depending on what your cut off is for made an impact, I'm sure there are dozens more from Stanford alone.
Gilber Strang's linear algebra and Stewart's calcus is the other two.