Someone without a formal CS education who's selling programming courses telling you that formal CS courses don't necessarily teach you programming (duh).
He has a point that if you do a Master's you might be going without the extremely useful basic algorithms and data structures courses, but that is easily avoided by choosing a program that lets you take that. For other topics that aren't strictly programming, eg AI, ML, game theory, cryptography, quantum computing, probabilistic methods, optimization,... it would take a significant amount of dedication to match the amount of understanding that a rigorous course with problem sets can give you in self study.
>For other topics that aren't strictly programming, eg AI, ML, game theory, cryptography, quantum computing, probabilistic methods, optimization,... it would take a significant amount of dedication to match the amount of understanding that a rigorous course with problem sets can give you in self study.
You mean, the same amount it would take to do it through a university program? I fundamentally don't believe the material is unavailable. If you've done a CS undergrad at a respectable university, you should have enough instinct to build a curriculum to become well versed in those subjects
I am going through the autodidact process for all of those subjects (except quantum computing) (yes, it's brutal, especially alongside a FT job). But with the amount of MIT open courseware, books, and frankly, just wiki pages available, it's been easier than doing it at the pace of a university course
Yes, following along online can give you as much information as sitting silently in lectures and recitations and completing problem sets while sitting in isolation in your dorm room; but there’s way more to be learned by thoughtfully engaging in discussions with your professors, TA’s, and classmates. In fact I’d argue the bulk of my education came from the latter.
In short: Getting A’s at MIT is simple. Challenging yourself to challenge your colleagues is difficult. Not to mention all the labs.
As a corollary, I’d go on to argue everyone who paid full tuition for online classes during the pandemic got ripped off on a massive scale.
I am someone who has a masters degree and I have taken Distributed systems, Computer Architecture and Advanced Algorithms at Bradfield. In my experience, I would rate Bradfield better than a master degree for these reasons -
1. Was more relevant and practical. It was taught by people who have worked building real things in the real world.
2. Better ROI on time spent learning. In about 8 weeks (roughly 2 hours every week) and some self study I had the confidence to apply the learnings and continue learning more. Time comes at premium to me as I am working full time and have a family to take care of
3. Cheaper than a university course. It costs 2k USD roughly for a course.
During the same time, I also tried OMSCS from GATECH. I felt I got way more from these Bradfield's courses by spending less time and money than OMSCS as well. That said, its just me and YMMV.
Well the material that the authors have collected on teachyourselfcs.com is:
1) free
2) taken from outstanding courses at Berkeley, Stonybrook (the illustrious Steven Skiena), MIT and Stanford.
As I see it, this is evidence that those schools have some good courses. Graduate systems courses are pretty great as well, and of course you can hit other areas such as computer graphics, AI/ML, etc..
For formal education, I'm a bit biased toward multi-year programs (because it takes a while to learn this stuff) and schools (like those above) with good CS courses. But there's nothing to say that the authors' unaccredited graduate program is bad - based on the above I'd expect it's probably decent, although I worry about the 15 hours x 52 weeks though because I don't think that's enough time. In my experience a good systems or engineering course takes about 200 hours.
For other topics that aren't strictly programming, eg AI, ML, game theory, cryptography, quantum computing, probabilistic methods, optimization,... it would take a significant amount of dedication to match the amount of understanding that a rigorous course with problem sets can give you in self study.
Also, you're going to have an extremely difficult time getting to work on that stuff without a degree, given that even PhDs often end up on regular business bullshit.
In my experience, it's not the degree that gets you the job, it's the research you got published in the niche you're applying for. The number of people with excellent publications in such a niche, but no degree, is awfully small.
On the Contrary: A Masters Degree is often easier than an undergrad degree. You get a lot more freedom to do whatever the fuck you want. In most Universities, Masters students can teach/do research and get tuition forgiveness.
If you want the experience of being in a college campus for a while without the bullshit involved in {undergrad, phd, jobs}, Masters degrees are the perfect middle ground.
I did a Masters degree. I didn’t do very well on some courses, really loved others, learned how to read academic papers and what research would look like if I wanted to do a PhD. But I also had enough free time to make friends, hang out and go on road trips and parties. It was a lot of fun!
Academia is running out of ways to convince people in that trap. Evolve already. I believe academia as we know it, is in stiff decline and the problems in CS are exarcerbated because of the faster reality of software.
Any degree, by the time you have it, might be deprecated or eroded to the point is pointless. If academia doesn't accelerate the pace between education/application, only wealthy people that can afford a lifestyle without income will seek them for entertainment and that's not sustainable to keep a society educated. The value proposition is just not viable.
Education has to go the tutorial way or it will continue to decline.
> Academia is running out of ways to convince people in that trap. Evolve already. I believe academia as we know it, is in stiff decline and the problems in CS are exarcerbated because of the faster reality of software.
That’s just CS you are talking about, getting into an emerging field that is very important like Material Informatics would be veeeery hard without an academic setting. Even with CS… honesty I still disagree that academia doesn’t have a purpose or value. Just reading a good research papers recently, one of my favorite recent topics being new ways to discover spy cameras using smartphones. Also some of our greatest minds have received a lot of benefits from an academic environments. I understand bachelor degrees from a lot of places are overpriced and I agree that those seem like a trap and the value might be questionable.
> Any degree, by the time you have it, might be deprecated or eroded to the point is pointless.
My degree cost me out of pocket about ~$5k for my tuition. I stayed in state and my state had a program that paid for tuition if you made above 2.3 GPA in high school. That gave me new friends, sports clubs to join, and understanding of subjects early on that I still use today like psychology and finance. Not to mention my CS degree has paid me back in dividends and I still use knowledge from it today. I have maaany friends in the same boat as me.
> Education has to go the tutorial way or it will continue to decline.
Sounds like a version of tutorial hell potentially but on the job training has its place. AI, ML, game theory, cryptography, quantum computing, probabilistic methods, the list goes on for sincere that would take a significant amount of dedication to match the amount of understanding that a rigorous course with problem sets can give you in self study.
I think you're missing the distinction between education and training.
On my Master's InfoSec degree, several students complained that the tech we were studying was really old. The professor simply responded that this is education, not training. You are expected to learn how to analyse the security of systems in general. The specific tech used is just an example.
This is pretty much the real reason I’d consider pursuing a Masters degree. Being in an academic environment for a while sounds fun, even if the resulting degree is fairly useless or redundant. I’d love a second chance to be exposed to people who could become new friends, while also learning a thing or two maybe.
It is possible to get a position as a graduate assistant or teaching assistant where the university pays for your tuition and pays your living expenses in exchange for you helping the CS department with various tasks (which can include teaching undergrad classes if you're a TA).
I did that for a year and then ended up a class short of graduating because of a scheduling mistake. Since staying a grad assistant wasn't an option, I ultimately dropped out and started working instead. Getting the GA job required me to take the GRE (I scored 170 on the math part, 165 on the verbal and 90 something percentile for both with only minimal studying for the test) and be interviewed about some undergrad CS concepts.
If you're able to do grad school in CS and you've already got permanent legal status in the US, it doesn't really make financial sense even if you're a GA or TA. You can make much more just getting a job. That said, I don't regret doing it because I probably enjoyed that year more than any of my undergrad years (this was pre-pandemic) and I didn't go into any additional debt to do it!
you can get a stipend along with tuition forgiveness, so all you're really losing is $150k+/year in opportunity cost by goofing off in academia rather than working
Lots of valid criticisms here against academia, but I'll take this time to get on a mini-soapbox about communication skills. The biggest value-add I honed over the course of my academic career is my ability to communicate technical topics. Now, that isn't to say every master's student or CS degree holder is definitely better than a bootcamp grad at tech comm, in fact the winning combo there would be a communications/english degree + bootcamp for entry-level positions.
And this isn't just making presentations and writing documentation. It's making effective use of your time in standups. It's conversations about your future relationship with your employer. It's knowing how to ask the right questions to weed out bullshit when digging through other technical documentation respectfully. It's taking a bold new idea that entered your head and figuring out which way would be the best to envangelize it at your organization.
I think in an ideal case, a master's degree in a computing discipline hits the intersection between intense training on a subfield of computing that can be difficult to break into (e.g. statistical machine learning, robotics, formal methods, cryptography) and your demonstration of your "mastery" is taking a cutting edge concept in that subfield and demonstrating mastery through effective technical communication in the form of a project/thesis.
A master's degree isn't guaranteed to make you a "better" programmer, but I would really hope that it would make you much more familiar with the field, and also teach you how to take that familiarity and leverage it to become a more effective communicator.
This post seems light on any actual data between the difference on individuals who have a CS masters and not, and instead argues against getting one on a few aux points:
* Opportunity cost
* two tweets of HM saying they don't find it useful
* cs masters aren't geared towards non-cs college grad students
* professors don't know how to handle online teaching
* programs are cash cows (mentioning two non-top-cs programs)
To address the core point about whether those with a masters from a top CS program (say top 20) vs self-taught _on average_ do know more about programming, in my experience they do! Topics like how programs work (interpreted vs compiled languages, memory management), how a program talks to the OS and underlying hardware, and even topics on AI/ML and how to do deep learning. Whether that makes them better at their job depends on what their role is - if it's making a web application or working on underlying infrastructure in C.
So I guess to provide an alternative view to the author's, there are good online programs like UT Austin and Georgia Tech where students don't have to take time off work as they can do it at night / weekends, both programs are low-cost (~$10k), and do have professors who do understand online teaching. I think it's important to make sure you pick colleges and classes geared towards learning, not go with the intention of specific "job training", and likely you'll come out ahead having done it.
What I found quite interesting were that these programs are both lower cost than the program the author was selling. Georgia Tech and UT Austin are both recognizable names as well.
I think the article has a sentiment that academia is wasted time. On the contrary. Of course it has its points, I agree.
I am a self-taught programmer who got later a CS degree. Also a self-taught entrepreneur who later (almost) got an MBA.
My experience is that these are much more valuable for people who already know how to do the stuff but missing the jargon and thoughts of a crowd of exceptionally brilliant people lived before us. I am talking about the scientists, inventors of algorithms, math, bookeeping, all the concepts. How faster you can tell complex thoughts when you just have the vocabulary to use. How do you even present a spectacular idea if you have to spend half an hour explaining something that you should have known has a name.
I never understand how some people get the hubris to think they can do better without all this. Of course if they feel they could not have used all these in their professional lifes - probably getting a degree would have been a waste of time for sure. Some of these folks of course are quite capable and can and will do big things. Also sometimes you have situations when you have to choose between your startup company and the degree. We know some successful dropouts. Still, I think there is big value here.
I've been self taught primarily in the computer networking space for a little over 10 years coming straight out of high school and I'd really really like to get more formal knowledge inside the CS field, both out of general interest but also to match the increasing scope of what I work with, but I've never been sure how to go about getting started mid career. Further education was never really something I planned for when I was younger and I barely even made it out with a high school diploma because I didn't value trying to chase education as much as I just enjoyed tinkering with computers. I was lucky to have an in to an internship via a high school robotics mentor to get my career started and now I'm in a principal role still on the upward portion of my career while also traveling a significant amount and working an irregular schedule so I'm not sure what education path is actually a worthwhile yet flexible enough. Do I just start looking up random universities with online courses? Is that even going to give me what I want or am I just going to end up qualifying for an official piece of paper instead of actually getting a useful amount of new knowledge? Do I have enough time to actually get a useful amount of new knowledge on the side?
So I haven't yet pursued a CS degree but whenever I hear of a story like yours it becomes more nagging in the back of my mind.
It appears a somewhat counter-argument is about getting high salaries and the corresponding positions without having to go into debt or at least invest the time and money into getting a degree. As we all know, student loan debt is a very real issue. Some of those people might feel they aren't missing anything, where in cases like yourself, there is a need to find out what they might be missing or a sense of wanting to be publicly viewed as a true equal in the field.
I think it should be noted that the Georgia Tech Online Master's in Computer Science (known as OMSCS) mentioned in the article costs about 7,000 USD, all-in for a 5-semester program. [1] This seems to me an amazing bargain for a Master's degree from a respected institution in the USA.
The article's argument against programs like OMSCS is that the only feedback is from teaching assistants. This claim is not true. There is (a) automated grading of programming assignments, so your program runs against unseen test cases, (b) informal feedback from TAs and other students on Slack and other unofficial channels, (c) formal feedback from other students as part of grades in certain classes and group projects, and (d) feedback from professors to student questions during video office hours (in some courses).
Yes, lots of assignment and exam feedback does come from TAs, but it is comprehensive and valuable. The TAs are well versed in the course material and relay information and questions from and to the professors. I completed the program and at no point did I feel like I was receiving too little feedback on my work.
I also completed the program. Some classes did better at feedback than others. But honestly, whats so bad about getting feedback from teaching assistants? The TA for Reinforcement Learning wrote the book Grokking Deep Reinforcement Learning and has experience using RL in industry. If that is the worst they can say about OMSCS its sort of a pretty big endorsement.
They also have an ms of cybersecurity. Depending on the track you take, it can end up being 9/10 the same classes needed for the CS degree.
I’m currently enrolled and will say the networking is the best part. I’ve found a group of about 50 people in a private slack scattered all over the world. We all bounce things off each other.
It’s amusing to me how this article ended. It really discredits much of the author’s opinion.
I’m a self taught software developer that went back to school to learn the engineering aspects of what I’m doing. Boot camps and online tutorials like what the author appears to be selling teach you how to do a task. Higher education teaches you how to think, learn, and communicate at a higher level.
Do you learn how to build a React app? No. But anyone can go learn that on any of the thousands of websites that promise to turn you into a computer scientist in 6 weeks.
I fully enjoyed my experience in grad school, and wouldn’t trade it for anything. It might not have been the deciding factor in getting a job, but it sure as hell made me more confident.
Readit News
He has a point that if you do a Master's you might be going without the extremely useful basic algorithms and data structures courses, but that is easily avoided by choosing a program that lets you take that. For other topics that aren't strictly programming, eg AI, ML, game theory, cryptography, quantum computing, probabilistic methods, optimization,... it would take a significant amount of dedication to match the amount of understanding that a rigorous course with problem sets can give you in self study.
You mean, the same amount it would take to do it through a university program? I fundamentally don't believe the material is unavailable. If you've done a CS undergrad at a respectable university, you should have enough instinct to build a curriculum to become well versed in those subjects
I am going through the autodidact process for all of those subjects (except quantum computing) (yes, it's brutal, especially alongside a FT job). But with the amount of MIT open courseware, books, and frankly, just wiki pages available, it's been easier than doing it at the pace of a university course
In short: Getting A’s at MIT is simple. Challenging yourself to challenge your colleagues is difficult. Not to mention all the labs.
As a corollary, I’d go on to argue everyone who paid full tuition for online classes during the pandemic got ripped off on a massive scale.
1. Was more relevant and practical. It was taught by people who have worked building real things in the real world. 2. Better ROI on time spent learning. In about 8 weeks (roughly 2 hours every week) and some self study I had the confidence to apply the learnings and continue learning more. Time comes at premium to me as I am working full time and have a family to take care of 3. Cheaper than a university course. It costs 2k USD roughly for a course.
During the same time, I also tried OMSCS from GATECH. I felt I got way more from these Bradfield's courses by spending less time and money than OMSCS as well. That said, its just me and YMMV.
1) free
2) taken from outstanding courses at Berkeley, Stonybrook (the illustrious Steven Skiena), MIT and Stanford.
As I see it, this is evidence that those schools have some good courses. Graduate systems courses are pretty great as well, and of course you can hit other areas such as computer graphics, AI/ML, etc..
For formal education, I'm a bit biased toward multi-year programs (because it takes a while to learn this stuff) and schools (like those above) with good CS courses. But there's nothing to say that the authors' unaccredited graduate program is bad - based on the above I'd expect it's probably decent, although I worry about the 15 hours x 52 weeks though because I don't think that's enough time. In my experience a good systems or engineering course takes about 200 hours.
Also, you're going to have an extremely difficult time getting to work on that stuff without a degree, given that even PhDs often end up on regular business bullshit.
If you want the experience of being in a college campus for a while without the bullshit involved in {undergrad, phd, jobs}, Masters degrees are the perfect middle ground.
I did a Masters degree. I didn’t do very well on some courses, really loved others, learned how to read academic papers and what research would look like if I wanted to do a PhD. But I also had enough free time to make friends, hang out and go on road trips and parties. It was a lot of fun!
Any degree, by the time you have it, might be deprecated or eroded to the point is pointless. If academia doesn't accelerate the pace between education/application, only wealthy people that can afford a lifestyle without income will seek them for entertainment and that's not sustainable to keep a society educated. The value proposition is just not viable.
Education has to go the tutorial way or it will continue to decline.
That’s just CS you are talking about, getting into an emerging field that is very important like Material Informatics would be veeeery hard without an academic setting. Even with CS… honesty I still disagree that academia doesn’t have a purpose or value. Just reading a good research papers recently, one of my favorite recent topics being new ways to discover spy cameras using smartphones. Also some of our greatest minds have received a lot of benefits from an academic environments. I understand bachelor degrees from a lot of places are overpriced and I agree that those seem like a trap and the value might be questionable.
> Any degree, by the time you have it, might be deprecated or eroded to the point is pointless.
My degree cost me out of pocket about ~$5k for my tuition. I stayed in state and my state had a program that paid for tuition if you made above 2.3 GPA in high school. That gave me new friends, sports clubs to join, and understanding of subjects early on that I still use today like psychology and finance. Not to mention my CS degree has paid me back in dividends and I still use knowledge from it today. I have maaany friends in the same boat as me.
> Education has to go the tutorial way or it will continue to decline.
Sounds like a version of tutorial hell potentially but on the job training has its place. AI, ML, game theory, cryptography, quantum computing, probabilistic methods, the list goes on for sincere that would take a significant amount of dedication to match the amount of understanding that a rigorous course with problem sets can give you in self study.
On my Master's InfoSec degree, several students complained that the tech we were studying was really old. The professor simply responded that this is education, not training. You are expected to learn how to analyse the security of systems in general. The specific tech used is just an example.
In what way is a Masters in Comp Sci going to be deprecated or pointless in the 1-2 years it takes to earn it?
I did that for a year and then ended up a class short of graduating because of a scheduling mistake. Since staying a grad assistant wasn't an option, I ultimately dropped out and started working instead. Getting the GA job required me to take the GRE (I scored 170 on the math part, 165 on the verbal and 90 something percentile for both with only minimal studying for the test) and be interviewed about some undergrad CS concepts.
If you're able to do grad school in CS and you've already got permanent legal status in the US, it doesn't really make financial sense even if you're a GA or TA. You can make much more just getting a job. That said, I don't regret doing it because I probably enjoyed that year more than any of my undergrad years (this was pre-pandemic) and I didn't go into any additional debt to do it!
And this isn't just making presentations and writing documentation. It's making effective use of your time in standups. It's conversations about your future relationship with your employer. It's knowing how to ask the right questions to weed out bullshit when digging through other technical documentation respectfully. It's taking a bold new idea that entered your head and figuring out which way would be the best to envangelize it at your organization.
I think in an ideal case, a master's degree in a computing discipline hits the intersection between intense training on a subfield of computing that can be difficult to break into (e.g. statistical machine learning, robotics, formal methods, cryptography) and your demonstration of your "mastery" is taking a cutting edge concept in that subfield and demonstrating mastery through effective technical communication in the form of a project/thesis.
A master's degree isn't guaranteed to make you a "better" programmer, but I would really hope that it would make you much more familiar with the field, and also teach you how to take that familiarity and leverage it to become a more effective communicator.
If you have an advanced degree and your job is making you do standups, you should get another job.
* Opportunity cost
* two tweets of HM saying they don't find it useful
* cs masters aren't geared towards non-cs college grad students
* professors don't know how to handle online teaching
* programs are cash cows (mentioning two non-top-cs programs)
To address the core point about whether those with a masters from a top CS program (say top 20) vs self-taught _on average_ do know more about programming, in my experience they do! Topics like how programs work (interpreted vs compiled languages, memory management), how a program talks to the OS and underlying hardware, and even topics on AI/ML and how to do deep learning. Whether that makes them better at their job depends on what their role is - if it's making a web application or working on underlying infrastructure in C.
So I guess to provide an alternative view to the author's, there are good online programs like UT Austin and Georgia Tech where students don't have to take time off work as they can do it at night / weekends, both programs are low-cost (~$10k), and do have professors who do understand online teaching. I think it's important to make sure you pick colleges and classes geared towards learning, not go with the intention of specific "job training", and likely you'll come out ahead having done it.
I am a self-taught programmer who got later a CS degree. Also a self-taught entrepreneur who later (almost) got an MBA.
My experience is that these are much more valuable for people who already know how to do the stuff but missing the jargon and thoughts of a crowd of exceptionally brilliant people lived before us. I am talking about the scientists, inventors of algorithms, math, bookeeping, all the concepts. How faster you can tell complex thoughts when you just have the vocabulary to use. How do you even present a spectacular idea if you have to spend half an hour explaining something that you should have known has a name.
I never understand how some people get the hubris to think they can do better without all this. Of course if they feel they could not have used all these in their professional lifes - probably getting a degree would have been a waste of time for sure. Some of these folks of course are quite capable and can and will do big things. Also sometimes you have situations when you have to choose between your startup company and the degree. We know some successful dropouts. Still, I think there is big value here.
So I haven't yet pursued a CS degree but whenever I hear of a story like yours it becomes more nagging in the back of my mind.
[1] https://omscs.gatech.edu/prospective-students/faq
Yes, lots of assignment and exam feedback does come from TAs, but it is comprehensive and valuable. The TAs are well versed in the course material and relay information and questions from and to the professors. I completed the program and at no point did I feel like I was receiving too little feedback on my work.
I’m currently enrolled and will say the networking is the best part. I’ve found a group of about 50 people in a private slack scattered all over the world. We all bounce things off each other.
I’m a self taught software developer that went back to school to learn the engineering aspects of what I’m doing. Boot camps and online tutorials like what the author appears to be selling teach you how to do a task. Higher education teaches you how to think, learn, and communicate at a higher level.
Do you learn how to build a React app? No. But anyone can go learn that on any of the thousands of websites that promise to turn you into a computer scientist in 6 weeks.
I fully enjoyed my experience in grad school, and wouldn’t trade it for anything. It might not have been the deciding factor in getting a job, but it sure as hell made me more confident.