I think there's plenty of content on Youtube about introductory CS topics, so breaking down advanced topics is something I'd like to see more of. I'm sure a lot of students know the feeling of wanting to learn more about a topic, finding a research paper about it, and then having no idea what's going on, so your videos might help bridge the gap! But if you were to make videos on introductory content, I think it would be nice if you approached it from a theory/algorithmic standpoint. A lot of introductory videos are geared towards coding interviews and problem solving, and I haven't seen anyone take the theory approach (outside of actual college classes, but those aren't as engaging as your videos 😁)
4 years ago
| 36
My opinion is to cover topics that have not been covered (well) on YouTube. As we probably all know, 3B1B has a series on the math behind ML. This was a very solid first step for many who wanted to gain an intuition on the subject, and I’m sure you could tackle another topic in the same manner. Also, I bet you get compared to Grant quite often and I apologize for perpetuating this. You’re already well into the journey of carving a unique path down the forest of CS education, so I’m sure you can do great things.
4 years ago
| 33
Glad to see you are so eager to integrate community feedback into the decision-making process for your videos! I can already tell that no matter what immediate content path you go on, your channel will end up amazing if you keep up this kind of approach.
4 years ago
| 4
Congrats, man! I'm new to the channel but everything I've seen so far has mesmerized me. Your videos are awesome!
4 years ago
| 4
I found your channel two days ago and I love it for the very reason that the topics you discuss are advanced, but you explain them in a beautifully visual way. I love 3b1b but your content stands out because his work is more introductory.
4 years ago
| 1
I think this is the only channel which I found covering both mathematical and programming concept in a single video(FFT) ,please continue such contents of mathematical programming. THANKS A LOT 🙏
4 years ago | 0
Your FFT video was so great I wish you would do more videos like that. Introductory videos are fine and all but you’re one of a few people can really do a fantastic job of breaking down harder topics and I think you will realize your full potential better making harder videos.
4 years ago | 1
I believe people like videos educational videos that either develop the core intuition for a subject or reveal a new and interesting way of visualizing (and by extension gain a deeper understanding of) the subject. The thing that brought me to your channel was obviously the FFT video... which probably also brought a lot of other science/physics people since it’s everywhere in our environment. I think part of the reason the FFT video attracted so much attention was precisely because it concerned a topic that wasn’t so simple on the surface. IMHO, arrays, graphs, data structures and recursion are fairly simple subjects... so the people that flocked to the FFT video did so because they were hoping to get insights that they didn’t receive in class and are still missing. Also, I’d be lying if I said that seeing you using Grant Thompson’s animation library didn’t help suck me in. Anyways, I would suggest giving a bit of a roadmap when you start a video. I think in education, people like to know where you’re taking them and how you plan to take them there FIRST. Even if it adds a few min to the video, I think it’s an invaluable addition that would help retain viewers for a longer duration of the video. When you started off with saying there were two classes of algos (beautiful and useful) I wasn’t really sure where you were going... then you just jumped straight into the explanation. I watched the whole video, but I felt like I was being taken on an adventure down a dark alley without lights. I think people are in general more receptive of educational videos when they start with a “First we’re going to see that... then we’ll see....” before actually being taken down the learning alley. It helps to be able to see the light at the end of the tunnel from every point in the tunnel. One final thing. If I had to make a specific request, it would be for an intuitive understand or visualization of eigenvalues and eigenvectors. I understand and can use them... but I do get the sense that there is much MUCH more to understand. - edited for grammar
4 years ago (edited) | 2
Whatever you make, just don't stop. It is very east to say that the internet is vast and you can find everything on it, but the fact that I've found no resource that explained the essence of the FFT better than you have in my entire engineering degree says a lot. I hope you keep presenting truly magnificent pieces of science in this way so everyone understands how beautiful and satisfying it feels when the concepts just click.
4 years ago | 0
I've only recently run across your channel, and the videos I've watched so far are great. Even though I will look forward to more advanced algorithms and data structure, I really hope you will also continue to cover more of the basic concepts in computer science. It's my observation that a quite a few programmers are weak in these areas, and it is usually the fundamentals that they're weak on, and most tutorial youtube videos I see concentrate on coding, which brings in a lot of unavoidable details that can obscure the fundamentals. I already plan to recommend your videos to a few folks I know who are working their way into computer science / programming from a non-traditional path.
4 years ago
| 0
I'd love to see content on random processes and probability theory, your explanations of material you've covered so far have been great. Suggestion: if you do cover advanced ML techniques, you might want to discuss their limitations as well. I feel like "machine learning" is being sold as a panacea right now, but it has limitations. E.g. deep learning works pretty well for any data you throw at it, but the tradeoff is you have basically no idea what the program will do in any given situation. So when your DL-powered self driving car drives into a parked firetruck, your root cause analysis end up being "idk lol". Basically any material on like, "when is it a good idea to use this algorithm", engineering ethics, etc, could be a nice addition.
4 years ago (edited) | 0
Pick a real hard topic, and create content for people to make their way there. That way you know what you need to do, and can always venture to something else when your imagination takes you for a spin. I'm a web developer who loves the mathematics of computer science. I have spent 3 months on average researching graph theory. I have no degrees, just research in my bedroom. As am audience member, I really want anything that is described in code and relevant to CS, as that's my primary background and expertise. Thank you for explaining and having a great time doing it too!
4 years ago | 0
I'm halfway through my course in Mechatronics Engineering and Science (sci majors in maths, cs) but would like to see both introductory and advanced topics. Obviously the advanced topics directly further my knowledge the most, but it's always interesting to see a different perspective on things I think I already know, especially if I'm teaching people that content. It's the reason why I keep going back to 3b1b's videos on calculus and linear algebra despite having taken multiple subjects in each of them. Sometimes I'll even go watching videos on early high school maths (stuff like solving algebraic equations) to find better ways of teaching it. So I think there's good value in doing both advanced and introductory topics. Introductory topics also give you a chance to structure your videos in a way that ensures your viewers have more prerequisite understanding for the advanced topics if you think they should have that.
4 years ago (edited) | 0
mostly want to learn about machine learning in a way that kind of makes sense
4 years ago
| 14
I really like how you explain the theory behind algorithms that you introduce such that producing the actual code is trivial. I think explaining what an algorithm does is much more effective than just showing code at getting the idea accross
4 years ago | 0
I strongly think there is ton of material available on basics. Very few good insightful resources on advanced materials. Please make more on those.
4 years ago | 0
I think it's really important that CS students not only grasp the complex material but also have strong pedagogical tools & skills to teach programmers with more informal training. So I'm arguing for both. Also, I'd recommend finding but not necessarily making good educational materials for the prerequisite topics before making a video on a topic with lots of prerequisites.
4 years ago
| 6
I thoroughly enjoyed the FFT video but also never knew how dynamic vectors work. I'd love to see those implementations of data structures but also advanced things, to name one i'd love to see a basic neural network implemented. Nevertheless, I love what you are doing.
4 years ago (edited)
| 1
Reducible
Hey everyone, this channel has had insane growth over the past 2 days! First of all, thank you all for all the support and kind comments. I really appreciate it!
I've had a few days to reflect on future content of this channel, and I want to share some of my thoughts with you all. Up until the FFT video, this channel has mostly covered standard data structures, algorithms, and problem solving methods that you'll see in most introductory computer science courses at the college level. This type of content is obviously still valuable to people just started out in CS and I absolutely still want to make some of this foundational content.
But at the same time, the positive feedback of the FFT video gives me confidence that you all also want to explore some more fairly advanced concepts you will likely not see unless you go into some "upper division" CS/EE courses. It also allows me to possibly branch into different subfields and related fields of CS. I already have a few ideas for content that can delve into some interesting content in computer graphics, probabilistic/random processes, signal processing, optimization, artificial intelligence, machine learning, etc.
I think this content is equally valuable and some of the most rich and interesting ideas in CS occur at this level. The tricky aspect of these videos, however, is by the very nature of topics covered, there is a lot more prerequisite knowledge that comes in to truly give justice to these ideas. When I eventually branch into some of these advanced topics, you will see some linear algebra, probability, multivariable calculus, etc. As always, I will always try to make the videos as engaging, clear, and accessible as possible, but there will be some concepts I will have to assume some prior knowledge. I truly believe, however, newcomers to CS will still find these videos valuable and interesting on some level.
I'm putting a poll to query interest as well now that the channel has grown a lot more, and as always feel free to let me know your thoughts in the comments.
4 years ago | [YT] | 358