Brother, as someone currently working in AI research, I got to inform you the best way to learn about it is to follow up with top 10 most influential work released within probably the past 6 months in the topic of your interest. The idea that this thing is mathematically intepretable is not going to help you go far in this field. Get your hands dirty by playing with the code of those projects and modify them with a target idea in your head. If you need any help with the idea or not knowing where to find good source for those projects, find someone who already work in the topic.
Hello, my friend. Ive been studying software engineering and got to college last year. I realized Im good with getting myself familiarized with technologies, but my abstraction is simply shitty. Im not a math genius, sorta speak. So I find it really difficult to chunk out a big problem into smaller pieces and was wondering if studying (actual) math would make me better at that department. Pen and paper. Would that help me at all? Thx in advance. EDIT: I forgot to mention that Im currently doing CS50x on my vacation.
This is inspiring, hearing that it's possible to get into AI research with just 6 months of effort. Honestly I'd assumed it would be years of effort to go that path, but your plan seems very solid. My next few projects are still going to be consuming existing models, but your advice does make me reconsider what I should prioritize after that
@@jeffrey_codes Sorry if the orginal comment is confusing. I wasn’t saying you can do research with merely 6 months of experience but the most influential works in the past 6 months pretty much covers all the practical knowledge you need to get hands on. You definitely need a wide range of knowledge to get a better grasp on the heuristics of the topic and that takes a long time of learning, most likely many years. However, even after all that, most of the time you will find the “top 10 most influential work in the past 6 months” pretty much covers most of the important points that align with the heuristics you gained in the process. So I was suggesting probably you could start with hands on and learn along the way.
@@sotvrno93 I would say what the standard requirement for a cs degree is definitely enough of math, try to understand those material and if you are interested in AI, pick up some signal processing courses ( in which you will find those math courses for a cs degree matters a lot ).
The point about math is probably backwards. Doing math is not about solving some differential equations. It's about gaining understanding and intuition. Some level of exercises is needed, but unless you are solving those problems on a regular basis, you are going to forget anyway. Intuition and understanding probably is going to stick for longer.
Absolutely. I am one of those codecamp software engineers and though I got ahead in my career, I always saw that people with CS degrees who did a lot of math had much more intuition for solvibg programming problems and they were way faster in getting new concepts and learning new technology stacks. I am now over 30 and thinking to go bto math and probably a CS degree to get my brain in this state and thinking.
I get your general idea, but the point in the video (that I agree with) is that when you merely read a math textbook or watch a course and simply reflect on it a little bit, you might get the "intuition" in the sense of "ah, I see more or less why that could be. Makes enough sense" but you don't get any sort of mastery or confident, deep understanding meaning that as soon as you were to go and try to apply it to a complex problem, you'd be completely stumped. Also "doing math" isn't just calculating things. What really gets you the deepest understanding is writing proofs yourself. Sadly he didn't specify exactly what type of exercises he did.
There's a balance. I agree that gaining understanding is the most important thing, but like @gaspaider said, if you don't ever do problems then it's very easy to trick yourself into thinking you understand when you don't. Of course, I'm sure that people with greater mathematical maturity than me have enough learning tricks/instinct built up that they can verify their understanding through proofs and other methods while avoiding calculations. I'm not there yet.
I've been doing math for sometime now. From contests in high school, to the International Mathematical Olympiad(IMO) and currently machine learning and silicon photonics, which are math intensive. I will say, having intuition is a farce. You never truly gain intuition until you solve a few problems yourself and rederive the formula or statement yourself. You'll be deluded into thinking you know but you actually don't. Yes, after the IMO, I've not been solving problems so I am rusty...but the intuition stayed with me. How do I know? well its so so much easier to pick up very difficult fields. It's almost like cheating. Recently I started learning category theory, a pHD level material...people say it is hard. I don't get why...it is so easy to pick up the propositions, intuition and ideas, and to solve problems and thus refine the knowledge and actually make it MINE, not the textbooks'.
Due to certain circumstances, I have not been able to code for the past six months or so and I could feel for the first time my studying/doing ratio alarm beeping hard after what, 6 books? on coding.
i am not even joking when i tell you this, i made an entire schedule to study like 60% of these topics by the end of this year (ml, math dsa (theory focus)), instead of data engineering i was going to deep dive haven't watched the video but wow the yt algorithm is good.
lmao same, i started doing the IBM data engineering professional certificate and bought the Fundamentals of Data Engineering Orielly book, my goal was to learn it, master Discrete Math then deep dive into C programming and Operating Systems along with DSA .....all before the end of this year
If you find all of those topics enjoyable to study, and you have the time, then absolutely go for it! I had a good time studying this year. But if you want to get ahead fast then I'd encourage you to start building in the area you're interested in, while studying only the subjects that are most relevant
@@josjos1847 oh no it definitely is! i learned all of these subjects in uni already but i think the data engineering knowledge i got in class suffices (plus its kinda boring for me)
The hard truth about math is that you still have to learn it the text book way. Visualization videos are nice but at the end they only touch a small corner of each topic
Thank you! I have the advantage that most of this is outside my day job, so I can be honest about setbacks without putting my family at risk. But still difficult to admit!
Completely Agree. I too, spent a lot of time building math fundamentals by studying alone. But looking back, theres really nothing to put on my resume, and I cant really market myself to prospective employers. As a recent grad, this is a non trivial issue. The resolution I came up with is to make this year the year of project-based learning. Cheers!
Good luck on your project-based learning! As a recent grad you have the advantage of tons of free time; you could move 3-4x as fast as me if you really go after it.
Ahh the old Theory Vs Practice dilemma, we all know this intuitively but we still cling to the "One more video, One more course, One more book" idea. There has to be a deeper reason why our brains think like this when we know tangible practice is the correct answer 80% of the time
Thank you so much man! This was exactly what I needed, as I was planning to do the same for quite some time now and for similar reasons (math and stats being foundational pre-req for problem solving especially in the CS domain). They say life is too short to learn from your own mistakes. But here because of an amazingly generous gentleman like yourself. It is possible as you are sharing your experience with us wholeheartedly, Hence the time and effort you spent are very much worth it as you are now impacting the lives of guys like me. Thank you once again. For the immense value you provided.
Great video jeffrey, I too had a similar realisation some days ago, when I participated in my first hackathon. Even though it was all here and there, I learnt a lot more in those hours than I could in weeks
Great video. I've recently decided to double down on programming and have found myself falling into the same patterns of scattered learning. You suggestion to take a top down approach and focus on specific goals is good advice. Good luck to us both in the year to come.
2:06 I literally was about to write a comment ranting about how you cannot learn math from math entertainment like 3blue1brown, but you already said it yourself, and I am glad you did. Some people think they can learn something after reading the "pop" version of it, but that's not the case. If you seriously want to learn math, I suggest you first learn how to write proofs. After you are comfortable with proof writing, you can jump into real analysis (Walter rudin's book, after which you can study proof-based multivariable calculus), topology (Munkres's book), and linear algebra (Sheldon Axler's linear algebra done right, which is way better than gilbert strang). Remember that discrete math is also heavily proof-based, and the best book for it would be Concrete mathematics by Knuth, which is claimed to be the "foundation of computer science". Also, after learning real analysis you can get into probability theory, which relies heavily on real analysis and measure theory. After you get comfortable with probability theory, stochastic analysis and discrete mathematics, you can study deep learning and stuff like that. I hope you find success in your journey, and I understand that my advice probably isn't that helpful since you are more oriented towards comp sci than mathematics.
I wonder how many people working in AI have gotten to that level of mathematics My guess for hardcore AI researchers is 10-20%, maybe as high as 50% - not because all of the math is necessarily needed (some of it is), but because people that smart just tend to enjoy math. For people who aren't innovating in model architecture (whether using existing base models, or training their own models with existing architectures and algorithms), my guess is < 0.1%.
I also studied a lot of deep learning this year and the only stuff I remember really is the stuff I built projects with. I learnt about CNN's, RNN's transformers, reinforcement learning, recommendation algos. I ended only building projects using CNN's and RNN's (mostly due to data availability and compute). I can tell you a lot about how CNN's and RNN's work and good network architecture depending on what problem you optimising for but nothing about other network architectures.
Congratulations on the two successful projects! By doing those you're way ahead of all of the consumption-only students. What's the issue with data availability/compute for reinforcement learning and recommendation algos? Is synthetic data (for recommendation algos) and self-play data (for RL) out of the question?
I am a uni student currently studying deep learning and computer networks, whenever I stumble upon a math concept that I am not familiar with I actually like to get lost in the rabit hole and explore more, but I kinda get your point, juggling all of this with a job and a family must take a lot of effort, on the contrary it is great that you were able to learn all of this within a year.
Good stuff, man. First timer here. Congrats on the baby and STICKING to a study regimen last year. More ppl need to produce videos like this as it properly portrays the fault in your original while detailing what you've learned from that and what you'll do to keep that from happening again
With regards to taking all these theoretical knowledge in "ai" further - Kaggle is a really good way to do it. Just like with leetcode, you will most likely be overwhelmed by how much is going on there, but a good start could be to start reading public notebooks, and/or start taking part in monthly playground competitions
The amount of stuff that you read and learned, considering full time job and a kid, is just amazing! I’m curious how much time on average you dedicate for learning in a week for example?
I usually do 1-4 hours/day, depending on my energy levels, what else is going on with the family, and whether it's a gym day. Baby wakes up at 5am, and my work is two timezones over so it doesn't start until 11am.
I mean you're doing pretty good though in terms of challenging yourself to learn more. I don't think it was that big of a mistake, you just need to apply it + you've now learned how you can improve in order to learn more effectively
The Deep Learning book can be done with an undergrad level of math, although I do have a math minor so I have some small "boost" from that. The problem is a lot of undergrad students don't actually get a firm grip on the math they are meant to learn because everything is so rushed and there just isn't time for yourself. That being said, from what I remember, after the 4th chapter or so it's just a series of techniques that need to be visited outside the book or put into practice. The first 4 chapters or so talk about fundamentals.
Yep, it can absolutely be done without a full undergrad math degree, and I could likely read through it now after just a year of study (4 months of which was math-focused). I also just checked and on the back it says it's suitable for graduate or undergraduate study. But some math background is important! If you read the first 4 chapters and this is your first time seeing the concepts they review, you're going to have a bad time
Hey Jeff, liked the video and wanted to ask whether maybe in the future you may change your mind on this journey you entail. You mention that your main concern was the lack of demonstratable output e.g projects, certificates, degree which is valid but perhaps this studying was necessary in that it covers the foundational requirements that will allow you to fasttrack your work that will have signalling/demonstratable value. Besides a course or going back to school, would it not have been hard to cover areas like ML Maths and CS in a demonstratable way. Really what I am arguing is that you may have sacrificed short term demonstrability to expedite long term demonstrability.
I'll admit I exaggerated a bit... While it's not the *best* use of my time, it was easily in the top 20% uses of my time. It was fun and, like you said, it lays a good foundation for future work. Despite the title, I don't regret doing the work However, I'd still rather have one completed house than three well-laid foundations 😂
Quality content! I can relate to a lot of this actually. Full-time job, 14 month-old daughter, 1st year full-time undergrad (online), etc. Been looking for a good math resource so will have a look at the math academy and recently read a sample of building data-intensive applications but wasn’t sure if I should buy it. And will get back to neetcode as well. I also have most of the books you mentioned. You did well to get through it all in year! Thing is, I really like studying but I prefer to build and apply so will take the project approach going forward. I think main thing I’ve been trying to get my head around is building production grade projects
You just promised to do three different things (MathAcademy, Neetcode, Projects). Focus on one at a time, especially since you have a full-time job, university courses, and a young daughter!
@ I’m a career-changer. I was an electrician for 8 years. Been working at a data analyst/scientist for 1, so I feel behind, hence trying to play catch up. Been coding since just before lockdown so have a software engineering background with python being my main language but it just feels like the tech world moves so fast. It feels like doing so many things at once is the only way to catch up but yeah, I’m trying to slow it down and pace myself. Get all of the things on my list done by the end of my degree instead of in one year.
This is exactly what I am feeling today! I’m thinking how can I do it differently this year. Last year and the year before was a rollercoaster of learning Python, DS, ML, DB and I have been contemplating on doing a deep dive into the Maths of ML while learning about Deep Learning and doing the Cuda and tensorflow stuff
Thanks for sharing your journey. As a senior math major, I agree it takes awhile to learn the math. Seems like some of the topics on your map were a bit scattered. Project ideas: building a data pipeline and running ml models or building deep learning models from scratch. Things like encoding data are complicated but help with solving custom ml problems. Everything can be solved with classification or regression.
Breadth is really important. Even though it might feel like you're scattering your attention all over the place, it pays off. The math is the main rabbit hole for which you have to make a responsible decision if you want to invest your time in it. Basic statistics (IMO) is what every single person should spend time on, its way more important than, say, english or biology. The other math topics might not be that important for most software engineers, but it's really hard for me to see how achieving one's life goals could not require a ton of math. It obviously depends on the goals you choose. Other things like CS, DSA, data engineering, sw architecture and so on are always worth the effort, unless you're digging too deep for no reason. Of course you should always work on a project. It's hard to work on a project, to have a job, a family, and to work on your math and cs related stuff simultaneously. No easy way around it. Sometimes it's really hard to push yourself, and even harder to judge if you'll be able to recover
You can do a surprising amount in the world with very little math! But I do agree that statistics is something that almost everyone would benefit from.
How did you accumulate these resources in the first place? And since your career is in data engineering like you mentioned, what are the most useful data engineering resources you’ve done (books, courses etc)
I like project driver learning. As you noted, this is difficult to do if you want to learn something more abstract. How do you do math projects? In macroeconomics? In physics?
I think the difference is fields where you *build*, versus fields where you *investigate*. If you're not building, a project doesn't really make sense. Maybe set out to re-prove a certain theorem, and then work backwards to build up the math you need to do that?
Yeah that is a good question, I think OPs initial reply to your comment is good. Personally if I was in that situation I'd combine the abstract concept with something I can build. For macroeconomics and physics, I would build programs that would simulate the concept and have a bunch of 'sliders' I could adjust to see what happens when different states are applied to the simulation.
@@johncunningham650 Interesting ! this implies that you can mathematically model those concepts, which is not always the case. But still, using simulation as a learning tool seems like a good idea to me.
Hi Jeffrey, thanks for the video How do I decide that the concept that I should learn using top-down approach or learning the concept from basic to advance level ?
If anyone wants to enter ML/AI, and scare about math part, you really do not need to solve multiple questions or memorize the equations etc. If you understand the intuition about the formula, that would 99% be enough
those are all the books I like by their title and have since been hoarding them but just dont have enough time to finish reading them (cover to cover).. that's a like a stone cast in the pond, i diverge so much I cant even finish reading one
I've been doing something pretty similar but with Physics, Math, and Computer Science. Physics allows mapping math on to the real world. Math helps with mental abstractions and foundations. Computer science allows for pragmatic calculations. The main point is to be become a better problem solver. You need to gain the tools and skills to recognize a problem, map it to abstract constructs, and solve it with the tools at your disposal. Also, I rather not look at things as a waste of time as it's narrowing down your solution space, and creating skill on how to approach knowledge.
I didn't go back to CMU-15445 - I got the Trie prerequisite done, but I still haven't done any C++ But I agree, doing the projects in these courses would have been a great improvement over just watching!
What you did is just 1st semester of MS in Data Science . Practical implementation is done in 2nd Semester , so dont worry . Whatever you learn is not waste . When knowledge saves us , we never know . All the best for future study .
This is insane! We chose the same route but with different Sources! Only difference! I Learn everything with mindset of rewriting all my previous understanding anytime so i never fully accept any concepts or information from my bottom of heart That's some wonderful feeling i can't explain it words. Freedom!
If you want to explore AI, start with linear algebra and calculus from 3Blue1Brown for visualization. Then grind Khan Academy so you can actually learn to solve problems, learn basic Python, and finally do Andrej Karpathy's Zero to One series. I think this is the best way because you'll actually learn by doing, and you'll pick up other necessary concepts as you go. What do you think?
I've been collecting research data for a project's initial topic modeling for over a year. Welcome to the world of AI Data. If you are dissatisfied with the information you reviewed, you learned a lot more than you think. Some never get that far. You'll be more than fine with that sense in your tool belt. I've been using Linux since mid 90s and now at 53 having to learn AWK. You will never have the "perfect knowledge set".
I focus on basics like algebra, real analysis, probability theory and classical algorithms for now. It seems like completely unrelated to applied side, but in fact it is the most effective way to prepare yourself for learning new algorithms / methods. For example, if you truly understand eigenvectors and orthogonal transformations, then PCA is not that hard to grasp in a day of study. Also, my advice is to work with only a good academic literature from renowned publishers like Springer. Sorry, but all those visualisation videos, O’reilly books, online course are not that effective than good old textbooks. PS. Lang’s linear algebra textbooks are bad. Idk why everyone recommends them.
@@jeffrey_codes math majors usually study the fundamentals of mathematical logic, which is binary by nature. People dont think that way. You have to learn it first to be able to read and understand theorems and proofs in later courses like analysis or linear algebra. This was the best investment of time and effort in my life. Also, I am not a math major. My major is finance.
My modest advice. You're trying to accomplish too much, in a period of your life which there are things much more important - your kid - . Don't be that absent parent, your family will be impacted 100%.
Thanks for the concern! I still spend time with my family every evening and most mornings. Study time mostly comes from the time other people spend on commuting and hobbies.
@@jeffrey_codes That's really good. Just commented since I've been there as well.For me it was just too exhausting too keep educating myself while being a parent + full time job. Against all odds, I managed to get a computer science degree by online education, but family & relationship took a toll.
It's definitely a challenge! If I was following someone else's schedule, like you did for a computer science degree, it would be even harder. Congrats on finishing!
Hang on it says on your twitter that u are data engineer already, so i guess those books are related to your field my question is if you have any related degree to what you are currently doing and how long it took you overall to get into industry. Im self taught did half of the books you presented now working on personal projects, i dont have a degree so strong Portfolio is must have.
I got a minor in computer science 14 years ago. After graduation I taught myself Ruby on Rails and Javascript and made a bunch of small games to practice. It was a year before I got a decent-paying gig ($30/hour), but if you're in a tech hub (I was in Arkansas) it should be shorter and the initial pay higher.
Hey Jeffrey, what was the Neural Network Backpropagation course? You mentioned that it's as good as everyone says but idk who the guy in the thumbnail is 😂 Great job on self study otherwise, I'm about to dive into some myself before going for a master's degree in DS. Keep it up!
More application and practicality instead of all the deep abstract theory, it seems, would have been better for you. Of course, a balance of both as you go along is best! Great insight!
I've learned, from experience (ironic, as you will see), that for many things in life, I can do all the research on something I want, and make all the hypothetical, theoretical conclusions and deductions I want, but I won't ever know what it's really like unless I experience it for myself. Whether it be having a child, being in a relationship, working with electronics, coding, writing, working a physical job vs a desk job, etc. Life experience is the best, the greatest teacher. A fundamental trait/part of our reality, and I think that's nice!
I typically study in the morning for a couple hours before work. Baby usually wants to be fed at 5am, so I just stay up after that. I've tried late nights as well, but those don't work as well as they used to
Could you elaborate on your goal with this self-study? If you have a clear, specific idea (e.g., dead-set on ML engineering), I can understand why you say that your approach is too scattered. Otherwise, I think what you've done is excellent for making you more flexible across your career, the math and lower level CS stuff especially. Though the benefits may not be apparent for years. Also, I believe you mention some regret in doing all this learning but having nothing to show for it, and that you'd recommend doing projects to better solidify understanding. I think that projects are probably the best way to learn specific things (e.g., building a Cuda kernel). However, I'd argue that other mediums are better for learning larger-encompassing subjects, given that you actively engage with the material. Passive consumption of books and especially lectures/YT videos doesn't lead to good learning, and it's nigh impossible to "passively" build out a project, so that is an advantage for project-based learning.
I appreciate the perspective from the other side of the issue, and I agree with most of what you say! My original goal was to get out of the "web development" rut - find something more intellectually stimulating and exciting. In the sense that I've studied lots of interesting things and expanded my horizons, it was a rousing success! In the sense of what I do every day at work, it was far from optimal (although not completely without benefits). Probably the biggest benefit is just knowing way more of what's out there, so I can choose better projects in the future - and have lots of the "basics" down so I can jump into other things quicker in the future.
While I was somewhat down on passive consumption in the video, as a corrective to how most people (including me this past year) approach the issue, I agree there are some benefits. My favorite analogy for passive learning is "intellectual tourism". Sure, you can't really say you know a place if you've only been there as a tourist (or only watched youtube videos from the place), but you have a much better idea of what it's like - and can make a better-informed decision on whether you want to explore deeper or not.
@@jeffrey_codes I think that's a fair perspective. So, what are you thinking of learning next? After doing this exploration, have you narrowed your interests? P.S. I am totally stealing that intellectual tourism analogy :)
@@kazzakistan My interests are not narrowed, but my short-term goals are 😂 The next project will be grabbing the low-hanging-fruit of working with LLMs in the context of a web app. It's not exactly a continuation of the more technical things I studied last year, but it lets me start positioning myself in AI using a strength that I already have After that, probably either PyTorch or something with agents
My plan for that is to do the "Deep Learning for Coders" book by Jeremy Howard + the relevant HuggingFace courses... and to actually do all the projects + expand on them to solidify understanding. But since I haven't executed on that plan yet, I can't say for sure whether it's the best way to go.
Build something that's a little bit beyond your current skillset. If you don't have any ideas on what to build, then you can follow a course/video, do the project, and then expand on that project in at least one meaningful way
It's a huge field. Everyone will have a different path. If you're not sure where to start then DDIA is a fun read, but it will just expand your horizons and not help you build a project
If you want a happy medium between a course and a project, CodeCrafters is good. This is an affiliate link but I wouldn't share if it wasn't quality app.codecrafters.io/join?via=jeffreybiles
Every web dev thinks the same thing at some point. "I'm bored, I'm going to hustle and grind my way into a ML role". Now all of those roles are swamped with crappy applicants.
@jeffrey_codes I code only because it pays better than other fields otherwise no way I had been doing this most unstimulating field made in the history of mankind. If I had generational wealth, i would have gone on to pursue English literature/ history/ geography. That type of intellectual knowledge really and in true sense help us to get a good and comprehensive understanding of the world that we live in and not this opening Google colab and constantly getting into errors 🤓
I don't plan on focusing on classical ML for a while, since 1) it's useful in fewer situations than it used to be, now that we're in the age of Deep Learning/LLMs, and 2) MathAcademy has an ML course coming out later this year, so I'll just do that. With that said, Kaggle looks great! My first projects will be about integrating existing models into web apps, and then later this year I want to go a layer deeper and start writing pytorch and training my own models
Whether it's useful or not depends on your goals. I enjoyed watching it; my biggest complaint is that set theory didn't make much sense, but that could be a skill issue. ua-cam.com/play/PLUl4u3cNGP60UlabZBeeqOuoLuj_KNphQ.html
It's too bad I have no interest in learning AI but AI is already replacing my web dev job. AIMl deep learning, math etc seem so boring to me I guess I'll just k m s
Its hard to study like this if you dont have a specific goal, like if you are doing it just for the job market. Thats very broad and tnh, if you already mean to learn something, you will have to build something, but then again, if you are able to build something, and create something on your own, why would you want to be a wage slave in the first place. Just my thoughts...
CS can absolutely be studied on your own! Just make sure to build projects along the way, and start getting work experience as soon as you can. See how University of Waterloo does it. For EE I think they still want you to have the piece of paper for those jobs, but not sure since I don't work in that field.
But imagine if you wouldn’t have learned the theory, your goal this year wouldn’t be to focus more on projects. Do you think one at least needs to take some basic courses, like computer architecture, operation systems, algorithms. Are they not prerequisite of working for a big project?
I've been working professionally as a software engineer for over a decade with almost zero knowledge of computer architecture, operating systems, or algorithms. In the real world, calling `.sort()` is almost always enough. HOWEVER Taking courses can widen your horizons and let you access *more interesting* projects (and after those projects, more interesting paying work)
@ What about someone working closer to hardware, building compilers, and optimizing neural networks for specific architectures? In that scenario, what are your thoughts?
@@uonliaquat7957 that's the "more interesting work" I was referring to 😂 so yes absolutely take the prerequisite courses if you want to do those things
@@jeffrey_codes right, i believe having a good theoretical knowledge of these subjects really help in building projects which are complex and critical, and that probably distinguishes a computer scientist from a computer developer.
@@mostinho7 Is your time really that precious if you are watching educational videos on youtube? Shouldn't you be making deals so that you can get a leg up on the stock market?
I think your experience resonates with a lot of us aspiring software engineers. In my experience, I've learned that consuming new information is only valuable with an objective in mind. otherwise, you're tricking yourself into thinking you're doing something productive, yet you're not producing anything tangible. thanks for your insights!
Summary of video, don't just learn theory but also practise what you have learned. Would like to test the ressources mentioned though. Title of the video is quite a clickbait to scare you and watch it.
Why learn all of these? I also like learning, but more of learning by deconstructing most of the skills that are in demand in the IT industry's job market of my country. By doing that I'm easily hirable, and I can just easily job hop from one company to another. Way better in being output based than scholar or learner. You forget what most of what you learn if you don't really apply it in a real scenario.
I think as humans we're not really meant to do something repeatedly for 10, 20, or 30 years. We should be reinventing ourselves, not make our career part of our personality. Life is a constant motion, but we don't have luxury of that cause our time is limited, and trying to change our career or take risk when we are at 40s, 50s is dangerous and kinda a suicide move. I don't know, maybe the answer is doing something entirely different. Some kind of business but even a business is like gamble, and only a few businesses actually become successful and stay after 10 years.
I did it like you just worst.. I did more LMAO. Worst thing ever. I've learned if your going to learn anything make sure you find a way to apply it in a short period of time. If not learning is just useless by itself.
Lots of people get stuck in the passive learning phase, and for much longer, so don't feel too bad! As long as you're moving the right direction now that's what's important
This is way too much content to cover in one year. It takes time to internalize these concepts. If you don't do exercises and/or get your hands dirty with the things you're studying, you don't really learn anything. Also the topics are very broad and learning about the internals of databases or distributed systems is not something an ML person actually needs. As you said the Kleppmann is amazing, but unless you're responsible for designing large scale systems or building databases, you don't really need to know any of the content of the book in detail.
Brother, as someone currently working in AI research, I got to inform you the best way to learn about it is to follow up with top 10 most influential work released within probably the past 6 months in the topic of your interest. The idea that this thing is mathematically intepretable is not going to help you go far in this field. Get your hands dirty by playing with the code of those projects and modify them with a target idea in your head. If you need any help with the idea or not knowing where to find good source for those projects, find someone who already work in the topic.
Yes.Tutorial hell is real. And wastes a lot of time. Your advice is golden.
Hello, my friend. Ive been studying software engineering and got to college last year. I realized Im good with getting myself familiarized with technologies, but my abstraction is simply shitty. Im not a math genius, sorta speak. So I find it really difficult to chunk out a big problem into smaller pieces and was wondering if studying (actual) math would make me better at that department. Pen and paper. Would that help me at all? Thx in advance.
EDIT: I forgot to mention that Im currently doing CS50x on my vacation.
This is inspiring, hearing that it's possible to get into AI research with just 6 months of effort. Honestly I'd assumed it would be years of effort to go that path, but your plan seems very solid.
My next few projects are still going to be consuming existing models, but your advice does make me reconsider what I should prioritize after that
@@jeffrey_codes Sorry if the orginal comment is confusing. I wasn’t saying you can do research with merely 6 months of experience but the most influential works in the past 6 months pretty much covers all the practical knowledge you need to get hands on. You definitely need a wide range of knowledge to get a better grasp on the heuristics of the topic and that takes a long time of learning, most likely many years. However, even after all that, most of the time you will find the “top 10 most influential work in the past 6 months” pretty much covers most of the important points that align with the heuristics you gained in the process. So I was suggesting probably you could start with hands on and learn along the way.
@@sotvrno93 I would say what the standard requirement for a cs degree is definitely enough of math, try to understand those material and if you are interested in AI, pick up some signal processing courses ( in which you will find those math courses for a cs degree matters a lot ).
The point about math is probably backwards. Doing math is not about solving some differential equations. It's about gaining understanding and intuition. Some level of exercises is needed, but unless you are solving those problems on a regular basis, you are going to forget anyway. Intuition and understanding probably is going to stick for longer.
Absolutely. I am one of those codecamp software engineers and though I got ahead in my career, I always saw that people with CS degrees who did a lot of math had much more intuition for solvibg programming problems and they were way faster in getting new concepts and learning new technology stacks.
I am now over 30 and thinking to go bto math and probably a CS degree to get my brain in this state and thinking.
I get your general idea, but the point in the video (that I agree with) is that when you merely read a math textbook or watch a course and simply reflect on it a little bit, you might get the "intuition" in the sense of "ah, I see more or less why that could be. Makes enough sense" but you don't get any sort of mastery or confident, deep understanding meaning that as soon as you were to go and try to apply it to a complex problem, you'd be completely stumped.
Also "doing math" isn't just calculating things. What really gets you the deepest understanding is writing proofs yourself. Sadly he didn't specify exactly what type of exercises he did.
There's a balance.
I agree that gaining understanding is the most important thing, but like @gaspaider said, if you don't ever do problems then it's very easy to trick yourself into thinking you understand when you don't.
Of course, I'm sure that people with greater mathematical maturity than me have enough learning tricks/instinct built up that they can verify their understanding through proofs and other methods while avoiding calculations. I'm not there yet.
I agree. I think intuition is the foundation, we build our house of knowledge on through rigorous exercise.
I've been doing math for sometime now. From contests in high school, to the International Mathematical Olympiad(IMO) and currently machine learning and silicon photonics, which are math intensive. I will say, having intuition is a farce. You never truly gain intuition until you solve a few problems yourself and rederive the formula or statement yourself. You'll be deluded into thinking you know but you actually don't. Yes, after the IMO, I've not been solving problems so I am rusty...but the intuition stayed with me. How do I know? well its so so much easier to pick up very difficult fields. It's almost like cheating. Recently I started learning category theory, a pHD level material...people say it is hard. I don't get why...it is so easy to pick up the propositions, intuition and ideas, and to solve problems and thus refine the knowledge and actually make it MINE, not the textbooks'.
Recap: shoot for a balance between learning and doing/coding to produce tangible projects
💯
Much appreciated 👍👍👍
Due to certain circumstances, I have not been able to code for the past six months or so and I could feel for the first time my studying/doing ratio alarm beeping hard after what, 6 books? on coding.
i am not even joking when i tell you this, i made an entire schedule to study like 60% of these topics by the end of this year (ml, math dsa (theory focus)), instead of data engineering i was going to deep dive haven't watched the video but wow the yt algorithm is good.
lmao same, i started doing the IBM data engineering professional certificate and bought the Fundamentals of Data Engineering Orielly book, my goal was to learn it, master Discrete Math then deep dive into C programming and Operating Systems along with DSA
.....all before the end of this year
If you find all of those topics enjoyable to study, and you have the time, then absolutely go for it! I had a good time studying this year.
But if you want to get ahead fast then I'd encourage you to start building in the area you're interested in, while studying only the subjects that are most relevant
Why no data engineering? It's not relevant anymore?
@@josjos1847 oh no it definitely is! i learned all of these subjects in uni already but i think the data engineering knowledge i got in class suffices (plus its kinda boring for me)
Same but I'm studying for web dev
The hard truth about math is that you still have to learn it the text book way. Visualization videos are nice but at the end they only touch a small corner of each topic
Very true! I find the visualization videos can help me learn it better/faster, but I still have to spend 80% of my time doing it the textbook way
this was a wonderful video. i don’t feel like this type of honesty is very common on youtube
Thank you! I have the advantage that most of this is outside my day job, so I can be honest about setbacks without putting my family at risk. But still difficult to admit!
Completely Agree. I too, spent a lot of time building math fundamentals by studying alone. But looking back, theres really nothing to put on my resume, and I cant really market myself to prospective employers. As a recent grad, this is a non trivial issue. The resolution I came up with is to make this year the year of project-based learning. Cheers!
Good luck on your project-based learning! As a recent grad you have the advantage of tons of free time; you could move 3-4x as fast as me if you really go after it.
@@jeffrey_codes What are some good project ideas to start doing? Where should I look to see people's sourcecode on their own AI models?
Thank you for sharing your experience.
Please consider adding links to the mentioned learning material in the video description.
Great idea! Added links to my favorites in the description.
Thanks!!!
Ahh the old Theory Vs Practice dilemma, we all know this intuitively but we still cling to the "One more video, One more course, One more book" idea.
There has to be a deeper reason why our brains think like this when we know tangible practice is the correct answer 80% of the time
We see more courses in one minute scrolling youtube than a medieval peasant would see in his lifetime. It makes sense to want to hoard the knowledge!
That's so empathetic to your brain, love this framing. We must learn to tame the knowledge-greedy peasant @@jeffrey_codes
Thank you so much man! This was exactly what I needed, as I was planning to do the same for quite some time now and for similar reasons (math and stats being foundational pre-req for problem solving especially in the CS domain).
They say life is too short to learn from your own mistakes. But here because of an amazingly generous gentleman like yourself. It is possible as you are sharing your experience with us wholeheartedly, Hence the time and effort you spent are very much worth it as you are now impacting the lives of guys like me. Thank you once again. For the immense value you provided.
Great video jeffrey, I too had a similar realisation some days ago, when I participated in my first hackathon. Even though it was all here and there, I learnt a lot more in those hours than I could in weeks
💯, now just keep working like you did in the hackathon and you're good to go!
How did you participate in hackathons?
Awesome video! Looking forward to your project videos
Great video. I've recently decided to double down on programming and have found myself falling into the same patterns of scattered learning. You suggestion to take a top down approach and focus on specific goals is good advice. Good luck to us both in the year to come.
Good luck on your programming journey!
Great video , Thanks for sharing!
2:06 I literally was about to write a comment ranting about how you cannot learn math from math entertainment like 3blue1brown, but you already said it yourself, and I am glad you did. Some people think they can learn something after reading the "pop" version of it, but that's not the case.
If you seriously want to learn math, I suggest you first learn how to write proofs. After you are comfortable with proof writing, you can jump into real analysis (Walter rudin's book, after which you can study proof-based multivariable calculus), topology (Munkres's book), and linear algebra (Sheldon Axler's linear algebra done right, which is way better than gilbert strang). Remember that discrete math is also heavily proof-based, and the best book for it would be Concrete mathematics by Knuth, which is claimed to be the "foundation of computer science". Also, after learning real analysis you can get into probability theory, which relies heavily on real analysis and measure theory.
After you get comfortable with probability theory, stochastic analysis and discrete mathematics, you can study deep learning and stuff like that.
I hope you find success in your journey, and I understand that my advice probably isn't that helpful since you are more oriented towards comp sci than mathematics.
I wonder how many people working in AI have gotten to that level of mathematics
My guess for hardcore AI researchers is 10-20%, maybe as high as 50% - not because all of the math is necessarily needed (some of it is), but because people that smart just tend to enjoy math.
For people who aren't innovating in model architecture (whether using existing base models, or training their own models with existing architectures and algorithms), my guess is < 0.1%.
I also studied a lot of deep learning this year and the only stuff I remember really is the stuff I built projects with. I learnt about CNN's, RNN's transformers, reinforcement learning, recommendation algos. I ended only building projects using CNN's and RNN's (mostly due to data availability and compute). I can tell you a lot about how CNN's and RNN's work and good network architecture depending on what problem you optimising for but nothing about other network architectures.
Congratulations on the two successful projects! By doing those you're way ahead of all of the consumption-only students.
What's the issue with data availability/compute for reinforcement learning and recommendation algos? Is synthetic data (for recommendation algos) and self-play data (for RL) out of the question?
Where do you get your data from?
I am a uni student currently studying deep learning and computer networks, whenever I stumble upon a math concept that I am not familiar with I actually like to get lost in the rabit hole and explore more, but I kinda get your point, juggling all of this with a job and a family must take a lot of effort, on the contrary it is great that you were able to learn all of this within a year.
Good stuff, man. First timer here. Congrats on the baby and STICKING to a study regimen last year. More ppl need to produce videos like this as it properly portrays the fault in your original while detailing what you've learned from that and what you'll do to keep that from happening again
Subbed. Super video!
With regards to taking all these theoretical knowledge in "ai" further - Kaggle is a really good way to do it. Just like with leetcode, you will most likely be overwhelmed by how much is going on there, but a good start could be to start reading public notebooks, and/or start taking part in monthly playground competitions
This is a good idea!
The amount of stuff that you read and learned, considering full time job and a kid, is just amazing! I’m curious how much time on average you dedicate for learning in a week for example?
I usually do 1-4 hours/day, depending on my energy levels, what else is going on with the family, and whether it's a gym day. Baby wakes up at 5am, and my work is two timezones over so it doesn't start until 11am.
I mean you're doing pretty good though in terms of challenging yourself to learn more. I don't think it was that big of a mistake, you just need to apply it + you've now learned how you can improve in order to learn more effectively
Super useful. Subbed 🥂
The Deep Learning book can be done with an undergrad level of math, although I do have a math minor so I have some small "boost" from that. The problem is a lot of undergrad students don't actually get a firm grip on the math they are meant to learn because everything is so rushed and there just isn't time for yourself. That being said, from what I remember, after the 4th chapter or so it's just a series of techniques that need to be visited outside the book or put into practice. The first 4 chapters or so talk about fundamentals.
Yep, it can absolutely be done without a full undergrad math degree, and I could likely read through it now after just a year of study (4 months of which was math-focused). I also just checked and on the back it says it's suitable for graduate or undergraduate study.
But some math background is important! If you read the first 4 chapters and this is your first time seeing the concepts they review, you're going to have a bad time
Hey Jeff, liked the video and wanted to ask whether maybe in the future you may change your mind on this journey you entail. You mention that your main concern was the lack of demonstratable output e.g projects, certificates, degree which is valid but perhaps this studying was necessary in that it covers the foundational requirements that will allow you to fasttrack your work that will have signalling/demonstratable value. Besides a course or going back to school, would it not have been hard to cover areas like ML Maths and CS in a demonstratable way. Really what I am arguing is that you may have sacrificed short term demonstrability to expedite long term demonstrability.
I'll admit I exaggerated a bit... While it's not the *best* use of my time, it was easily in the top 20% uses of my time.
It was fun and, like you said, it lays a good foundation for future work. Despite the title, I don't regret doing the work
However, I'd still rather have one completed house than three well-laid foundations 😂
This reflection is FABULOUS THANK YOU
Quality content! I can relate to a lot of this actually. Full-time job, 14 month-old daughter, 1st year full-time undergrad (online), etc.
Been looking for a good math resource so will have a look at the math academy and recently read a sample of building data-intensive applications but wasn’t sure if I should buy it.
And will get back to neetcode as well. I also have most of the books you mentioned. You did well to get through it all in year!
Thing is, I really like studying but I prefer to build and apply so will take the project approach going forward.
I think main thing I’ve been trying to get my head around is building production grade projects
You just promised to do three different things (MathAcademy, Neetcode, Projects). Focus on one at a time, especially since you have a full-time job, university courses, and a young daughter!
@ I’m a career-changer. I was an electrician for 8 years. Been working at a data analyst/scientist for 1, so I feel behind, hence trying to play catch up. Been coding since just before lockdown so have a software engineering background with python being my main language but it just feels like the tech world moves so fast. It feels like doing so many things at once is the only way to catch up but yeah, I’m trying to slow it down and pace myself. Get all of the things on my list done by the end of my degree instead of in one year.
I'm not saying to slow down - I'm saying to focus. You'll see progress much faster, and you can always switch your focus if you get bored
I really loved this bro :)
"if your goal is math entertainment" damn brother went deeeep hahah glad you were able to make it back man
Keep up the good work
This is exactly what I am feeling today! I’m thinking how can I do it differently this year. Last year and the year before was a rollercoaster of learning Python, DS, ML, DB and I have been contemplating on doing a deep dive into the Maths of ML while learning about Deep Learning and doing the Cuda and tensorflow stuff
Good luck, and remember to use whatever you're learning in projects if possible!
Thanks, this is useful information
Thanks for sharing your journey. As a senior math major, I agree it takes awhile to learn the math. Seems like some of the topics on your map were a bit scattered. Project ideas: building a data pipeline and running ml models or building deep learning models from scratch. Things like encoding data are complicated but help with solving custom ml problems. Everything can be solved with classification or regression.
I am a student and it was insightful. Thank you for sharing your experience!
That was inspiring. Thanks for sharing
Too bad I haven’t had the chance to see a video like this earlier, I made the same mistakes, but ok, I am glad we are on the same page
I am in my first year of a data science masters degree and this si so helpfull. Thank you!
Breadth is really important. Even though it might feel like you're scattering your attention all over the place, it pays off.
The math is the main rabbit hole for which you have to make a responsible decision if you want to invest your time in it. Basic statistics (IMO) is what every single person should spend time on, its way more important than, say, english or biology. The other math topics might not be that important for most software engineers, but it's really hard for me to see how achieving one's life goals could not require a ton of math. It obviously depends on the goals you choose.
Other things like CS, DSA, data engineering, sw architecture and so on are always worth the effort, unless you're digging too deep for no reason.
Of course you should always work on a project. It's hard to work on a project, to have a job, a family, and to work on your math and cs related stuff simultaneously. No easy way around it. Sometimes it's really hard to push yourself, and even harder to judge if you'll be able to recover
You can do a surprising amount in the world with very little math! But I do agree that statistics is something that almost everyone would benefit from.
This is really nice display of the roadmap and its correction. Learn how to learn is so important! Thanks for sharing❤
I just highly resonate with you.
Haven't watched the video but did the exact same thing last year. I can say it's worth it.
How did you accumulate these resources in the first place? And since your career is in data engineering like you mentioned, what are the most useful data engineering resources you’ve done (books, courses etc)
Great video and advice.
Respect for learning do much while working and having a baby.
I appreciate the video
I like project driver learning.
As you noted, this is difficult to do if you want to learn something more abstract.
How do you do math projects? In macroeconomics? In physics?
I think the difference is fields where you *build*, versus fields where you *investigate*. If you're not building, a project doesn't really make sense.
Maybe set out to re-prove a certain theorem, and then work backwards to build up the math you need to do that?
Yeah that is a good question, I think OPs initial reply to your comment is good. Personally if I was in that situation I'd combine the abstract concept with something I can build.
For macroeconomics and physics, I would build programs that would simulate the concept and have a bunch of 'sliders' I could adjust to see what happens when different states are applied to the simulation.
@@johncunningham650 Interesting ! this implies that you can mathematically model those concepts, which is not always the case. But still, using simulation as a learning tool seems like a good idea to me.
Hi Jeffrey, thanks for the video
How do I decide that the concept that I should learn using top-down approach or learning the concept from basic to advance level ?
If anyone wants to enter ML/AI, and scare about math part, you really do not need to solve multiple questions or memorize the equations etc. If you understand the intuition about the formula, that would 99% be enough
those are all the books I like by their title and have since been hoarding them but just dont have enough time to finish reading them (cover to cover).. that's a like a stone cast in the pond, i diverge so much I cant even finish reading one
So what's your advice on learning Machine Learning? How do I keep learning until I become a hireable engineer?
Great video, waiting for the video of you're project! I'm on a similar track.
I've been doing something pretty similar but with Physics, Math, and Computer Science. Physics allows mapping math on to the real world. Math helps with mental abstractions and foundations. Computer science allows for pragmatic calculations. The main point is to be become a better problem solver. You need to gain the tools and skills to recognize a problem, map it to abstract constructs, and solve it with the tools at your disposal. Also, I rather not look at things as a waste of time as it's narrowing down your solution space, and creating skill on how to approach knowledge.
Did you finish projects in CMU-15445? I think that serves your needs mostly regarding the "accomplishment" you want.
I didn't go back to CMU-15445 - I got the Trie prerequisite done, but I still haven't done any C++
But I agree, doing the projects in these courses would have been a great improvement over just watching!
What you did is just 1st semester of MS in Data Science . Practical implementation is done in 2nd Semester , so dont worry . Whatever you learn is not waste . When knowledge saves us , we never know . All the best for future study .
yep, can share a very similar story
This is insane! We chose the same route but with different Sources!
Only difference! I Learn everything with mindset of rewriting all my previous understanding anytime so i never fully accept any concepts or information from my bottom of heart That's some wonderful feeling i can't explain it words.
Freedom!
If you want to explore AI, start with linear algebra and calculus from 3Blue1Brown for visualization. Then grind Khan Academy so you can actually learn to solve problems, learn basic Python, and finally do Andrej Karpathy's Zero to One series. I think this is the best way because you'll actually learn by doing, and you'll pick up other necessary concepts as you go. What do you think?
Math academy looks good but it's so expensive, wish they had parity option for some countries
I've been collecting research data for a project's initial topic modeling for over a year. Welcome to the world of AI Data. If you are dissatisfied with the information you reviewed, you learned a lot more than you think. Some never get that far. You'll be more than fine with that sense in your tool belt. I've been using Linux since mid 90s and now at 53 having to learn AWK. You will never have the "perfect knowledge set".
I mean even as a student it is overwhelming, I thought I was the only one stuck there for me it was : graphics programming , ML , low-level diving
I focus on basics like algebra, real analysis, probability theory and classical algorithms for now. It seems like completely unrelated to applied side, but in fact it is the most effective way to prepare yourself for learning new algorithms / methods. For example, if you truly understand eigenvectors and orthogonal transformations, then PCA is not that hard to grasp in a day of study.
Also, my advice is to work with only a good academic literature from renowned publishers like Springer. Sorry, but all those visualisation videos, O’reilly books, online course are not that effective than good old textbooks.
PS. Lang’s linear algebra textbooks are bad. Idk why everyone recommends them.
Ah yes, the basics, like Real Analysis 😂
There's definitely a math majors vs everyone else split in how people think
@@jeffrey_codes math majors usually study the fundamentals of mathematical logic, which is binary by nature. People dont think that way. You have to learn it first to be able to read and understand theorems and proofs in later courses like analysis or linear algebra. This was the best investment of time and effort in my life.
Also, I am not a math major. My major is finance.
My modest advice. You're trying to accomplish too much, in a period of your life which there are things much more important - your kid - . Don't be that absent parent, your family will be impacted 100%.
Thanks for the concern! I still spend time with my family every evening and most mornings. Study time mostly comes from the time other people spend on commuting and hobbies.
@@jeffrey_codes That's really good. Just commented since I've been there as well.For me it was just too exhausting too keep educating myself while being a parent + full time job. Against all odds, I managed to get a computer science degree by online education, but family & relationship took a toll.
It's definitely a challenge! If I was following someone else's schedule, like you did for a computer science degree, it would be even harder. Congrats on finishing!
Hi, great video. A weekly release if similar content would be very appreciated
thank u mane, im in the same situation rn lol
Hang on it says on your twitter that u are data engineer already, so i guess those books are related to your field my question is if you have any related degree to what you are currently doing and how long it took you overall to get into industry. Im self taught did half of the books you presented now working on personal projects, i dont have a degree so strong Portfolio is must have.
I got a minor in computer science 14 years ago. After graduation I taught myself Ruby on Rails and Javascript and made a bunch of small games to practice. It was a year before I got a decent-paying gig ($30/hour), but if you're in a tech hub (I was in Arkansas) it should be shorter and the initial pay higher.
Hey Jeffrey, what was the Neural Network Backpropagation course? You mentioned that it's as good as everyone says but idk who the guy in the thumbnail is 😂
Great job on self study otherwise, I'm about to dive into some myself before going for a master's degree in DS. Keep it up!
Andrej Karpathy. He was a leader at both Tesla and OpenAI
Be sure to actually use what you're studying! 2 years experience as a data scientist is just as good as a master's degree
More application and practicality instead of all the deep abstract theory, it seems, would have been better for you. Of course, a balance of both as you go along is best! Great insight!
I've learned, from experience (ironic, as you will see), that for many things in life, I can do all the research on something I want, and make all the hypothetical, theoretical conclusions and deductions I want, but I won't ever know what it's really like unless I experience it for myself. Whether it be having a child, being in a relationship, working with electronics, coding, writing, working a physical job vs a desk job, etc. Life experience is the best, the greatest teacher. A fundamental trait/part of our reality, and I think that's nice!
It’s *Ammazing* what you’ve been able to achieve with a full-time job and a baby! What time did you study usually? What worked for you? Thanks.
I typically study in the morning for a couple hours before work. Baby usually wants to be fed at 5am, so I just stay up after that. I've tried late nights as well, but those don't work as well as they used to
Could you elaborate on your goal with this self-study? If you have a clear, specific idea (e.g., dead-set on ML engineering), I can understand why you say that your approach is too scattered. Otherwise, I think what you've done is excellent for making you more flexible across your career, the math and lower level CS stuff especially. Though the benefits may not be apparent for years.
Also, I believe you mention some regret in doing all this learning but having nothing to show for it, and that you'd recommend doing projects to better solidify understanding. I think that projects are probably the best way to learn specific things (e.g., building a Cuda kernel). However, I'd argue that other mediums are better for learning larger-encompassing subjects, given that you actively engage with the material. Passive consumption of books and especially lectures/YT videos doesn't lead to good learning, and it's nigh impossible to "passively" build out a project, so that is an advantage for project-based learning.
I appreciate the perspective from the other side of the issue, and I agree with most of what you say!
My original goal was to get out of the "web development" rut - find something more intellectually stimulating and exciting. In the sense that I've studied lots of interesting things and expanded my horizons, it was a rousing success! In the sense of what I do every day at work, it was far from optimal (although not completely without benefits).
Probably the biggest benefit is just knowing way more of what's out there, so I can choose better projects in the future - and have lots of the "basics" down so I can jump into other things quicker in the future.
While I was somewhat down on passive consumption in the video, as a corrective to how most people (including me this past year) approach the issue, I agree there are some benefits.
My favorite analogy for passive learning is "intellectual tourism". Sure, you can't really say you know a place if you've only been there as a tourist (or only watched youtube videos from the place), but you have a much better idea of what it's like - and can make a better-informed decision on whether you want to explore deeper or not.
@@jeffrey_codes I think that's a fair perspective. So, what are you thinking of learning next? After doing this exploration, have you narrowed your interests?
P.S. I am totally stealing that intellectual tourism analogy :)
@@kazzakistan My interests are not narrowed, but my short-term goals are 😂
The next project will be grabbing the low-hanging-fruit of working with LLMs in the context of a web app. It's not exactly a continuation of the more technical things I studied last year, but it lets me start positioning myself in AI using a strength that I already have
After that, probably either PyTorch or something with agents
love the video! I want to learn deep learning, specifically cv and nlp, but I'm feeling lost.
Do you know any good material/advice?
Studying NLP these days with a new LLM each month is useless unless you are building that models
My plan for that is to do the "Deep Learning for Coders" book by Jeremy Howard + the relevant HuggingFace courses... and to actually do all the projects + expand on them to solidify understanding.
But since I haven't executed on that plan yet, I can't say for sure whether it's the best way to go.
What would you recommend for backend engineering (senior+ level) what should one study?
Build something that's a little bit beyond your current skillset. If you don't have any ideas on what to build, then you can follow a course/video, do the project, and then expand on that project in at least one meaningful way
@@jeffrey_codes thanks for the reply! Are there courses that implement backend/data engineering projects that you could recommend
It's a huge field. Everyone will have a different path. If you're not sure where to start then DDIA is a fun read, but it will just expand your horizons and not help you build a project
If you want a happy medium between a course and a project, CodeCrafters is good. This is an affiliate link but I wouldn't share if it wasn't quality app.codecrafters.io/join?via=jeffreybiles
Every web dev thinks the same thing at some point. "I'm bored, I'm going to hustle and grind my way into a ML role". Now all of those roles are swamped with crappy applicants.
If you think ML roles are swamped with crappy web devs, wait till you see the web dev roles!
3:48 "not the best use of my time" is what sums up my entire outlook on the undergrad i did in CS. Too deep and really not worth the time
Yep, it's really interesting stuff but only a good use of time if you have a place to practice what you've learned
@jeffrey_codes I code only because it pays better than other fields otherwise no way I had been doing this most unstimulating field made in the history of mankind. If I had generational wealth, i would have gone on to pursue English literature/ history/ geography. That type of intellectual knowledge really and in true sense help us to get a good and comprehensive understanding of the world that we live in and not this opening Google colab and constantly getting into errors 🤓
What is the learning resource next to Coredumped (blue square) ?
Branch Education. Interesting videos, but VERY high-level
Haven’t even started watching the video but I was dreaming of this
Thank you, I have experienced the same thing. For ML, what projects are you doing? Have you thought about Kaggle competition?
I don't plan on focusing on classical ML for a while, since 1) it's useful in fewer situations than it used to be, now that we're in the age of Deep Learning/LLMs, and 2) MathAcademy has an ML course coming out later this year, so I'll just do that. With that said, Kaggle looks great!
My first projects will be about integrating existing models into web apps, and then later this year I want to go a layer deeper and start writing pytorch and training my own models
to do all this with a job and a BABY, incredible
Solid video
what is the playlist of discreate mathematics. and is it useful to watch.
Whether it's useful or not depends on your goals. I enjoyed watching it; my biggest complaint is that set theory didn't make much sense, but that could be a skill issue.
ua-cam.com/play/PLUl4u3cNGP60UlabZBeeqOuoLuj_KNphQ.html
great video
1:40 The last days I am seeing gilvert strang in my dreams. His series is really good
It's too bad I have no interest in learning AI but AI is already replacing my web dev job. AIMl deep learning, math etc seem so boring to me I guess I'll just k m s
+1
AI/ML is the most boring dogshit ever
Its hard to study like this if you dont have a specific goal, like if you are doing it just for the job market. Thats very broad and tnh, if you already mean to learn something, you will have to build something, but then again, if you are able to build something, and create something on your own, why would you want to be a wage slave in the first place. Just my thoughts...
Bro im thinking about doing the same thing. CS, EE, and Physics on my own. So i can save on college😭😭.
CS can absolutely be studied on your own! Just make sure to build projects along the way, and start getting work experience as soon as you can. See how University of Waterloo does it.
For EE I think they still want you to have the piece of paper for those jobs, but not sure since I don't work in that field.
@ alright Thanks!
Your attack on Angular was out of pocket...
Mathematics is the abstract science of number and measure.
When you stop measuring, like in Set Theory and Topology, you stop doing mathematics.
But imagine if you wouldn’t have learned the theory, your goal this year wouldn’t be to focus more on projects. Do you think one at least needs to take some basic courses, like computer architecture, operation systems, algorithms. Are they not prerequisite of working for a big project?
I've been working professionally as a software engineer for over a decade with almost zero knowledge of computer architecture, operating systems, or algorithms. In the real world, calling `.sort()` is almost always enough.
HOWEVER
Taking courses can widen your horizons and let you access *more interesting* projects (and after those projects, more interesting paying work)
@ What about someone working closer to hardware, building compilers, and optimizing neural networks for specific architectures? In that scenario, what are your thoughts?
@@uonliaquat7957 that's the "more interesting work" I was referring to 😂 so yes absolutely take the prerequisite courses if you want to do those things
@@jeffrey_codes right, i believe having a good theoretical knowledge of these subjects really help in building projects which are complex and critical, and that probably distinguishes a computer scientist from a computer developer.
you have a baby. i am gonna subscribe to you. 🎉
Video Starts 11:55
not worth watching?
@@mostinho7 Is your time really that precious if you are watching educational videos on youtube? Shouldn't you be making deals so that you can get a leg up on the stock market?
Agree 100% 2:00
I think your experience resonates with a lot of us aspiring software engineers. In my experience, I've learned that consuming new information is only valuable with an objective in mind. otherwise, you're tricking yourself into thinking you're doing something productive, yet you're not producing anything tangible. thanks for your insights!
💯
Watching the video got feeling you are talking on behalf of "Math" =)
Summary of video, don't just learn theory but also practise what you have learned.
Would like to test the ressources mentioned though.
Title of the video is quite a clickbait to scare you and watch it.
Why learn all of these? I also like learning, but more of learning by deconstructing most of the skills that are in demand in the IT industry's job market of my country. By doing that I'm easily hirable, and I can just easily job hop from one company to another. Way better in being output based than scholar or learner. You forget what most of what you learn if you don't really apply it in a real scenario.
That's what I've been doing for a decade+. I needed to break out of the routine
I think as humans we're not really meant to do something repeatedly for 10, 20, or 30 years. We should be reinventing ourselves, not make our career part of our personality.
Life is a constant motion, but we don't have luxury of that cause our time is limited, and trying to change our career or take risk when we are at 40s, 50s is dangerous and kinda a suicide move.
I don't know, maybe the answer is doing something entirely different. Some kind of business but even a business is like gamble, and only a few businesses actually become successful and stay after 10 years.
why didn't you study algorithms instead, those will always be useful
5:25
But you are not an AI research scientist🤔
??? I never said I was ???
real YMs(yung mathematicians) WILL rawdog 2025 learning advanced math topics.
😂
he like me frfr
denji-grabbing-tv.jpg
I did it like you just worst.. I did more LMAO. Worst thing ever. I've learned if your going to learn anything make sure you find a way to apply it in a short period of time. If not learning is just useless by itself.
Lots of people get stuck in the passive learning phase, and for much longer, so don't feel too bad! As long as you're moving the right direction now that's what's important
This is way too much content to cover in one year. It takes time to internalize these concepts. If you don't do exercises and/or get your hands dirty with the things you're studying, you don't really learn anything. Also the topics are very broad and learning about the internals of databases or distributed systems is not something an ML person actually needs. As you said the Kleppmann is amazing, but unless you're responsible for designing large scale systems or building databases, you don't really need to know any of the content of the book in detail.
I'm glad you agree with the video!
web dev is a mistake hahahahaha
A profitable mistake, but a mistake nonetheless
@@jeffrey_codes Devops and Data Engineering and MLops are other very profitable mistakes =P