Don't learn machine learning

Should you learn machine learning? Yes, you should. But how? By avoiding being a donkey.
There is an abundance of machine learning resources out there, too many.
The good thing is, most of them are pretty good.
But remember, having too many options is the same as no options. Instead of trying to learn them all, pick something, use it, make something with it and if it doesn't work, move onto the next thing.
If you're stuck being a donkey, maybe the procrastination is a sign it's time to move on or find a better option.
Thank you to Sk for sending through the article. If you've got something you want to be reviewed, send it through.
Links:
Don't learn machine learning article - towardsdatascience.com/dont-l...
Building an X-ray classifier - blog.tensorflow.org/2020/03/u...
Teachable machine - teachablemachine.withgoogle.com/
42 days: a cure for shiny object syndrome article - www.mrdbourke.com/42days/
My machine learning course - dbourke.link/mlcourse
Get email updates on my work - dbourke.link/newsletter
Support on Patreon - bit.ly/mrdbourkepatreon
Connect elsewhere:
Web - dbourke.link/web
Quora - dbourke.link/quora
Medium - dbourke.link/medium
Twitter - dbourke.link/twitter
LinkedIn - dbourke.link/linkedin
#machinelearning #datascience

[@MinecraftLetstime]

I am currently enrolled in a MSc Machine learning masters here in the UK, and we are learning both the theory and putting it into practice using the assignments. There really isn't a problem of doing it both. Stop telling yourself you can't.. If you put in the time and effort with the right learning mindset and method, you can learn to eat and drink at the same time. Just need to be smart about it. Stop putting these limiting thoughts on yourself. It really isn't that hard.

[@mrdbourke]

Great reminder.

[@JanessaTechZhao]

That's because you registered a ML master where you seldom encounter the donkey problem. I am a self-taught MLfor 1 year and am going through the donkey problem recently, and I am not the only one case since I met a lot of self-taught ML learners who have the same issue

[@abdelrahmanabdelhadi4195]

Agree , I am self studying ML and I did believe that I should learn alot of stuff to do smth , I guess I might be a donkey 😂 but I'm still at the start of my learning journey and I will work on changing my mindset , thanks for the advice !

[@HelloThere-xs8ss]

Yeah BUT doing many things at once takes longer. Being able to focus and use your strengths (and time) wisely will inevitably be better for you.

[@niranjantamhane3602]

Just understand the tools and find a great alternative for traditional ones', till there is no need don't code from scratch that is something time wasting. And if you are a creator of algorithms no one gonna recognise till you become a problem solver using same algorithms. And one day someone gonna use your tool to solve problems of other and become famous.

[@TheActualTed]

KZbin's algorithm is driving me crazy. Everytime I get interested in a specific field and watch a few videos, 2 days later I get 5 recommendations with the title "DON'T LEARN ..." "STOP DOING ..." "DID YOU ...???. It's usually a clickbait title, which leads to a point "of course you should learn it, just mind some minor detail".

[@amiar7993]

What's clickbait means?!!

[@yuxingchen2237]

Lmaooo actually fun

[@TheActualTed]

@@yuxingchen2237 Is it? It makes me rather angry.

[@amaanvzer6923]

@@amiar7993 a title that is misleading and does not represent the content of the video. The title is such purely to get clicks

[@xtramindofficial8227]

it's also raping me😭😭😭

[@fktudiablo9579]

I have this exact disease... e.g I spend more time learning ways to be more productive than doing concrete things...

[@madcalm2024]

It's normal. You just have some extra time & brain resources to feed your curiosity ))

[@prod.kashkari3075]

Back when I was learning to code I didn’t know anything or what to do and how to do it. I asked a friend of mine how I could what books I should read etc. first thing he did was stop me in my sentence and say “you don’t need to research all these functions and data structures and take notes from textbooks, you just need to know what is possible” and that really reasonated with me because once you know what is possible, you can think of a problem and as the video states “learn by doing”, my friend said to work I’m projects I want to solve or replicate and implement what I have learned that way. Rather than sitting around reading these books.

[@WellBeing999]

Great advice!

[@fanaticalscribe9671]

I’ve just started exploring this field. And this was a really good reminder to get so early on the journey. Thanks a lot!

[@thevictoriaabia]

I agree with you, choose one and adjust as needed

[@taihaozhang869]

There is one exception that proves the rule: when you want to solve a problem without a well matched theoretical model and you need the mathematical understanding to heavily adapt existing theories to your problem (and devise new theories along the way). In this case, you solve the problem by studying machine learning theory, and you learn theory by solving the problem. Sometimes you have to learn enough of the "how" before you can even clearly articulate the "what", especially if the domain defies logical analysis.

[@nguyenhoanglong420]

KINDA LIKE QUANTUM MECHANIC IN PHYSICS ! QUANTUM MECHANIC ALSO DEFIES LOGICAL ANALYSIS AND COMMON SENSE

[@nguyenhoanglong420]

"ABSTRACT" DATA

[@bigboi9049]

Exactl why Im majoring in statistics and then will learn the coding later on in a job

[@artr0x93]

I've worked for about 2 years with ML/computer vision and a lot of the things in this video resonated. When I started out I was frustrated that I couldn't dive directly into designing new advanced mathematical models right away but I realized there are a lot of "boring" things that need to be in place before that's even relevant (data curation, annotation, testing frameworks etc just to name a few). There are a lot of interesting problems to solve along the way that I probably wouldn't even realize existed if I stayed in academia for a PhD. I think there is a point to be made for switching back and forth between learning "tools for the sake of tools" and solving problems though. As long as you spend enough effort to make some progress each time it can really help keeping things fresh

[@brainoverflow98]

I came for a machine learning related talk but I got a life lesson, thanks Daniel nice content! PS: I was thinking about going into ML field but I now realized it's due to engineer disease. I'm sticking with web apps at least for some time until I can decide what I love and what I don't.

[@aguantemotorhead]

Hey Daniel, just discovered your channel a couple days ago and I'm learning lots. Thank you!

[@mrdbourke]

So stoked to hear. Plenty more to come.

[@higormonteiro7636]

I come from a physics background, and usually we spend a lot of time concerning with theory, trying to enrich our theoretical knowledge as much as we can. A lot of colleagues in my department don't understand that trying to learn everything is an infinite loop. During my masters I realized how important is to just start doing things with curiosity, which led me to collaborate in two amazing projects. Now that I'm finishing my masters and starting to get into ML engineering I see how this mindset is even more crucial to let you accomplish the things you want. Amazing content you put there, mate! Thanks from Brazil!

[@francovega7089]

Im happy to listening you, always you make me think about my carreer. Thanks from Argentine

[@mrdbourke]

Thank you from Australia Franco

[@hizokadarkwolf]

thank you for your advice, I'm glad not to be alone on this idea... of completing lots of learning without ever using it. Or getting stuck because you don't know where to start. We just need to start somewhere, it will work eventually.

[@jaredbeckwith]

“Learn by building” exactly what I’ve been doing with simple neural nets

[@victorliendo2010]

That's the approach I have adopted these days

[@jaredbeckwith]

Mike 1412 here’s a video where I explain the simplest neural net -> kzbin.info/www/bejne/n56ZfmCAeduphLs

[@albertog2196]

Same here. I´m following Grokking Deep learning, playing with the code, trying things, and understanding everything from scratch.

[@FirstLast-gk6lg]

@@jaredbeckwith ty for sharing that

[@yokeshkp3797]

What did u learn by building?

[@JanessaTechZhao]

Thank you. You changed my life! In fact I learned a lot of knowledge related to ML, but as time went away I found the more I learn, the more I don't know. This feeling is terrible. Thank you again. I will know where to head next step.

[@ivoriankoua3916]

I'm going to disappoint with the article and even this video even if I agree with some points. Let's say it AGAIN , Machine Learning (mostly Deep Learning in this video) is just Mathematics not coding , so if you're really serious about learning it , you have to take courage and start learning Linear Algebra before tackling ML and when you are about to start Artificial Neural Network you could then start learning about MultiVariable Calculus And if you're about to give some sentence like Driving a car requires no knowledge of the engine. I don't really think it apply in this , suppose that you're the mechanician (and in AI you're) how are you going to repair the car when you have zero knowledge about the car engine ? And also you don't have only do one at the time , for example you could start with a practical course on ML and once you face a concept you don't understand , you search for the Math foundation and come back to your course later ..... It's quite long but I write all of it as a WARNING , AI is not PROGRAMMING and you're going to face a lot of Imposter Syndrome / Dunnel Kruger effect if you plan to go with the No MATH road or rather a lot of copy and paste (but if you're comfortable with working with code you not understand no problem) My final note , I wouldn't be really one side if you aren't really serious also about learning it and just want one feature for example CV , you could grab OpenCV and implement it in your project , for any serious work (like for example searching an actual solution for the Corona Virus ) , come tell me how you do it without the Math Foundation (I really want to know).

[@shreyasdangare4610]

Spot on!

[@tejaswinid1022]

Exactly i too faced same problem. Even if it is tough to do both learning and practical

[@wondays654]

Ivorian Koua I’m currently in college studying software engineering, and most of my professors don’t even expect us to know the math. One prof in particular always says “you don’t have to know all the math, that’s what math majors are for. If we as engineers knew all the math, we won’t need the mathematician”. I’m still early in the field so I’m really conflicted on if I should learn the math or not.

[@TheRcCrazyFan]

@@wondays654 your prof is right that you don't have to learn ALL the math, but you should learn the math of CS well. If ML, no doubt you need to have strong math background to proceed further

[@pllpsy665]

@@wondays654 How correct your professor is depends more on your temperament and your goals. I am kind of in a different position. I graduated from a program that made most math in ML look basic and trivial. When I started on ML on my own after graduating this was both good and bad. Whenever any math was presented I felt compelled to do it. I had to go over all the derivations and verify everything on my own because I was stuck in that mindset. Although that was fun and satisfying It was also a waste of time for little returns in terms of useful insight . I had to force myself out of that mindset and work at a more "high level" . If you are comfortable with black boxes I think you can get pretty far without knowing everything.If you are unable to do that because you really, really need to know what is in the box you will need to learn the math.Keep in mind that once you know it you will need to balance your need to know the details with your need to be productive. That said I think most CE programs should and do require you to take the math courses at some point.

[@skyeparker5822]

Indecision. Donkey problem. KZbin recommended me this video so it must know. Great advice.

[@mwangimukuha]

Thanks for your content. I'm doing data analytics and machine learning. And your content has helped me a lot. Thanks brother!

[@mrwhitedev2655]

I loved the donkey story, I believe many people are gonna relate, I also started studying Computer Science and I know where I am heading to. I have a background on programming but now computer science is my new favourite thing, especially machine learning. Peace mate!

[@saroshpadikkal2362]

Thanks a lot ,I was doing same mistakes, you changed me, and gave me a direction.

[@gregtomka6995]

Excellent video. Great way of describing the donkey problem. It’s easy to stay there too if you’re not aware of the problem.

[@AmCanTech]

Just stumbled upon this video. Thank you for putting links in the description, most creators don't

[@mrdbourke]

That confuses me too haha

[@deepanshurawat4449]

Really nice and a great video. I was too suffering from that so called DONKEY problem. But today you have really shown me the true path. THANKS

[@ricardosierra9281]

Excelente video con una pregunta filosófica profunda. Gracias por compartir esta reflexión sobre la base de tu experiencia. Pregunta: cuáles son las tres herramientas (framework, data base and code writing) de las que hablaste en el vídeo que tú usas?

[@hargobindsingh9763]

You have actually helped me bring clarity in my life and for that Thank you Do well man, and try to keep your videos as concise as your article (Very well written).

[@mrdbourke]

Hey Hargobind, thank you so much for the kind words As for future videos, I'll do my best to speak as clear as my written words

[@ekaterinazakimatova1184]

Hi! Thx for the nice videos. Btw, I love the macbook stand you’re using. What’s the name of it and where did you get it?

[@rgb5521]

Your advice got me thinking...a lot. Thanks!

[@anushlaila]

Real eye opener mate! Thanks for the advice🤘🏻

[@mrdbourke]

You're welcome Anush!

[@kshitijbithel3238]

What was holding me back was that I thought I had less of the basics and should learn them prior to implementing ,but what I saw is that by applying it , the key questions arise about basics which strengthen them ..

[@andrewchang7194]

I resonate with what’s in the description, having too many resources is the same as having none. I’ve had interest in so many things and I’ve pursued all of them through my undergrad years. Started with physics, took everything up to general relativity and dipped my toes into quantum field theory before focusing on math. Studied all undergrad maths and focused on algebra and commutative algebra hoping to learn algebraic geometry. Realized I didn’t like doing any of this so I switched to re-exploring real and functional analysis, and I realized that the premise of machine learning is just optimization of convex functions (most generally) and transforming different data (booleans, strings, int, etc) to a subset of Euclidean vector spaces. I’ve spent so much time studying the background that I’m envious of people who can implement it flawlessly. Good motivation to start practicing implementation. I’ve always wanted to explore machine learning, and now’s a better time than ever before I apply to grad school. Needed something like this much appreciated.

[@isos6534]

why did you don't like Algebraic geometry?

[@andrewchang7194]

IS OS without going off on too big of a tangent, algebraic geometry is extremely long and tedious of a subject. The prerequisites are huge (commutative algebra, analysis, category theory, homological algebra, etc) and there’s no room to do other things because it’s such a deep and nontrivial subject. I wasn’t even that good at it (more of a mathematical physics person here), so it wasn’t worth my time. If you like it, you should definitely study it, or if you’re already studying it I’m impressed! You could definitely tackle anything after getting to that point.

[@kbehari1995]

KZbin just throws wonders at you sometimes. You Sir are just that.

[@azwaraziz6996]

Your example of hardware tools shop enlighten me, thanks !

[@Rusvi1]

I want to learn everything, but long time ago, when I worked on a big game by myself, I realized I have no capacity to know everything. And now, even I WANT to learn everything, I just take stuff people build and modify it for my projects. There is no way I can learn everything, but I have to fight donkey problem every day.

[@Rusvi1]

@Victoria Bergman Maybe you misunderstood, time is an important variable here, and I have to earn money to survive. So when I have two or three projects with deadlines, and they are all in different languages and frameworks, there is no time to fu*k around with stuff I want to learn. Also, who memorizes syntax of over 8 different languages, and I'm not counting html, css, xml and all other crap I have to deal with on front-end side. O, and I forgot database design and administration. Why would I pay for courses? It's all out there for me to learn. No time to do it.

[@xcloudx01alt]

It certainly is hard to say no to wanting to learn new things. It's so exciting and fun, but you're totally right. You end up in a situation where you have too many options, and then you don't even do any of them at all. I'm like that with my projects now. I have so many fun interesting ideas, but i have too many ideas, so i end up overwhelmed and i can't start any of them.

[@MrAbhithepandey]

I practice the ML theory on Kaggle. Simple. I try to solve problems using ML. Period.

[@ikechukwunigelogbuchi4031]

This is a really nice and thought provoking video. We must all avoid wasting unnecessary time being indecisive (the donkey problem) and focus on identifying the real world problems we want to solve and solving it.

[@jorislimonier]

YOU ROCK ! You inspired me into making my own channel, thanks mate 💪🏻💪🏻

[@mrdbourke]

Thank you brother, keep getting after it legend.

[@actzful]

very wise advice. thanks a lot!

[@DJ-bo4pz]

You nailed it man. Lot of content (lectures,videos,tutorials) out there discuss about technical aspects of compSci - straight to logic and code. But no one ever talks about this psychological how-to-go-about-learning things, the correct way to approach learning something new. Hats Off!! Excellent video!! Today I discovered that I have been suppressing my desire to build something, just because in my mind I have been searching for an optimal path, and literally gathering tools, without ever using them, so that when I put in the effort, it would be on the most effective project.....This is just the wrong way. Golden Comp Sci Wisdom: "First find a what, then a how!!!" 99% people do the opposite.

[@mrdbourke]

Thank you so much Dipunj, glad you enjoyed it. All the best my friend

[@DJ-bo4pz]

@@mrdbourke Thank you for the best wishes. Stay safe and healthy. Keep inspiring!!

[@webdev2072]

@@DJ-bo4pz Is it good Idea to do MSc comp sci with Specialization in Machine learning over MCA in India??

[@sidharths9416]

Hey Daniel. Your course on machine learning along with Andre is super awesome!

[@mrdbourke]

Hey Sidharth! Thank you brother, I'm so happy you're enjoying it

[@sidharths9416]

@@mrdbourke As always. Pls continue with such quality courses master💥🙏

[@abhaysoni3146]

Thank you so much Daniel... For realizing that I am donkey and your hardware store keeper example touched me.... Thanks a lot for inspiring.... Cheers!!

[@Trazynn]

I get it I get, have a project to contextualise the theory. But the theory is still incredibly important and without a global idea of what the theory is like, a project can become an obstacle to acquiring it as well. Learn theory as fast as you can, gloss through it, and then start on the projects so you can ground it.

[@Rod24e]

Great Video !! my favorite "first what for im learning ? and then how ?"

[@HVAC-EDUCATION]

Brilliant logic. We have so much curiosity but we need to be careful. Learn only what you need. Otherwise it is forgotten

[@laohein4699]

Thanks god I found this video. This really helps a lot for beginners and prevent them from becoming donkeys.

[@charliema4336]

Your video is very helpful. I am an E&E engineer working with Diesel Engine electronics at the moment(though I got master in Mechanical Engineering). Last year I got interested in ML and took me long time to decide actually do something about it. Then I took a 3 months online course delivered by MIT via EdX(not to advertise but the course is really great and challenging). Now I have a small project which I decided to carry out potentially can be something really cool in the future after I saw the video about using deep learning to learn how to park in Unity simulation. Now I have started learning how to use Unity(yes, it is a game engine) which is really fun. Not sure how the project will turn out but I felt positive and confident about it after watching your video. Many thanks.

[@rusticagenerica]

Totally great video. Well done !!

[@freq-of-nature7699]

This is great. Focus is key. A playful approach can also speed things up depending on how you learn and of course patience. For those interested in R and python, the latest R studio IDE is great and allows you to run both side by side with reticulate. It can seem a steep curve but those with a grounding in R and statistics you can run ML scripts side by side and learn the similarities and where they differ.

[@sanketparab7509]

Too good. Such an eye opener. Thanks Daniel.

[@mrdbourke]

Glad you enjoyed Sanket!

[@bhargavacharan2262]

Thanks for sharing this Daniel. It really helped me in understanding the different perspective.

[@mrdbourke]

Glad you enjoyed Bhargava!

[@theailateshow368]

A great video. I have always thought about these same things

[@sonofkalpana]

Eye opener!!!! Thanks a lot

[@muhammadmehmoodyousuf6235]

Amazing analysis on how learn anything. Well done brother

[@justinrice388]

I really like the point about engineering toolboxes. I think that’s where a lot of the burnout comes from. I work in DevOps and teach full stack so going to school for machine learning sounds like a large leap. Learning how to build better ML pipelines are much more worth my time.

[@aimamalik5126]

You're my new favorite. Absoloutely right.

[@nigelfulton3296]

Needed that donkey metaphor!! Thank you !

[@mrdbourke]

you're welcome Nigel!

[@TechGriva]

Machine Learning is one of the first programming MOOCs Coursera put online by Coursera founder and Stanford Professor Andrew Ng. Although Machine learning has run several times since its first offering and it doesn’t seem to have been changed or updated much since then, it holds up quite well. This course assumes that you have basic programming skills. Assignments also require many vector and matrix operations and slides include some long formulas expressed in summation notation so it is recommended to have some familiarity with linear algebra. You don't need to know calculus or statistics to take this course, but you may gain deeper insight into some of the material if you do. The course uses the Octave programming language, a free clone of MATLAB.

[@pladimir_vutin]

I'm going to make an AI, and call it donkey, then it'll decide whether it's hungry or thirsty...

[@tera6245]

will enrol on your course.. thanks for the reminder..

[@marziehzargari4940]

Thanks.I needed that.

[@mrdbourke]

You’re welcome Marzieh!

[@sivareddyarikatla7672]

glad i watched this video... to remind...learn what is needed...not enough time to learn everything...

[@khanaftab3852]

I love u men ur videos always answer my questions which i couldnt even frame properly

[@mrdbourke]

Thank you Khan!

[@Bueolmutz]

You make a very good point in the and though. Thank you. 😁

[@mrdbourke]

Glad you enjoyed, thank you for the comment

[@anshshrivastava9107]

What I am facing rn is that even though I am watching ML tuts, I do not have a place to apply them. EXACTLY the donkey you told me. I also feel the same way as in LEARNING BY PROJECTS. I do NOT want to be a researcher. I want to provide software that solve problems. You definitely are making my path correct EACH day. Thanks a lot man..!!

[@mrdbourke]

So glad you enjoyed Ansh! All the best legend

[@gero8049]

That is the number one channel that keep me motivate to study

Im a music producer/engineer and my approach to making music is now exactly that "practice and more practice" I barely now check theory but as an engineer im filled with theory and no practice

[@aizensosuke5825]

That was a real opener.... Nice example brother 👊... Subscribed

[@mrdbourke]

Thank you Aizen, glad you enjoyed

[@persebra]

It took me 10 minutes to understand what you was trying to say, but I got it. thanks

[@MasoodOfficial]

very good advice!

[@bharathchand8670]

Thank you for this video. I am also having this problem. We don't have to do the process of 're-inventing the wheel'. Wheel is already there. We just have to make the cart/vehicle only.

[@mrdbourke]

You’re welcome Bharath! Keep learning & creating legend

[@chacho5907]

Amazing video and perspective.

[@jerrychang3253]

Love your point of view, I’m currently working on a project but I really don’t know where to start to improve my model. I already used a RNN model with about 3 or 4 hidden layers and still couldn’t get perfect results. Where would you recommend me to change ??

[@angellahoz1985]

I’m not him, but please stop seeking perfect!!

[@abdelrahmanabdelhadi4195]

The reason I personally want to learn a lot of stuff is that fear that I am not enough to be accepted into a job or a team, after all I am just a high school student who wants to study medicine (a decision I can't change as I only took the A-Levels required already) . I always want to add to my portfolio by learning more stuff and taking more courses to able to prove myself. I don't have a clear idea on how machine learning engineers make money alone without being employed by a company, do you propose an idea to potential clients or build models and find people ready to invest in them? I love machine learning, In fact, I fell in love with data science, my main goal of learning machine learning is to eventually work as a biomedical data scientist but I am always in fear that I won't be able to make use of the things I learn before I graduate from university( medical school takes 7 years in Egypt !!) Whenever I check an ad for a data scientist vacancy there is always the requirement of having a bachelor in CS or math (IDK if this is the case for Egypt only ). Now I know you would ask why did I start studying ML from the first place now rather than waiting till I graduate? Well, I had no idea about these obstacles and I was told by my Python teacher that anyone could find a job as a machine learning engineer no matter what was his or her background. I am sorry the comment was too long but I really wanna benefit from your experience and hopefully resolve my concerns.

[@vinhthekid_mt]

Do ML projects on your own to show that you've actually put your education to use

[@francistembo650]

Give me framewoooooorks!!! Lol 😂This is me seeing everything coming out being built using Pytorch. Thanks for this.

[@mrdbourke]

Hahahaha, you and me both!

[@thefabcoder8637]

@@mrdbourke me too bro

[@adybrain7693]

It’s like martial arts- you learn a arts A-Z of tools over many, many years before you understand what works for you and what doesn’t before you discard what’s useless and use what’s useful....but as you practice overtime a punch isn’t a punch any longer as first thought....continuous learning and making changes to your tools are part of improving and might show you a more effective way forward. So it’s important to remain open minded and flexible to new ways so try new ways out (new tools) as you may find a better way and not waste years on a less effective approach.

[@mrdbourke]

Great analogy.

[@adybrain7693]

Daniel Bourke ....loving the material mate, keep up the good work!

[@alirezamogharabi8733]

Very very useful advices, thanks a lot ❤️❤️❤️❤️❤️❤️❤️🙏🙏🙏

[@ansharora3248]

You've just solved a dilemma. I've explored various channels. Nothing beats you.

[@mrdbourke]

Thank you Ansh!

[@tarikkindell6819]

I have a what, but I have no idea where to start.

[@vibhorsharma7371]

After listening the part from 14.46 , I did two things First smashed like button like hell, second subscribed immidiately as these lines were very relatable Well done sir 👍

[@cornevanzyl5880]

Thanks, I needed to hear that

[@mrdbourke]

You’re welcome Corne! All the best

[@andymounthood]

You mentioned engineer's disease. Wanting to explore data science and machine learning as possible career options, I found three Udemy courses that introduce both of these subjects and look thorough enough to be useful to me (i.e. have more than 12 hours of video). Your course and one other look like they will help me to understand these professions and the problems that they solve, sufficiently to help me decide if either specialty interests me. The third course looked like it took a tool-focused approach rather than a problem-solving approach, so I decided against that course. I think the instructor had engineer's disease.

[@MARIUTSKI13897]

Amazing point of view!

[@AnkitKumawat-ce6ut]

This video appeals to what I actually thought. Great value giving video 😊

[@FirstLast-gk6lg]

What if i am the opposite of you? I want to build build build but dont want to learn specifics because i am afraid it will tamper my creativity. Is it possible for me to stumble through it and end up with a project that uses ML for something without following some tutorial or studying specific theories/algorithms?

[@quackdoescrck3009]

Hey man I'm trying to make products for android applications but i really don't want to learn ml for my upcoming project, please guide me

[@kristiandilov5249]

YOU ARE THE BEST!!!! THANK YOU!!!

[@vishva8kumara]

When you go production with most machine learning related app - pipelines as you mentioned, and batch processing I would say becomes a very important part as well as the ML part itself.

[@vishva8kumara]

I have made myself a solid and versatile batch processing framework that throttles up/down according to available resources - and I integrate that with multiple ML APIs and tools that are already out there for the most part. And that scheme has worked very well.

[@mrdbourke]

That's really cool Vishva, what did you build it with?

[@vishva8kumara]

@@mrdbourke I made it with Node JS. I find asynchronous programming very efficient for batch processing that involves calling external resources (APIs, Databases and even shell programs).

[@VladiGGeorgiev]

Nice title. Just action, focus, exploring.

[@akhilshekkari5023]

You are just awesome brother!!!

[@FloWoelki]

8:57 this moment is everyday when you are a js fullstack engineer xD

[@mrdbourke]

Hahaha, there are definitely a lot of JS frameworks out there.

[@DontKillAnts]

I'd say I'm a chronic donkey across the board, but I definitely have it currently with learning R. I am getting into stats for psych research, so it's a given I am going to choose R to begin with (and ML is more of a hobby I want to pursue, just for fun and the sake of following developments in this area). But even in learning R, I'm stuck between learning Base R and then moving to working with tidyverse and other related packages, or just jumping into these packages from the get go. I've heard advice both ways as to which is better for getting started as both a coding and R noob. I relate to this video, man. And I'm not even in ML yet, nor do I want to do it for a career/profit and it resonates with me.

[@089sailboot1]

DontKillAnts a basic course in r won’t take you too long, then you can directly jump to all the helpful packages out there. Normally one Semester in university makes you a pretty good user

[@michaelnurse9089]

Reading a book called "Broad" at the moment, which takes a somewhat contrarian view on the donkey problem - being that most people who advance specialised fields forward are the donkeys (they have some research to back it up etc) - mostly because experienced donkies can zig and zag way faster than those who are monorailed and can 'import' seemingly useless knowledge into unrelated fields which are game changers. Steve Jobs, Vincent van Gogh, Roger Federer were all donkeys at the start of their careers. Ultimately you are right though - at the start of a career you bring yourself a ton of pain if you do the donkey thing. But over the entire career, you will bring yourself a ton of pain if you do NOT do the donkey thing as you find it too difficult to switch to greener pastures when the time comes and too many opportunities are invisible to you.

[@diegopdomingos]

Hi, is it possible to buy this book@Amazon? Do you have the link? Thanks

[@mrdbourke]

Great insight Michael. Thank you very much for this. A reminder once again balance is achieved by operating at both extremes at the same time. Thank you for the book recommendation too.

[@DontKillAnts]

Who is the author? I can't find the book you are talking about.

[@soblessedongfrfr]

I think the book you're referring to is called "Range" by David Epstein. Great book.

[@garrettsmith2256]

Hey Daniel, thanks for the video! I think you provide a great perspective. My main concern is the amount of math necessary to get a job in ML. What do you think is the essential math knowledge someone in ML should know?

[@garrettsmith2256]

Never mind, you answered my question in your ML Roadmap video. You're the best!

[@mrdbourke]

Haha glad you figured it out legend! All the best

[@archardor3392]

I hate your clickbaits but you speak wisdom so I subbed.

[@mrdbourke]

Thank you for holding on

[@veganboss93]

Great advice!

[@mrdbourke]

Glad you enjoyed Tunç!

[@adeolaogunleye7965]

The only problem I have with some of the object detection models is the false positives. They show up when you least need them no to show up.

[@mrdbourke]

You and me both!

[@RAHULROY-sb9nt]

Good advice...for everyone..but not for me.. nowadays in of billions of people you have willing that to do hardest things,which makes you unique.

[@nazeerasheik4051]

I have been trying to learn ml but I think I wasting my time mostly on where and how to learn

[@escx1613]

Thanks man!

[@benjamingraham2727]

im an investor and i want to use machine learning to analyze financial statements