Ask a Machine Learning Engineer Anything (live) | April 2020

Oh, we might be live on Twitch. Am I? We might have Twitch. What are we doing? Are we live on Twitch? Oh, here we go. This works way better. I need to change the title. Why isn't the title coming up? How's this looking? If you're in the Twitch chat, come and join. How do I get a link to this? I'm not even sure how Twitch works. If you join again, come and say hi. Oh, we're on! Fish on! How does this look now? I got construction going on in my backyard. Everything's happening. YouTube, YouTube doesn't want us. So we on? So we— We on Twitch. I'm gonna share this out. I need to share this out somewhere. Oh, so, uh, I will repost this, don't worry. Let me adjust my face as well. There we go. I think that's a good, a good place. Welcome, welcome everyone. I'm so— I need to upgrade my streaming game. Would you like to see like live streaming, uh, like coding sessions? That'd be pretty fun, like working on, on projects and whatever. I need to just post this to YouTube so people know that what's going on. And because YouTube kicked us off, my old faithful YouTube, what are you doing? Is, uh, is this recording for me? Hold on, I need to check Streamlabs. There we go. Boom. Need to make sure that it is also recording. Yes, so I can repost this later. So let's go here. Oh, everyone, where you from? Let me know where you're from. This is a third attempt. Third attempt. Let me tweet this out as well because YouTube kicked us off. Alright, alright, alright, I think we're on. I think we're on. Refresh chat. Why doesn't my chat refresh automatically in here? Can I send a message? Welcome everyone. Okay, if you have any questions, We are going to get started. How can I view? I'm such a rookie when it comes to this. But thank you so much for joining in. How's the quality of the stream? Is it working? Alright, alright, stuff this. We're just going to get straight into it. Question answering, because who knows when we're going to get cut again. I can't believe this, it's taken us 45 minutes to get this stream working. I promise I will fix this, I'll fix this for a future one. I don't know why. If anyone has an idea of why YouTube was kicking, kicking me off, um, it's never done that in the past. It said I was violating some sort of copyright, but I have no idea. Anyway, Thank you, thank you, thank you to everyone who's joining in, and a big thank you to, um, as you all know, we said this, this is the third time, but it's worth saying again, thank you to everyone who's playing their part with what's going on in the world. Let's have some fun. Okay, welcome, Cass Magic and Code, good to see you again. Thank you so much for joining in. Um, Cass is from South Korea. Amazing. Sweden, third time's a charm. You are absolutely right. It's great. It's working perfectly. Wonderful. Beautiful. That is so good to hear. Okay, what I'm going to do, as we said before, I'm going to switch back and forth between— I'm going to switch back and forth between the chat and the questions I've got saved. So, we'll start off from the top. We'll go in the chat. How about we do that? What have we got here? We'll go right from the top. Cass says, hello Daniel, do I need a GPU to train and deploy machine learning models, or do you have any recommendations for GPU cloud? No, you don't need a GPU to deploy machine learning models unless your inference runs on, on a GPU. Now, but for training machine learning models, it kind of depends on what what kind of machine learning model you want to train. So if you're training anything on unstructured data, such as images or audio or video or something like that, you're definitely going to need a GPU. For example, let me show you, actually, this would be pretty fun. I want to, I want to use Twitch. I want to use Twitch to, let's see if we can get this going. Here we go. I'm gonna come over here on Streamlabs. Watch this. Watch this wizard. Oh, nah, failed on that one. There we go. Here's, here's an example. So this is, this is a recent project that I just built. It's called Airbnb's amenity detection. And so I trained the machine learning model on, on a GPU. Now what it does is it detects furniture in different images. So if you imagine, if you're Airbnb, um, you might want to figure out what features a house has. So come here and I'm going to upload a photo of my bedroom. That's my bed over there. And what it's going to do, it's going to take this image, it's going to run a machine learning model over it, and hopefully this works because it is a live demo. And look at that! Oh, how cool is that? So we've got bed, pillow, stairs. So that's, that's a wrong prediction. That's my bedroom door. And so you can imagine how this goes in an Airbnb situation. But this is kind of a long-winded story of training a machine learning model using a GPU. I trained this model, the model that does this, on Google Cloud's GPU as well as Colab's free GPU. So Google Colab you can use for a free GPU. So if we go to Colab, this is actually what we use in the, in the machine learning course, is we use Colab because it gives you a free GPU. So when you have big datasets of unstructured data, so like data that doesn't come in an Excel type spreadsheet, you'll probably want to use Google Colab. And so here you can go Runtime, Change Runtime Type, and and you can use a free GPU. It's actually a really powerful GPU in Colab, so be sure to check that out. I— if you are, for example, now they do have a time limit on Colab, so if you do want to train machine learning models for a long, long time, you are— you probably want to start investing in your own GPUs or set up some sort of cloud service like Google Cloud or AWS for a longer-term thing. Now, but when you're first beginning, Google Colab is more than enough. All right, this is cool. Let me know if this is, uh, if this is working. Um, but I'm having some fun. Okay, so Cass, thank you for the great question. Now we have— man, Twitch is working so much better than, than YouTube. This is just running smoothly. And you can ask anything related to anything. This is ask me anything, machine learning, life. We got here, um, where we at? Oh, we go back to the Notion. I'll show you my Notion document of all the questions that I have. So let's go here. We'll go to SK. What's your daily routine? Okay. So, this is— maybe I zoom in a little bit here. Can you see that? There we go. That's actually pretty good quality if I'm watching this on Twitch. Okay, so, SK asks from Twitter, thank you so much SK, "What is your daily routine?" Now, I'll be straight up, I try to avoid doing the same thing over and over and over again, because otherwise— now, look, this is going to be a little bit of a story. If you've got 2 or 3 minutes to hear it out, I'll tell you. The reason being is because I don't want to get too stuck into a routine, because when you get too stuck into a routine, if something affects that routine, like the events that are going on in the world right now, you might struggle to deal with that. So, most days I will begin by starting off with this piece of paper. So, see here? I wonder if you can actually see that. Is it going to autofocus on that? Probably not. No, but anyway, let me grab my pen. So today I wrote down the date. I have the first one on the list is livestream at 9 AM, which that was a fail, wasn't it? It's now 10 AM and we've only just started. I've got livestream, collect livestream questions. So this is what you can see in the Notion document. So these are the, these are the questions that I collected. And then I just write out a whole bunch of other things that I have to do. So, that's how I will start most days, with a sip of coffee, usually, usually with my dad and we're having a chat. And what we do, what I do then is I just work through whatever it is there. And so, wake up, get moving, some sort of movement, figure out what I need to do for the day, do that. And then whatever I need to do for the day is usually, usually pretty different. So, but most days at the moment I'm doing some sort of learning, so studying machine learning, some sort of creating. They're the two that are, that are constant throughout every day is learning and creating. And at the moment I'm making sure every day at least to answer as many questions as I can in the Discord chat for my machine learning course. As well as the Udemy QA. So I try to get to them at least once per day and have them all answered. So that's my, that's my routine at the moment. Some sort of creating, some sort of learning, some sort of moving, and helping out the students of my machine learning course. Everything else is whatever, whatever happens during that day happens. Now, where we at? Let's go to the chat. Boom! Magic and Code, you've asked a question a few times. I want to get to yours. I saw it. I don't— I'm not ignoring it, trust me. Do you want to repost it again and I'll try answer that? But anyway, so let's go to the chat. So, Kavlax asks, is it realistic for me to think— actually, I want to put this here so people can see what the question is. So let's go here. I'll get back to this one in a second, but this is— we'll go. Thank you, Kavlax from Twitch, is asking, is it realistic, is it realistic for me to think I can get a job in something data science related from just going through your course and making a portfolio in my free time? So, let's take care of this question. So, Kavlax, I'm going to be straight up with you. I don't know the real answer to that. So, this is what I'm going to be— I'm going to be real straight up with you guys. If I don't know something, I'll tell you. But if I do know something, I'll just— the only way I can answer things is from my experience. So, is it enough to just go through my course? I would say no. My course will give you a foundation of machine learning and data science. What will set you apart, right, from this is using those skills that you've learned in the course to work on your own projects in your free time. So that your second half of the question of asking to get a job in data science by going through the course and then using the the skills you've learned to make a portfolio in your spare time, I would say that is, that is more than enough. Now the most important part there is not my machine learning course. It is you taking what you've learned in something. Now this is what can't be taught in a course and it's why it's so valuable. It's you taking the skills you've learned somewhere and applying them to something. Now this something is quite hard to do because it, it in nature, it has to be something that might not work. Now you might be thinking, well, that's, that's strange, but here's the point. Every project— so the projects in my machine learning course, every project I went through and made sure that they worked before posting them. So all the code works if you just copy and paste it, it all works. But where in reality what happens is you'll run into errors that you haven't foreseen before. And something that— because a course can't contain every single error, right, or every single problem. But what I tried to do, I made sure to leave some errors in to show you how I would search and try to fix them. Now, that is a real skill to be able to do that on your own, because when things don't work, it's easy to be motivated when all of your code is running. It's very hard to go, you know what, everything is broken, I'm facing errors, the problem I'm trying to solve is not working. And then still go forward and try to fix it. That's what happens in your own projects. Now, for example, my Airbnb project that I'm working through at the moment, it hasn't been all smooth sailing. I've run into, I've run into a bunch of different errors, but I've learned a lot on the process and it's almost finished. I'm almost ready to deploy it. So you'll be able to go to a link and run it on, on your phone. Hopefully. Might not be a phone. It might have to be a desktop because I'm not sure of how people can upload things. Anyway, the best knowledge comes from working on your own projects. It's also the most difficult to do because our education system is set up so they discourage you from finding the wrong answer. But real life is just purely figuring out what doesn't work and then going to the next thing, to the next thing, to the next thing. Experimentation. So, What can you do? I would, if I was you and you wanted a job in the data science field, I would use a course like mine to get foundational skills. I would then apply that to some sort of project on my own. You can use my 42 days project principle, write about it, write up, make sure you have some sort of documentation of what you've actually done outside of a course. And then I would get in contact with people and companies of places you'd like to work for. I would start the job before you have it. That is, that is part of, I think it's Section 15 or something of the course, and the basic principle is, is find an ideal role you'd like to have, figure out what that person does on a day-to-day, and then do it on your own accord. And then when you go to apply for that role, you can say, "Hey, I've already been doing the exact thing that this person is going to be doing," and then you can show them that. Will it always work? Probably not, but that's life, right? Not everything always works, but the important thing is to try things that might not work. So hopefully that answers your question. And what have we got here? So we're going to go back to the Notion board. I'm going to click that one off. Boom, boom, boom. Now we might go back to the chat. Back to the chat. I need a drink of water. Magic and Code, I'll probably, um, I, I promise you, I promise you for yours. Here we go. Have you been stuck in tutorial hell? Oh yes, here we go. Let me copy that one. There's a good one. Thank you, Magic and Code. And again, thank you to everyone who's joining in. This was originally going to be on YouTube, and I was going to say thank you for 25,000 subscribers, not that I have any sort of subscriber goal, but boom. Let's answer this one. So Magic and Code says— oh no, this is from Twitch, but YouTube is ditched me today. I feel heartbroken. Have you been stuck in tutorial hell and how did you get out? Everyone says build stuff but would need something more than just that. Okay. Tutorial hell, you're right, my friend. I, I got very stuck in, in tutorial hell. What you said, you're gonna need something more than just Build That, but I'm probably gonna disappoint with this answer because I actually wrote an article that was how to work on your own machine learning projects, and this backs on to the question that I just kind of answered just before, is that all the best things start out as toys. Things that someone's building out of their own interest that they have no idea that's what its potential is. Because as soon as there's a tutorial for something, those skills are commoditized because 1,000 people can go through that same tutorial. What skills will never be commoditized is being able to go through a problem and figure something out on your own. Now, I would in fact ask— think that you can't even— you can— it's like a— it's like a— being able to work on your own projects is like a muscle. So, the two hardest points, what are they? Starting, maybe three. Starting, starting something of your own, that can be exciting. Some people find it really easy to start things. I'm a real easy starter of things. And then the next part, the hardest part, which will come over and over again, is when it gets tough, when you don't know what the next step is. And so, here it's really convenient to jump ship and just start something new. Maybe that'll work out, maybe it won't. And then the third hardest part is when you're like right near the end and you're just trying to— you can see the end in sight, but you know how much you've got to go. And so, there's a saying, it's like, when you're 90% done of something, You've actually got 90% to go. And so, when you're stuck in tutorial hell, what I found, I was— I didn't have the confidence in— when I get stuck in tutorial hell, it's I didn't have the confidence in myself to figure things out. Now, why was that? Because I hadn't done it enough. So, where does all fear originate from? Fear originates from fear of the unknown. Anytime— the reason why a lot of people are afraid at the moment, which is understandable in the current world situation of coronavirus and whatnot, is because there's a lot of uncertainty. We haven't faced these kind of events before, and so we're figuring it out on the, on the way through. But we're gonna— when you do that, when you figure something out on the way through, you come out on the other side. Having learned something that can't really be taught. So that's what I, I implore you to not just build stuff, but to figure things out. That's, that's where the best knowledge comes from. If you, if you're always stuck in tutorial hell, maybe it's time for you to build your own tutorial of something. Where you've worked it out on your own. So, take what you've learned in other places and combine that into some sort of project of your own. So, that's what I'm going to be doing for, for my Airbnb project, is I've taken a lot of resources from other, other, other places, and what I'm going to be doing at the end is combining it all and creating my own version of it. So that if someone wanted to do what I did, they can. But I would never be able to create my own version of a tutorial if I didn't try things out for my own. Alright, let's go on the next one. Where we at? We have— let's go back to the Notion board and we're going here. SK asks, how do you have work-life balance? Now, this is a— let me, uh, let me just tick that one off. How do you have work-life balance? So that's the, that's the question we're gonna, we're gonna tackle now. Now, work-life balance, I thought about this the other day. I actually spoke about it in, in one of my Reps For Honor videos. Was that they each require each other. Now, a lot of Western philosophy is that they should be— a lot of modern philosophy is that they should be separate. I think mine is different. I think they're one and the same thing. Work and life are the same thing. Now, here's the caveat. Of course, there are times when I'm working and there are things I don't really want to do. But that's the sacrifice that I make because that sacrifice that I make enables me to live the life that I want to live. And so, when I see it like that, working through things that are a challenge or difficult or I think are not the way, that most ideal way I'd like to spend my time at the given moment, when I see it as a go, okay, I take on this sacrifice and I get to live the way I want to live. Now, when you see it like that, work and life become the same thing, and then work becomes play because every time I'm going through, I'm like, I'm just jumping between challenges. I'm dancing between these challenges which allow me to live the life that I want to live. So, for me, my work-life balance is that I see work as play. Like, it's a good thing. I'm very lucky with that. And so, what I think is very important for someone to set up their life is to design the life that you would want to live first and then figure out how the work fits into that. I think a lot of people do it in reverse. They see the work they want to live, or the— sorry, the job they'd like, and then they They merge their entire life to work around that. I see it from a different angle. It's like, oh, do I want to have to wake up at a certain time? No. Do I want to have to be anywhere at a certain time? No. Do I want to have an event on my calendar? No. So, I was like, well, that's the kind of life I want to live. How can I design a job or work around that? And now, There are days, and this might sound like, "Oh, Daniel, you see, workers play. You must work all the time." But there are like days that I just go, "You know what? I wrote down this list today and I'm going to blow it all off. I'm just going to do nothing. I'm going to sit around." Like, when was it? Friday? My brother's birthday? Friday afternoon, just did nothing. Sat around, sat on the deck, ate, drank a couple of beers, and then just chilled with my brother. That was a good relaxing time. And I don't beat myself up about it. Now, there is another caveat to this one, is that— or a sidetrack— is that why is it okay? I've been asking myself this question or thinking about it. Let me know what you think about it. Why is it okay for work to interrupt life but not life to interrupt work? And what I mean by that is say you've got a job that's Monday to Friday, 9:00 to 5:00 or something like that. And you, it's like Wednesday afternoon and it's like 2 PM and you're like, you know what, I'm gonna go watch a movie for 2 hours. But that would be seen as frowned upon, right? Because you're like, no, I've got to work till 5, I can't just go watch a movie for 2 hours. But if it was Friday afternoon and it's 5 PM and it's like, oh, we've got this extra thing that we need you to work on, can you stay till 7? That's an extra 2 hours. So that 2 hours is taking away from the life you would spend to spend with your family, why is that suddenly okay? Or more okay? I mean, think about that. Why is it seen okay as work to interrupt life? I think at the moment with technology and whatnot, we've got the curse of the modern individual is that we almost work too much. People forget to have fun. Like you, you, you talk to someone these days and they ask a lot of the default people that It might be different where you live, but I see a lot of people, I run into a lot of people and they ask, "How have you been?" And they say, they almost say like, "Have you been busy?" As if busy is a good thing. And I don't think busy is a good thing. I think busy is a lack of control. Like, it means you don't have— busy to me sounds like you don't have, you don't have control of your time. And so, that the idea of going back to the work-life balance, the full circle. Is I decided I want to live a life where I can have as much control over my time as, as I like. Of course, there are things I have to do that I could go, well, I wouldn't really like to be sitting here for 2 hours doing this, but when I view it from the perspective of going, okay, I make this sacrifice. That's the only choice you get in life, right? You choose what you sacrifice. And so, otherwise, your sacrifices will choose you. So I do the things that I don't really want to do so I can spend the rest of the time doing whatever I want to do. So hopefully that, that clears it up about work-life balance. That was a bit of a long-winded answer, but let's, uh, let's mark that off from the, from the Notion document. And now we're going to go into, we're going to go into the Twitch chat. Here we go. Oh wow. Thank you everyone for joining in and saying hey. What's it— there's a good quote there from Wagon. What's it say? I must not fear. Fear is the mind killer. Fear is the little death that brings total obliteration. Alliteration. Obliteration. Anyway, I will face my fear. I will permit it to pass over me and through me. And when it's gone past, I will turn the inner eye to see its path. Where the fear has gone, there will be nothing. Only I will remain. Frank Herbert, Dune. Wow, that's a good book actually. I haven't read Dune, but I'd like to. What's your favourite book? Leave it in the chat. Mine at the moment is Antifragile. Um, DMT. Oh, this is a great one. How do you know if something is worth pursuing? I got a great— let's, uh, let's talk to this one. DMT Trooper. Twitch to-do. Let me type this out. So this is— I'm leaving this documentation here so if someone wants to come back and watch this video at a later date, I'm going to upload it or something. And this is so that people can see what questions we've talked about and I can timestamp that. That. How do you know if something is worth pursuing? There we go. Now the truth, the truth for this, I could, I could spin out a long-winded answer for this, but the truth is you don't. You don't know if something is worth pursuing because how well, how good is your track record of predicting anything? Seriously, think about it. How, how well is your— is anyone's track record of predicting the future? So, there have been things like, let's take it, the example that jumps to my mind is the Segway. That was thought to have revolutionized transportation, but has it? Like, everyone thought the Segway was going to be revolutionary. Then I don't see many people riding Segways anymore. That's just one example. I mean, you could pick heaps from history. What you must understand and what I need to come to terms with as well, so don't worry, I have this problem as well, is that 99% of effort is wasted. So, the fact is, whatever you work on, Eventually, the nihilist view is that it's going to amount to nothing, which, which in 1,000 years is probably the likely case because that's just how history goes, right? It's just, that's just life. But I view it from the optimistic nihilist view, is that even though whatever I'm working on will likely, will potentially turn out to fail, and if it does succeed, well then that's, that's luck. Right, that's just luck. That's— luck is prep when preparation meets timing. So, if I can be prepared to work on something and the timing of the universe is great and then I show up at that time with my preparation, with my skill, and I decide to work on this thing and it goes boom, whatever boom means, well then that's pretty damn well lucky. But the chances of that happening Can you calculate that? Like, I'm not— I can't calculate that. And so how I deal with this, if something is worth pursuing, it doesn't— what another fallacy that I've sort of noticed is that a lot of things, a lot of people need an explanation for something, and I needed this too, an explanation for doing anything. But I realized after writing, and I wrote an article on this actually, I wonder if we can find that, because I think that's a, that's a good one of my favorite ones of the past, the past few months or the past year of writing. Let's go Daniel Bourke, because is enough of a reason to do anything. Oh, maybe it is here. There we go. Let me bring this over here. 26. I wrote this when I turned 26. Oh no, number 16. Boom! There we go. So if you want to work on something, how do you decide if it's worth doing? Well, if you want to do it, is that enough of a reason? Because is enough of a reason to do anything. A lot of the time people will hold themselves back from doing something because they need a great reason. Would— would— I've— you know, it makes me sad the things that I've missed out on, the opportunities that I've missed out on because I was waiting for a good enough reason to work on something. I was sitting there going, "Oh, should I do this?" What I already knew, I already knew, I already knew deep down that I wanted to work on that and I didn't. Because why? Because I was waiting for a great reason to appear or I was waiting for someone to give me permission and now that's what I want to avoid. It's, for me, something is worth doing if you can imagine yourself in the future 10 years in the future thinking, "I wish I did that." That, that is probably the best reason for doing anything other than because it is that regret minimisation framework. Is that think of yourself in the future and go, "Am I going to look back at this time and go, 'You know what? I had that opportunity. I really wish I did that thing.'" That's enough of a reason. First of all, because it's enough, or are you going to look back and think, did I really want to do that? Okay, where are we at? We're going back to the Notion document. So Abdel Rahman from YouTube asked how to make money as a freelance machine learning, learning engineer. How to approach potential clients. Here we go. Let's go back. I'll grab a sip of water. So how to, how to make money as a freelance machine learning engineer and how to, how to attract potential clients. Now I can't answer this question. The reason being is because I can only tell you what I would do, but you must know that I haven't made money as a freelance machine learning engineer aside from creating educational content around machine learning. My— I, when I was a machine learning engineer, I wasn't a freelancer. I was working at a, at a small technology company in Brisbane, so I had a full-time role. But we did approach potential clients, so I can answer that how to approach clients part. But as a freelance machine learning engineer, you must know that I've only done creating machine learning freelance work by creating educational, because I believe that's my— I'll never be the best engineer. I'm not trying to be. I'm trying to be— every time I've tried to be someone else, I've failed. I'm trying to be me, and I realized, okay, my— what I enjoy doing most is learning to create and creating to learn. And part of that is communicating things in an understandable way. I have a real— I love doing that. And so that is where I use my machine learning skill to, to make an income, is I take concepts or complex documentation and teach it in a way that people can understand. So that's my machine learning course. I made something for DataCamp. I've written articles that have been paid for in the past. I've made videos, you might have seen them on YouTube. But anyway, how would I approach potential clients? The first thing you need is skill. So this is the first thing with anything. This is what makes— this is what gives people imposter syndrome. Imposter syndrome comes from a lack of confidence. A lack of confidence comes in from a lack of skill in something. So you see, that's where it comes from. You can deal with imposter syndrome by getting better at something, because when you're good at something and when you have confidence in yourself, then imposter syndrome goes away. Now, it'll always be there on some level, but it's healthy. It's healthy to have it because it means that you're conscious of your ability. And so how would— how do we approach potential clients? One way, the best way I think, is to share your work online. The reason being is because it's infinitely scalable, you're able to demonstrate your skill, and the best people, the best clients to work with are the ones that come to you. Otherwise, or no, sorry, there may be great clients that you seek out, but just beware, it's good to have business that comes to you. So first and foremost, I would ask you if you want to attract potential clients is, where can I see your work? If you don't have any position, place I can see your work, well then I would say first things first, make somewhere where people can see your work. A website at a bare minimum, a blog bare minimum, GitHub repo bare minimum. Those 3 things, you need that, you need that. Everyone should have their own little corner of the internet. Now, what about outreach? So that's inbound. So then you would just create relentlessly about the things that you're interested in working on and let the internet machine do its work and have people come to you. Now, what about if you wanted to do outreach of potential clients? Now, this is where it takes some consideration from yourself. It's to say, this is what we would, we would do is we would know, we would have a sample of projects that we'd already worked on at our machine learning business. And we would go, okay, we have this portfolio of projects, there's maybe there's 6 of them or something like that. This is where it's important to have a place where you can showcase your work. And then we would search out clients, businesses that were similar to that sort of realm. Who we thought may like our services and we expected to be denied most of the time. So the second caveat would be, or the second point to this, I keep saying caveat for some reason. Number 1 is where's your place to show your work? And number 2 is how many people have you asked? Sheer numbers game after that. How many people have you asked for to offer your services? That's, that's the— I would say, like, for example, let me tell you what I told my brother who wanted to get into photography. So this is, this is almost universal for no matter what skill you want to offer. First and foremost, anything, you need some sort of skill in, in whatever it is that you're trying to offer, because for people to pay you, you need to have some kind of skill. So that's, that's zero. We're talking in Python index terms here, so that's 0. 1 is after 0, once you have some skill, you need a place to showcase that skill. That can be wherever you share your work, some way to demonstrate your ability. And 2 is if you, if no one is coming to you from your, from your demonstrations, you need to go to them. And then I would ask you from here, is, how many people have you, have you spoken to? And I said to my brother, he wanted to take photos for local businesses. He's been building skills, so he has some skill. He takes some great photos. You might have seen them. Some of them are on my articles and some of them are on my YouTube thumbnails. So he does have skill in that aspect. He does have a place to show his work. And finally, I asked him, how many people have you asked? And he said, "None." And I said, "Well, what do you think the answer is going to be for people to pay you? It's always going to be no if you never ask." And so, he said, "Okay." And then he went out and asked 10 people. You know what? All nos. And I said, "Well, you're probably going to get 100 nos." And so, that was— I told him, "That's your goal. Get 100 nos. And then, if you're still getting no, you either need to get better at it is what you're doing, or change the way that you ask. So if you look at sort of anything, 1 in 100 nos is probably— sorry, 1 yes out of 100 nos is probably— what can you control? That's the thing, the main thing you need to consider. What can you control in this situation to get any kind of freelance work, whether it be machine learning or some sort of skill? You can control whether you build skill or not. You can control whether you share your work or not. And you can control how many people you ask. You cannot control whether someone else says yes to you or someone else comes and offers you work. So there are 3 things I would take care of, the things that you can control: build skill, build somewhere to showcase your skill, and then ask people if they need your services. There are 3 things I would take care of. The other things, don't, don't relinquish what is not in your control. So there we go, we go here. How do you potentially approach potential clients? You got to search for them and you got to figure out where your skill matches their needs, and that's where the ask comes from. So let's go here. We'll come back to the Notion document. That was a good question. All right, now let's go back to the chat. Holy goodness, we've got a lot going on. I'm so sorry, I've been missing all this out. I'll quickly scan through here. What do we got here? So here we go. So we have NRE Cobb, NRE Cobb, us. I'm not even sure, RE Cobb. I'm not even sure I'm pronouncing that correctly, but forgive me. Um, N-R-E-K-O-B. Let me copy it into the Notion. Oh wait, oh wait, that's— sorry, that's not really a question, that's a statement. Here we go. Canaan, Canaan asked, Canaan, I— let's go here. I'm going to put that in the Notion board. That way we can, we can keep track of the questions that we have answered. So Kanan says, hey there, Burke. Hey there, Kanan. I've recently started deep learning, but in the process I feel I don't know even DSA, data structures and algorithms, or even sometimes I have even learnt ML. I have not even learnt ML in the case. This demotivates me sometimes. What are your views on this, and please can you answer? Um, it depends what you mean by data structures and algorithms. So if you ask me what are the data structures, I couldn't really list them off for you. And you know, I know there's like things like dictionary and, and list and array and hash table and lookup tables and linked lists and all that sort of stuff, but I don't Honestly, I don't really think of it like that. Or algorithms, I know there's a whole bunch of this. See, knowing them off by heart, all these things, that's like memorization learning. What I value far more is being able to approach a problem and then break it down and then figure out the how along the way. And what I mean by that is how do you look at something and break it down to its essence rather than trying to— and figuring out what tools you need along the way rather than— I'm just going to move this out of the sun— rather than just grabbing, like, trying to fit a whole bunch of solutions to problems. See, that's, that's the engineer's fallacy, is having too many tools and then not knowing when it comes to actually solving a problem, spending more time thinking about what tools they should use than actually using any of them. And so I get this question a lot, and the truth is, how much like data structures do I need to know, or how much algorithms do I need to know? And I honestly, I can't answer that. It's how— how can you answer that? My, my biggest thing is what are you working on right now and what can— what's holding you back from fixing the current problem that you're working on? That, and learn that. That's, that's what you need to know if you're stuck somewhere, because otherwise you just become— you try— that's trying to learn everything, it's like trying to boil the ocean. This is a saying. It's like my mate George told me that saying and it stuck in my head. You try to do everything and you end up dedicating 50% to— 20% to one thing, 20% to another thing, 20% to a third thing, and all of a sudden you spread in 5 different directions. I have that curse. So I'm spread in 5 different directions. What I would say here is that the reason you think you don't know And I, I, this is how I've realized myself, the reason you think you don't know enough about something is because you keep trying to learn new things rather than use the knowledge that you have. And so all you're doing, right, is you're just jumping to the next thing instead of cementing what you've already laid. So it's like you're just adding, adding like you got a plant growing and all you're doing with trying to learn new things is if your plant of knowledge is growing, just jumping to the new thing, continually learning, is just adding more and more fertilizer to that plant rather than just letting it grow. It's almost overkill with knowledge. Your brain just becomes— it's a— we have— so, the curse of learning anything at the moment in especially in the technology field, is that the resources are abundant and they're all amazing, except there's probably too many now. And I say this being the creator of a machine learning course. What is lacking, a severe amount of lack, is people taking what they've learned and build things. And this is from— I say this out of my own like reflecting on my own, by myself, right? When you're continually jumping to the next thing, you feel like you've never actually learned anything because you've never spent time improving the skill of a certain— now it's not of a certain topic. Now it's not a bad thing to want to be interested in different things. I would say it's actually a plus. But where, where we're going at this, you were talking about data structures and algorithms. I would say rather than trying to learn every single data structure and every single algorithm, start with a handful and then build something with that handful and then understand those really well. And then once, once you've understood those really well, move on to something else. Or if you want to keep pursuing that project, whatever you're working on, and that's helping you understand a certain data set algorithm, push forward with that. So there we go. Boop. Thank you, Kanan, for the question. So we're going to go back to the chat— I mean, sorry, to the Notion documentation. This is from Sai on YouTube. Let me know what's happening here. Sai from YouTube, thank you for the question, Sai. We'll grab a little bit more coffee. Is everyone having fun here? Because I can keep going for a little while. As long as you're happy, that's what I want to make sure is that I'm answering the right questions, we're covering the right topics. 'Cause this is what it's all about, is that I'm having fun. So, as long as you're having fun, we will keep going. We might have to take a little break to do some dancing in a second because I've been sitting down for a long time. But Sai says, "You said you're currently running a business other than YouTube, blogs, and your courses. What are you up to?" Too much. I'm up to too much. "What about your idea where you wanted to bring health and machine learning together?" Now, Here's a great example. So this is what we're going to answer, this one, these two. So there's, there's a great indication from Sai that it's, uh, I'm currently running a business other than YouTube, blogs, and your courses. What are you up to? So that's, that's almost too many things, you know. So YouTube, blogs, courses, actually that's probably what I need to limit it down to. And that's, that's really all I want to be up to at the moment is creating and creating. I mean, sorry, creating and learning. Learning to create and creating to learn. Now, I make YouTube videos, like where this is going to go, because I enjoy making videos. I make blogs because I enjoy writing because it helps me think, and I make courses because I enjoy learning and then using that learning and somehow somehow to potentially help others. So that's all I'm doing at the moment, or all I want to be doing. My, my main goal with the business, uh, my main— so you must— I have a few sources of income. I have YouTube revenue, so that's not, not actually that much compared to other sources of revenue. So YouTube ad revenue, I've never actually claimed. It's just sitting in a Google account somewhere, so I've never spent a dollar from YouTube. I have affiliate partnerships with Coursera and Udemy. So that means if someone clicks on a link that I've spoken about and they purchase a resource from Udemy or Coursera, I will get some money back from that. I actually have DataCamp as well. I have an income from Medium's partnership program, and I have, of course, the, the funds that I get from my Udemy machine learning course. The Udemy machine learning course is the number one. So my first and foremost focus is is making sure the students— I'm helping the students in that course. So that is what I'm driven by at the moment, is how can I make— how can I address the needs of my customers better? So that's all I want to be doing is helping students learn machine learning, answering questions where I can. And then of course learning what I need to learn in my, my own sense to create better, better content, better educational content in the future, better videos, and write better articles about whatever needs to be created. So that's, that's my business in a nutshell. Now, what about your idea where you wanted to bring health and machine learning together? Now, this is another a great example of a solution looking for a problem. So, the fallacy was that I learned machine learning and suddenly my hammer, my tool is machine learning. This can be my hammer and now every problem I look at is a nail. What does that mean? Is that I try to think of every single problem in a machine learning sense. So that was, that was the challenge, is that I was, I had this skill of machine learning, or was building this skill of machine learning, and I was also interested in health. Like, health is my first love, you must know that. I have done a nutrition degree and whatnot, but I really, despite the formal nutrition degree, I failed medicine for the first 2 years of studying that, despite that I learned more studying on my own, and that's kind of what got me started studying machine learning on my own. And now, the problem with learning one sort of skill is that you can get stuck trying to solve every single problem that comes into your life with that one skill. So, this is why it's important, I said before, to, to yes, have a wide variety of different ways that you look at things. There's a great book actually called Loser Think, I'm going to recommend that to you actually. Let me just get that up. Loser Think. Here's a book recommendation. So, it says how untrained brains are running America, but what it goes through actually is how different people think about different problems. So, you've got like an engineer, like myself, like maybe you, who looks at the world in a certain way. You've got maybe a lawyer who looks at the world in a certain way. You've got a healthcare professional who looks at the world in a certain way. But this book talks about how— it gives you an idea of how each of those ways of thinking comes about. And so, leading back to the point is that where was I trying to relate health to machine learning is that despite its promises, machine learning might not always be the best solution to a problem. So the number one rule in machine learning is don't use machine learning if you can solve a problem with a simple rules-based system. And so let me give you an example. Um, if I wanted to build a machine learning app that that could help give you information from food. One of the ideas, actually I'm not going to talk about the idea because I don't want to spoil it for myself, because when you talk about an idea and you don't fulfill it, it sounds like you've already done it. So, let me just give you a very high-level overview because I'm actually going to build something over the next couple of weeks. My next 6-week project is to build something with machine learning and health. I'm pretty excited for that. But let me just give you an example. Um, say you wanted to know what to eat for food, and my idea, uh, would be— because machine learning was my, my hammer, right— was to just look at everything from a machine learning point of view and go, uh, this camera on my phone should tell me whether I should eat this meal or not. Now Mm, could I build that? Probably. But could a simple rule-based system be something like, don't eat anything from a packet? Done. That's a rule. Don't eat anything that requires an advertisement. Done. There you go. You've taken care of 90% of your problem with 2 rules with helping people eat better. Don't eat anything from a packet. Don't eat anything that requires an advertisement. Eat single ingredient foods. 3 rules and you've solved 95% of nutrition problems. So, there's the hard part of figuring out how to merge 2 industries, but that being said, the fun part is also figuring out how you can merge 2 things. And so this is where it helps, and this is why I'm biased towards those people who are machine learning practitioners, right? Who— so I don't want a job in research or working at a big technology company. I'd rather let smarter people do those things, and I'm going to take their work and apply it to things where I can. So that's my goal. So that's a long-winded story. Of saying, "I don't know how to do it yet, but who knows, I might figure it out." I've got two ideas, figure it out, well, two skill concepts, one in health, one in machine learning, and they're dancing around my mind at the moment. Who knows when they'll start to tango, you know? So, let's come back here. We will cross these two off. So, we might come back to the chat. To the Twitch chat. Boom! I love how there's a conversation going on in this Twitch chat. Thank you all so much for joining in. What have we got here? We got any more questions? Um, Kavlax, I totally agree. Learning the basics of computer science via Python is arguably the path of least resistance. Let me show you something actually. I want to show you a resource that I found. So speaking of learning more things, why don't we, why don't we learn more things? And what's it called? I need to find it. It's the Missing Component of Your CS Degree. So this is something that I'm probably going to be going through because it's one thing to code in a Jupyter Notebook, but it's another thing to code like an actual programmer. So if we go here, Missing Link. Here we go. Let me open this up. Can you see that? I wonder if I can link this in the Twitch chat. Can you link things in Twitch? Look at this. So classes teach you all about advanced topics with CS, from operating systems to machine learning, but there is one critical subject that's rarely covered, and it is instead left to students to figure out on their own. Proficiency with tools, because if you know how to use your tools, you can use your tools to build things. I'm going to link this in the Twitch chat. We go here, boom, put in there. So check that out. I'm going to also put this in the, in the Notion document. So we, we can go here I'm going to check this out over the next couple of months. Actually, I'm supposed to be in Vietnam right now. Did you know that? I had a— but the whole world kind of got shut down, so I had a holiday planned with, uh, with a couple of friends, and we were going to go to Vietnam from April 3rd, which was 2 days ago, till the 12th. But now I'm gonna— I've been given an opportunity instead to— I'm gonna take, after I finish my ABB project, probably tomorrow, I'm gonna take some time to reflect and think about— just chill out for maybe a week and just think about what I want to work on for the rest of the year. I've kind of got an idea of what it is already, but instead of just jumping straight into the next project, I'm gonna take probably just a week off of just reading, writing, journaling, and and figuring out what I want to do next. But that's enough about me. Let's get back to the questions. We got a question here. Kanan, here we go, down there on Google. Okay, Kanan, great question. So I'm gonna bring this over here. Question number 3, thank you so much. So, Kanan Twitch, so here's the question here. Hey, Burke, I have doubt all this time whenever I take up some ML/ML project down there, here on there on Google Git, but I barely copy-paste the code, or sometimes when I see through and write the same code, it would just be barely like a typing practice. I don't understand how to proceed further, but can you, can you please help? Well, actually, one of the best, one of the techniques I use is literally doing that thing to learn something. Like, you can— here's the thing, right? We're a lot smarter people This is what you need to remind yourself of. People are a lot smarter when they're doing instead of thinking. So look at the animal kingdom. Does an animal sit around and think about whether it's going to run away from a lion? This is a very high-level metaphor, but think about this in your own, in your own life. If you sit around thinking about something all day, how much better at it do you get? Like for example, you could think about, learning to swim all you want. Imagine how many books you could read on learning to swim, and then all of a sudden you hit the water, and what happens? Now, coming back to what you just said, if you want to learn something, don't just sit there thinking about all the code that you can look at on GitHub and, and, and copy— just copy and paste it. Instead, write it all out yourself. Now, of course, yes, you might think it is, uh, writing exercise, but what you're going to do is slowly and slowly absorb the way different programs are written. So that's, that's one way. Let me— I actually took this down notes, uh, notes before, how to learn programming or something like that. Oh, maybe I did, maybe I didn't. I'm pretty sure I did. And maybe it's in— ah, see, this is another one of my problems of using too many tools. I take notes in too many things. Here we go. Anyway, I'll just, I'll just talk about it because I know it, I know it sort of I'll bring it out what comes in my soul. I'll listen to my soul. So, the way I learn things is I think about it, I act about it, I react about it. So, those 3 things. Now, what is that? I think about it. Now, I spend— I try to minimize this time of when I'm thinking because I've realized that once I start to do something, it's so much— I'm so much better than when I'm just thinking about it. It's like why I write, write a lot. Why I write journaling, why I write articles about machine learning. You're a lot smarter doing than you are thinking, and writing is, is translating thoughts into doing. Now, same thing goes with code. You could look at someone's GitHub repo and go, wow, this is easy to fork. Import, import it all, and now start to run it. Or you could take a hard approach, which is a rote learning approach. And I did this the other day. I found a Detectron2 training script and I wrote it out verbatim because I wanted to understand. And what I mean by verbatim is I wrote out all 500 lines of code or whatever by hand, like line by line, because I wanted to figure out. And each time I did, I I annotated what was going on. And now the original code didn't have any annotations. It took me about 3-4 hours maybe to write it all out. But what I wanted to do was kind of reverse engineer what they were using for their training script. And then once I did that, did I understand it fully? No, hell no. But did I understand it more if I just imported it all as one hit? Yes, I understood it way more. And did I start to figure out how I could research where each part came from in terms of when I typed out a single function, looked at the docstring, okay, connected the dots. So it would be like, write out this function here. Okay, what does this function do? Oh, I would find this one. Alright, well, what does this one do? And yes, I was jumping all over the place, but this is again how some of the best knowledge that is, knowledge that is built, it's by following your curiosity and continually asking questions. So, do I advise copying and pasting code? No. Do I advise rewriting code? Yes. Now, of course, there are little snippets and whatnot that you can feel free to copy and paste and whatnot, but when you really want to understand something, start By one, reading it for a little bit. And now I say a little bit because you're going to confuse yourself if you spend too long reading, and I know this because that's what happens to me. Then rewrite it in the way that they did it, and then for a third time rewrite it in the way that you need it. So, think, act, react. They're the three steps that I use. So, great question. Let's tick that off. Bom bom bom. Thank you for that, Kanan. Now, what we might do is come back to the chat, have a look at what's going on on Twitch. We're still sort of open, but I'm self-isolating just because I don't really have to go out, so might as well. Great idea. Yeah, Ricob understands things by drawing things in Paint. That's a great way. It's all about figuring out, it's learning how to learn. It's learning how you best understand things. That's a, see, education resources can only be made in really one way when everyone learns in a different way. So, this is why I stress so much that it is so important to start working on what you want to work on. Education systems, Also, discourage the answer "I don't know." "I don't know," aside from "I love you," are like, "I don't know" is like the 3 most powerful words you can ever say. Why? Because it gives you, not only, it gives you an awareness of what you don't know, which is actually probably more important than what you do know, and it also gives you an opportunity to learn more or to figure out what you need to know. So I feel sorry for the person that never says I don't know because they, they probably don't know more than they think. Now here, I think the most difficult part of doing your own project is not to know if it's good. I don't know if heading straight in the wrong direction. Or if it is quite passable. That uncertainty, as you mention, is scary and petrifying. I totally agree. Uncertainty is scary and petrifying, but that's— Jake, that's a great quote. The context or feedback comes from sharing your work. So that's why Magic in Code actually here is saying, without any context or feedback, how will I know? You start by being your own, own biggest fan and your own harshest critic at the same time. Now what that means is you have to, again, this is balance achieved from operating at both extremes at the same time. You not only have to be your biggest fan, so this is having confidence in yourself, so you're supporting yourself because no one's going to support yourself, you, as much as you. So you need to support yourself, but you also have to— again, this is— no one's going to care about your work as much as you. So you have to ask yourself, hey, is this something good? Is this something— you have to be your harshest critic at the same time. Is this something I would use? Is this something I would share? Is this something I would, I would, I would— if I looked at it, it was someone else's work, is this something I would go, wow, that's pretty cool. So Or could this be better? And of course it can always be better, but that's what— in the beginning when you're working on something that you're uncertain about, you have to be your biggest fan and your harshest critic. Then when you set it out to the world, you must know that when you set your projects or whatever your creations are out into the world, they're no longer yours. So they're no longer you. People who— and people who critique your project They're not critiquing you. They're critiquing your project. So it's— this is what you have to— your role in that sense is you're the performer. And when people give a critique on a performance, they're critiquing the performance rather than the performer. And the ones who critique the performer, well, they're a waste of time anyway. You don't need to listen to them. Here we go. That's— that is— whatchamacallit— that is— the benefit of the internet is Cunningham's Law. So let's go to that. Here you go. This is Cunningham's Law. This is what I tell myself. If I make anything on the internet, I'm taking advantage of Cunningham's Law. Cunningham's Law. Check this out. Cunningham's Law. This will help you make anything on the internet. There you go. Cunningham's Law states the best way to get the right answer on the internet is not to ask a question, it's to post the wrong answer. So all that fear Have you been wrong on the internet? I have it too, don't worry. But it works in your favor. Reason being, if you can handle it, is because you're going to find the right answer. So how's that for a great benefit? Okay, so someone asked another question. Someone said, have I taken that Nissan to Fraser Island? No, but I have taken it to Morton Island. So, that's our family 4-wheel drive. We take that to Morton Island every year. It's closed at the moment though, so hopefully it's reopened by the end of the year. So, where we go here? Let's get back to the chat. Someone asked, "What does knowing a language mean?" I actually don't know, seriously. Like, think about it. Let's think about it in a programming language. Let me think about it in a— or talk about it in a speaking context. Here's another reason to get hands-on with anything that you want to build, because let's think about how do you learn how to speak. Is it from a book of grammar or is it from speaking? I'll wait. So, I tried to learn Japanese and Chinese once, and we did a lot of like rote, like, learning and all that sort of stuff and not as much speaking. And of course I could start to— I've actually probably forgotten all of the things that I— all of the symbols that I learned. And so, did I know the language? I knew it in one sense, and the one sense would be that I kind of knew the grammar, grammatical format and structure and all that sort of stuff. But could I have a proper conversation with someone? Is that— is knowing a language being able to have a conversation with someone? Or is knowing a language being able to write a story in that language? Or is knowing a language being able to read a book in that language? Same with programming, right? Is programming able to read a language? Is programming able to write a language? Is programming able to take any application and build it in that language. So, I would say it's very hard to answer that question, what does knowing a language mean? It comes from a lot of— see, these questions about— I try to avoid questions that can't be— that really don't have— it just can't be answered at all, like in terms of how much should I know about something or how much of a language is considered knowing it. Like, I would say, put it this way, I can write Python, I can read Python, I can code programs in Python, I can build machine learning models in Python, but at the same time, I would say I still don't know Python very well. English, I can speak English, I can write English, I can read English. As in, I could, I could learn more Python, right? But could I recite all of the, the rules of English even though I can read, understand, and write write it? Probably not. Same with Python, right? There's a lot of things that I know about it that I probably, if I really sat down and tried to think about, regurgitate all of my knowledge on it, I could probably think of a fair few things. But there's also a lot of things that I don't know. And so again, it comes back to the point, can, what, with what your current knowledge is, Can you solve the problem that you're currently working on? And then if you can, the next part of that is what is the quality of your solution? And that's, that's, that's tough to measure. So because you can build a crappy solution or you can build a really great solution, and the only really way to figure that out is if it's crappy or, or, or great is to set it out in the wild and see how it performs. Forms, then you'll know. So we go here and Cunningham's Law is a great, a great way to get good answers on Stack Overflow. But let's, ah, haha, someone's writing in Japanese. Alright, now what we're going to do, we're going to answer one more question in Notion. And then we are going to probably wrap this up as long as anyone doesn't have any more questions. I will finish off with this question here and then we will do a little bit of movement and rock and rolling. So let's go here. So Grundet11 from YouTube asks, "Hey man, I won't be able to watch it live." That's alright. The livestream got, got a bit twisted, so not many people can watch it live. So, have you spent any money in the 42-day project? So the 42-day project he's been talking about is the 42-day project I'm working on replicating Airbnb's object detection. The answer to that is yes. So I spent about $150 on a GPU training a machine learning model for 18 hours. So, Colab. Oh, did I use Colab for free? Yes, I used all the free resources I could until I couldn't. So, this is another part of working on your own projects, right? Is that, yes, do it for free as much as you can, but one real way to tell if you're actually into it is, are you going to invest financially into whatever you're working on? So, that's a real, a real way to tell if you're taking something seriously or not. Now, of course, some people might be limited in that aspect, but it's all about thinking about what can you— if you are dedicated to something, what does that mean to you? And for me, it means skin in the game in my project means if I need to, if I'm being held back by not dedicating certain funds, well then, I'm not really committed to it. Could I have trained it all for free on Colab? Yes, I probably could have, but Colab's timer times out on 8 to 12 hours, so I would have had to train it, wait for the timer to go off, train it again, wait for the timer to go off, train it again. Now, that wasn't really something that I was— that would have probably taken me more time than just paying for the fee. And because that is part of my business, is working on, is learning to create and creating to learn. It's a business cost to me. So the answer to that is yes, I have spent some money. It's about $150. I'll break it down, uh, towards the end of the project, how much the— how much it cost, what the business cost is. Um, but it's still cheaper than what Airbnb would have spent, so put it that way. And let's go here. And one more question to finish off in the, in the Twitch stream. Let's, let's take this off. Let's take this off and let's go here. Kanan says— here we go. Kanan says, do I listen to music while working on projects? And I would say I used to a lot. But now I don't really. Now I just, just, just, uh, I kind of go back and forth, right? So I listen to things and when it's distracting, when it's distracting, I stop, I stop listening to music. I use Brain.fm sometimes, or I go to YouTube and search, um, just, uh, like binaural beats. And listen to that, but I don't— it confuses me when I listen to music with lyrics and code at the same time. So, the music can't have lyrics. It's just like background white noise and whatnot. So, but at the moment, for the last couple of months, I haven't really listened to anything while I'm working. I just code away and think about different things and stare out the window and whatnot. But Here we go. Oh, Kanan says question 4. All right, question 4. We're doing one more question and then we're wrapping this up. This is from Twitch as well. And so, what are your views on rushed learning, like a DLML course on Udemy, put the videos on 1.5 to 2.x, completing a course? Okay, I love this one because I used to do that shit, and don't do that, right? Take your time. Learning anything worthwhile takes time. Yeah, sure, you can put a course on 2x videos and complete more courses, but are you going for— what are you going for? Are you trying to complete more courses and get more certifications, or you're trying to build skill? I mean, I could sit on Coursera with autoplay on, Udemy with autoplay on, all on 3x speed and complete a 15-hour course in, in 5 hours or whatever it is. But did I learn anything from that? So I would say remind yourself, learning any skill takes time. And what are your goals? To just go burn through more courses in a short amount of time, or to actually do the hard work of building a skill? Watching videos on 2x speed and not really paying attention to them because they're going too fast for you to comprehend is easy. It's an easy way to get through videos, but taking your time and being patient and trying to learn something, that's hard. So, if you take shortcuts, you're going to get shortcut results. And the reason why I can say this is because I used to do that, and it— and that's all I would do is just jump, jump down, jump from thing to thing and complete And I thought, I mistook, I made the mistake of thinking completing more courses meant more knowledge. When it was in fact almost the opposite. I was just, the more I completed, the more I was just, say you have a tower of knowledge, let me draw this out. A plant of knowledge, right? All I was doing, can you see that? Not really? But let's say here, this is your knowledge stack and all I was doing was just trying to add on layers and layers and layers of knowledge. You can't really see that, but put it this way, I'll just, I'll just talk about it. You have a stack of knowledge. All I was doing was trying to layer on layer on layer on knowledge of doing more courses on 2xSpeed. And all I did was I just weighed down. So all of the knowledge that I'd built above ground was just getting pushed down into the ground and all of a sudden, All I knew were these things that were the layers that were going on top and I was forgetting all the fundamental things that I'd learned before. So, my root system was badly affected. And so, do not watch videos as fast as you can. Learn what you need to learn when you need to learn it at your own pace. Learning something new by definition is hard and that's what separates people who do mediocre shit The people who do great shit. The ones who, like me, who watched videos on— watched ample amount of videos on 2x speed and don't learn a thing because they're just— all they want to do is just grab a certificate and share that on LinkedIn. Don't learn very much. The ones who watch— I would rather someone sign up to my course, watch 2 hours out of the 42 videos, 42 hours, and then take that 2 hours and and then spend 40 hours, the rest of 40 hours working on their own project. That is seriously, I would rather less people get certificates from my course than to watch 3 hours of videos and go, you know what, I've learned enough here, I'm going to go use this, this knowledge. Because skills are worth more than certificates. So there's an answer to question 4. Don't rush learn. Anything you rush I've never met a happy man who's in a rush, right? What is the rush? These skills are going to be around for a long time. Let's just, uh, let's just put this here. I can't— for some reason Notion isn't turning this into a to-do, but we'll mark that off. That's the last question. What do we— thank you all so much for joining in. I had a great time. Thank you for your patience. I hope you all had a great time. Please, please leave a message in the chat if you have any feedback or send me a message somewhere. You can find me on the Discord channel for the machine learning course, or you can message me anywhere @MrDBurke, or you can find— you can— you're smart enough to figure out how to contact me. If you'd like a question answered in a future stream, please let me know. We usually do these on YouTube, but for some reason YouTube kicked us off twice today, so we're doing it on Twitch. I'm going to re— this video is being recorded on Twitch, so I'm going to repost it to my YouTube channel so you can watch it again. Or if you're watching this in a future date, please leave a comment below of how you think this could be improved, or if you have a question that you would like answered. But what I'll try to do I would like to do these at least once a month, and I think I missed last month because just too many things going on. But I do, I do like to do these at least once a month. So with that being said, another big thank you to everyone who's playing their part with what's going on right now, the coronavirus. My appreciation could not be, and gratitude could not be higher for the people who are on the front lines, the people who are working supermarkets, the people who are doing deliveries, who are doing the things that they need to do to keep society running. My heart goes out to you all. I thank you, thank you kindly. And to everyone who's joined into the stream, I thank you, thank you, thank you very much. If you're new here, welcome. And if you're coming back, I appreciate you returning. Hope you've had a great day. Enjoy the rest of your time. It is a beautiful sunny Sunday here. I'm going to go for a— probably do a little bit of movement, a little bit of food, and a little bit of exercise. Size later. But, um, as always, keep learning, keep creating, and I'll see you next time.