How I keep myself on track | 100 Days Of Code 8

Day 32 of 100 Days of Code. Time to start the day off with a cold shower in the middle of winter. Wish me luck. What's good, y'all? Still day 32. I finished my quiz on machine learning. I did 3 attempts, could only finish at 80%, so 4 out of 5. Um, been doing a whole bunch of mindless tasks the last couple hours, like cleaning my email inbox. Updating my blog, doing a whole bunch of, I don't know, just shallow work, not deep work, not too much study, not too much code, which is what I'm sort of, I don't know, trying to think, how can I not put as much time into the shallow tasks? But then again, I guess you need a balance, right? Like if I want to have my blog and stuff looking good and my answers on Quora and my Medium and whatnot accessible, these are the things I have to do. So, Hmm, I'm not sure. But I'm also thinking about doing a podcast later this evening or tomorrow sometime. I think today might be a write-off in terms of any more tasks. My brain is a bit, I don't know, shutting down. I'm getting a bit annoyed and stuff like that. Well, I don't know. How do I explain it? I'm complaining. I'm complaining too much. But I got dodgeball later, so that's going to be really fun. I'm going to go have dinner with one of my best mates and yeah. Trying to check off all the things on my whiteboard. I cleaned down here. It was just too much stuff. Too much stuff going on and I want to get back to this. I want to get back to this. Yes, machine learning, deep learning, code, workouts, writing, AnyGym people. No to everything else. I'm really enjoying writing and I'm thinking about changing my blog from Squarespace to WordPress, but that's a whole nother topic. Um, but what else? I can't remember. I'm gonna give technology a break for the rest of the night, I think. Maybe I'll do some writing later, but that's, that's about it. And then I'll get back into study tomorrow. I think tomorrow is a deep learning day. If not, you'll know anyway. So we'll see you in a few. Day 33, 100 Days of Code. What's going on, y'all? I'm doing some Udacity transfer learning this morning. Uh, sorry, Doing deep learning study via Udacity, learning about transfer learning. That's correct. That's my goal today. I want to get fully up to date with the Udacity deep learning course I'm doing, um, and then if I have some free time later this afternoon, I'll do a whole bunch of other, uh, I know, less brain-intensive tasks that I want to get done, like making YouTube videos, writing, canceling a few tasks, emails, etc., etc. But main goal for today, just get all the Udacity stuff done. And yeah, that'll be, uh, that'll be it. But I just had a cold shower, and it's winter and it's raining outside. I don't know if you can see that very well, but it was hard. That's the hardest thing I'll do today, that cold shower. Cold shower, 60 seconds. Try it if you want to, I don't know, increase your willpower. Have a cold shower every day. That's what I'm trying to do. Um, Not to mention all the major benefits that cold exposure has for the body— immune system, weight loss, etc., etc. You can Google all these. But yeah, try it out. Even if you just have a warm shower and then at the end turn the hot tap off, stand out of the cold water for like 5 seconds, 10 seconds, and then slowly build your way up. Try it out, I dare ya. So, learnt all about transfer learning today and deep learning. What is transfer learning? Well, it's the part of machine learning which gets you to sort of apply one set of knowledge you've learned on one topic to another section. So for example, what's an example in real life? Well, for you and I, let's just say we know the difference between a good apple and a bad apple, right? And if we take that knowledge from good apple and bad apple, we can tell the difference between a good potato and a bad potato. Well, not exactly the same thing, but similar. And that's the same with, with deep learning and with machine learning. They transfer Transfer learning is taking— so if you were to train a massive neural network on a big dataset, it could take weeks on hundreds of GPUs. So instead of doing that, why not take a pre-trained neural network like VGGNet or, yeah, VGGNet or something, or ImageNet, that has already been trained, take that information and use it on your own dataset. So it's just applying one, one already pre-trained neural network to another dataset that wasn't particularly the same as the original dataset it was trained on, but is somewhat similar. And I think transfer learning is what Andrew Ng believes is where the next frontier is in deep learning. Because yeah, that's what we do. That's how we learn, right? And that's what deep learning is based off. It's based off how the human brain works, how one neuron talks to another neuron, talks to another neuron, and it comes out in a thought. So really cool stuff. I'm still at the base level of it, lots more to learn, but tomorrow I'm gonna be getting into some Python as well as doing a bit more creative stuff. But yeah, we'll see what tomorrow holds. For me, it'll be about 8 to 12 hours. For you, it'll be a couple of— Day 34! It's the 14th of June, 2017. What is going on, y'all? Check out this weather out here. It's been raining the past few days at my place. It's like sunny and raining at the same time. I got my favorite combo. I just had a cold shower, but this coffee is going to warm me up. I've got more here. I'm about to learn a whole bunch about object-oriented Python on Treehouse. Oh, and we got a YouTube video uploading nice and slow on the Australian internet, but I'm not complaining. At least we've got it. Um, yeah, that's my main plan for today. All Python, and then I'm going to work out later this afternoon. But that's what it is. Study is just Groundhog Day. Same thing, same thing every few days. So machine learning, deep learning, Python. And, uh, yeah, hopefully some of this knowledge retains in my head over the next few months because I don't know, what do I want to do with it all? Well, hopefully make the world a better place. How do I do that? I'm not entirely sure yet, but I'll work it out. See you soon. What an epic day of programming. Got some Python done, Treehouse, Python the Hard Way book, cleaned out all my emails because I had a whole lot of back— well, actually, that's a lie. I still got a lot more to do, but, um, no, I had a really good day. I worked out in the middle of today instead of at the end. I think that was good to break it up. So I did a few hours of study in the morning, and then I went and worked out with some friends, saw them, had a nap after working out, and then did a whole bunch of study. In the last few hours, mostly on Treehouse and mostly on the Learning Python the Hard Way book. I'm up to chapter 42 as of today, and it is the end of day 34. I've had enough time in front of a computer today, so I'm going to shut everything down and start to relax, have some dinner, and I don't know, I think my friend's coming over. We're going to watch a movie, and I think it's actually to do with some sort of artificial intelligence, which is exciting. Um, oh, speaking of that, there's a research job going at Oxford University that I'm going to apply for. I'll, yeah, I'll tell you more once I've actually applied for it, I think. Wish me luck. This right here, dangerous place. I nearly got caught in the nap trap. I had a 20-minute nap and then went for another 15 to 20 minutes or so after waking up to my timer and pressing stop without actually realizing it. Um, maybe I needed more sleep. I don't know. I don't think so. I slept enough last night. But what I've been up to, I've been up to do machine learning. It's week 7 of the machine learning course I'm doing on Coursera. So far so good. I've got to do the programming assignment for the week and then I'm done until week 8. Maybe get a start early on week 8 instead of, uh, right after the deadline. Um, but also been learning about blockchain and cryptocurrencies and whatnot on YouTube and TED, just watching, trying to learn about these new technologies coming out. Apparently blockchain is going to be as big as the internet itself and I don't know enough about it to sort of Well, I know enough about— I know more about it than I did a couple weeks ago, but I don't know enough about it yet to sort of see how much it can replace. Like, the internet replaced a whole bunch of distance. Like, in— for me to communicate with someone across the world now is so much easier. Blockchain, I think, will replace trust. Well, not replace, but certainly bring trust to a new level, whereas now we depend on intermediates like middlemen to sort of trust to give our things value and prove that, I don't know, we're not lying to each other. Like banks and stuff, like, hold the value of currency. Blockchain will replace that. I don't know, as I said, I don't know enough about it yet to sort of explain it and give, I don't know, give examples of where it could be used, but I will learn more about it and I'll let you know about it. Otherwise, check it out for yourself. Maybe you could teach me something about blockchain. But for now, I'm going to do my machine learning assignment I'll check back in when it's finished, if I can finish it this afternoon, but I will. I'll make sure I do. And we'll see you in about 3, 2, 1. I passed! Machine Learning Week 7 assignment done. Still getting used to using MATLAB. I mean, I'm, I'm still a novice in Python, and I think to me MATLAB is a bit more confusing. There's a lot more going on, but I essentially programming is all the same sort of stuff, right? Ideally, you want to take an input, do something with it, create an output, solve a problem. That's, that's what programming is to me. It's just a whole bunch of telling, using commands to get a computer to solve a problem for you. But that's me done for the day in terms of coding. Done Machine Learning Week 7, that's over. So now I'm gonna do a few other little tasks. I might do a podcast, I'm not entirely sure. Still got to work on that Oxford application. Maybe that's a task for tomorrow, but we'll see you then.