Learning to code has changed

Key Takeaways

Learning to code has fundamentally changed. It's a lot different now than it was even one or two years ago. And even myself, as someone who makes long- form tutorial content, has been changing my perspective on what the most effective way is to actually learn this technology and what's important and what you can completely skip. So, in this video, I want to break down how I'm thinking about learning to code in 2026, what I'm doing personally, and what I would recommend to you if you're someone who's trying to get into this industry. Let's dive in. So, I want to start by kind of reminiscing on the old way of learning how to code, which some of you may be using right now. And I think a lot of you watching this video are going to remember if you've been in this industry for three, four, five, you know, six plus years. When I started learning how to code, it was 2012. Back then, we didn't have AI tools. AI wasn't even really a thing yet. We hardly had machine learning. We didn't have, you know, millions of YouTube videos by thousands of different people teaching different frameworks. We didn't have interactive documentation. And if you wanted to learn something, you pretty much just had to sit down, scour the internet for any decent resource, and that's kind of all you had. I remember for me, I was learning from channels like The New Boston. You guys have probably heard of that on YouTube. It's really the OG in the coding space. People like Sendex, which have been around for, you know, 10 plus years. And even looking at like example projects or just random, you know, crappy websites to try to learn something. And a lot of the focus back then was actually about really understanding the syntax of the language. what is a for loop, what is an if statement, what can you do inside of HTML, you know, what is a metatag, what are these types of things? And being able to kind of confidently sit at a code editor and generate some code without having to purely, you know, read it off another screen or look up every single component. We're going to Stack Overflow. There was bugs that would take literally days to solve. And to generate what you can do now in AI in 5 minutes could literally take 3, four, 5 days, sometimes a week long. Now, even though it's objectively much easier to learn how to code than it was 3, four, 5, 10 years ago, I think because of that, a lot of people are really losing the fundamentals. And I actually say that I'm really fortunate that I got into this industry before AI took over because I was really able to become a strong, competent developer based on the absolute fundamentals and actually experienced writing literally millions of lines of code before I started crutching on AI tools. Now, with that said, I absolutely don't miss the endless grinding to get a basic application up and running. But if you're a beginner today or you're someone who is trying to learn some of those fundamentals, a lot of problems are popping up which I want to dive into now. So, over the last 6 months, I've personally worked with over 200 developers through my program, Devaunch. Now, in that program, I'm able to really understand kind of the key competencies of these developers, what they're good at, what they're not good at. We run them through mock interviews, all of that kind of stuff. Now, I say that not to advertise it, but just to give you some context, that by working one-on-one with people, it gives you a real perspective into what they're doing right now and where their kind of strengths and weaknesses are. And what I've seen, especially with more of those kind of junior, even mid-tier developers, is they've really lost the ability to think critically. Now, this is something that has always been the most important skill when it's come to software development. And without AI, you really did have to think quite deeply even about relatively simple projects. You'd have to come up with an architecture. You'd have to solve, you know, all of these different problems. How do I loop over this list? How do I store this state? How do I back this up? How do I connect to the database? Everything you did required a lot of brain power, a lot of searching, a lot of figuring stuff out. Eventually, it would become second nature, but at least at the beginning, it was really difficult and there was a lot of problem solving you had to do. Now, with AI, if you don't know something, you just ask it and it can just do it for you. It can set up an entire project. It can write a ton of code and you kind of get this feeling where it feels like you're getting something done. You're building something, but you don't really have any idea what's going on. And most importantly, you're not exercising that critical thinking skill, which I believe is what makes someone a good developer. So a lot of the people I work with today, especially if they started learning this during this kind of AI era, they have that issue, right? Where they have kind of this confidence because they've built some projects, but they don't really have any ability to think critically to compare different alternatives to explain why they made certain decisions because the truth is they didn't make decisions. They outsourced their thinking to an AI model and they simply prompted it with a little bit of an understanding based on, you know, four hours of a YouTube course they went through. And to be honest with you, I don't blame these developers. If I was just learning how to code as a complete beginner, I would probably lean into that, too, because I can just get things done so quickly. And it's just speeding up, you know, so much of that hard work. But that hard work is also something that, at least for me, really got me into coding in the beginning. you know that really rewarding feeling of spending hours working on a bug and solving it and always remembering that solution in the future or spending, you know, three weeks working on application to have it finished and having something you're really proud of. That's what made me start to get more confident as a developer to really get passionate about the field and to want to write code, you know, to want to do this as a job and to walk into an interview like, "Yeah, I built this. I'm really proud of it. Here's the things I did. Here's why I did it." the devs that I work with today, most of them I just don't see that anymore. I'm missing a lot of that passion that I remember having and again a lot of those fundamental skills where, you know, even midlevel devs are struggling to write basic code. They don't know a for loop. They don't even know what type you're supposed to use in a language like Java or C++. And it's just mind-boggling to me. But again, I get it. Now, with all that said, not all of this is a bad thing, and not every developer is falling into this trap. I have worked with many developers who don't have these issues, who actually did start learning in the past year or two and have made an insane amount of progress because they've leveraged things like AI tools effectively, right? They've kind of learned how to balance themselves and to make sure that they're actually picking up the important information that they need while being able to skip some of the more monotonous things like memorizing every piece of syntax but still getting the fundamentals down. And overall, the trends that I'm kind of seeing is that the people that use AI effectively in their learning, but don't crutch on it, are absolutely skyrocketing in terms of the amount of information that they can learn and how good they can get. But the people that completely lean into it and really aren't able to do anything on their own are just left behind further than they were before. So, what I've been seeing is just this massive divide between the good developers and the bad developers, and it's just expanding rapidly as AI gets better and better. So, with all that in mind, let me go over how I would actually go about learning how to code in 2026 considering all of these factors. Now, the first and most important thing that I think a lot of people do actually skip is just having a goal in mind. You know, if you just want to get a project out there, if you want to have a SAS, if you want to build a startup, you probably don't need to grind the fundamentals. You don't even need to learn data structures and algorithms. There's a lot of topics you just don't need to do because you're purely focused on the output, right? Or the outcome. And if you get to a point where you know you have a million-dollar startup, you can hire some developers, then know about those things already. However, if your goal is to land a job, if you know that you're going to have to pass, you know, rounds of technical interviews, the way in which you approach this should be different. So, I would really challenge you if you're watching this, ask yourself, what actually is my objective when I'm learning? Do I just want to get something out there? Do I just care about the outcome? Or am I doing this because I want to build a long sustainable career? Now, after you answer that question, that's going to help you determine what to do next. Regardless, for everyone, I do really recommend having some kind of structure or road map. A big mistake that I see a lot of people making, especially now when there's new things coming out every single day, is they just jump between so many different technologies, they think they're going to be an AI developer, then they're going to do front end, then they're going to do backend, then they're going to do cloud, and they never build any depth in one particular area. They might be this kind of general developer who's built a bunch of cool stuff, but if you don't have any expertise in one particular area, you're not really that useful, especially when anyone can kind of be decent in all of these fields by utilizing AI. So, I recommend creating a structured road map and a plan that outlines first of all again what the goal is and how you plan to achieve that over the next 3, four, 6 months. There's all kinds of resources and videos online. If you were to work with someone like me inside of Dev Launch, I would provide that roadmap to you. my team would build it for you so you know exactly what to do. But the point is you need some structure so that when you wake up in the morning, you know what it is that you're going to be working on. Do we need to work on APIs today? Do we need to work on databases? You know, how many projects should we have? Just knowing what to do is literally half the battle when it comes to learning how to code. Now, the next component, which I think is increasingly more valuable, especially today, is constant evaluation. You don't want to fall into the trap of feeling like you're learning because you're using these models, you know, to generate these cool projects or you're getting some interesting outcomes, but you don't actually really know what's going on. Now, before if you built a really cool project, you know, in 2017, you could say like, I'm a developer, right? I know how to code because I built this whole thing and I actually had to write the code by hand. But now in 2025, 2026, it doesn't really matter what you're able to generate unless you can actually prove that you did that and that you're the reason why this is the outcome, not some random AI model. So when I say constant evaluation, I mean actually testing yourself much more than you would have in the past to understand, do you actually know what's going on? So ways that I like to do this is obviously following, you know, like a structured course where it would have quizzes or assessments or projects you do actually do on your own. But if you're just going to work on this, you know, self-paced, going to AI, asking it to quiz you, asking it to give you, you know, multiple choice questions, mini projects, code snippets you have to write on your own, going through, you know, files that maybe an AI model is generated and seeing if after reading it, you could reproduce that. Turning off the internet on your computer and trying to write even, you know, 10 minutes of code by yourself. I think that really exposes where the gaps in your knowledge are. And just the more you can evaluate yourself to get an objective metric of what you know and what you don't know, the better off you're going to be and the less surprised you're going to be if you get into an interview situation and they start quizzing you on things that you genuinely don't know anything about. And then the last step is really just utilizing AI tools. And I know this goes without saying, but as much as it's great to write code on your own and you should be writing lines without using AI, not crutching on it, you know, it would be stupid to say don't learn these AI tools because this is where the future's going, right? prompt engineering, MCP, all of these things are really important topics to know and you definitely shouldn't skip them, but you just want to make sure that you're balanced in terms of picking up the fundamental skills, actually deeply understanding what's going on, learning about things like system design, making decisions, you know, questioning the AI models, but also learning how to use these tools to increase your productivity and efficiency and your learning. You know, AI is great because it's literally a coach and a tutor sitting beside you, but you have to use it like that and not just kind of cheat with it effectively because ultimately you're only cheating yourself. Now, I know it's not the most concrete road map in the world. really depends on what you're trying to do and again what your goal is but generally speaking I think that the people that do follow this kind of plan and really are trying to challenge themselves ensuring that they do deeply understand what's going on and questioning the things that are being generated by things like AI models are much further ahead than those that don't and if you use this approach you know fortunately you can learn to code probably five times faster than I was able to because the access to information is just so much more instant and it's less searching and more can I really comprehend this and retain this information long term. Anyways guys, that's all that I have for you in this video. If you are focused on learning, then a great platform that I do recommend is the sponsor of today's video, Brilliant. Brilliant is where you learn by doing with thousands of interactive lessons in math, data analysis, programming, and AI. They adopt a first principles approach, ensuring you understand the why behind each concept. Every lesson is interactive, engaging you in hands-on problem solving, which has proven to be six times more effective than simply watching lectures. The content is developed by top-notch educators, researchers, and professionals from renowned institutions like MIT, Caltech, and Google. Brilliant emphasizes enhancing your critical thinking abilities through active problem solving rather than memorization. As you learn specific subjects, you're simultaneously training your mind to think more effectively. Consistent daily learning is crucial, and Brilliant makes it effortless with their bite-sized lessons, allowing you to acquire meaningful knowledge in just a few minutes each day, which is perfect for replacing idle screen time. Additionally, Brilliant offers a comprehensive range of computer science and Python courses, as well as extensive AI workshops guiding you from a complete beginner to an expert through practical hands-on lessons. To learn for free on Brilliant, go to brilliant.org/techwithtim. Scan the QR code on screen or click the link in the description. Brilliant has also given our viewers 20% off an annual premium subscription which gives you unlimited daily access to everything on Brilliance. [music]