I have a real soft spot for The Coding Train, because it does something that almost no other resource manages, which is to make the act of programming feel like the most fun thing in the world rather than a grind you have to survive on the way to a job. Daniel Shiffman is the entire secret here, and his teaching style, all excitement and happy accidents and thinking out loud as he debugs live on camera, turns concepts that would be dry in a textbook into little adventures you actually want to follow to the end. That matters more than people give it credit for, because the biggest reason beginners quit is not that the material is too hard, it is that it is boring or intimidating, and Shiffman removes both of those obstacles at once by being so obviously delighted by the thing he is showing you. Where the channel really earns its place on a site about AI is the creative coding and machine learning material, because when Shiffman walks through building a simple neural network from scratch, or plays with ml5.js to do pose detection or image classification in the browser, he gives you something a lot of formal courses fail to, which is a genuine gut level feel for what these systems are doing rather than just the equations.
You watch the network learn, you see the art it generates, you break it and fix it, and that visual intuition sticks in a way that a wall of notation often does not. For a nervous beginner, or for someone who found maths heavy ML explanations alienating, this is a wonderful on ramp that quietly builds confidence. I do have to be honest about the limits, though, and they are real. This is not a structured curriculum with a clear beginning, middle, and end, it is a vast and slightly chaotic library of explorations, so you have to supply your own direction and discipline, and people who want the most efficient route to a specific job skill will find the tangents and the exploratory pace frustrating rather than charming.
The stack is also mostly JavaScript, p5.js, and Processing, which is lovely for visual work but is not the Python, NumPy, and PyTorch world that almost all serious AI and machine learning careers actually run on, so the skills do not transfer as directly as you might hope. And because the deep learning content leans on the friendlier, more visual corners of the field, it is light on and in places behind the modern state of large models and current tooling, which is understandable for a channel built around creative coding but worth knowing before you rely on it as your main AI education. My honest take is that The Coding Train is best understood as the thing that makes you love programming and gives you real intuition for how code and simple learning systems behave, and it is genuinely excellent at that job, which is not a small thing. What I would not do is mistake it for a career track.
Use it to get hooked, to build confidence, and to actually understand what is happening inside these systems instead of treating them as magic, and then, when you are ready to get serious and employable, move on to a structured Python based machine learning path while keeping Shiffman around for the sheer joy of it.