I'm Starting a YouTube Series Where I Code With Zero AI. Here's Why That Feels Radical in 2026.

Developer Alex Cloudstar is launching a YouTube series where he builds software entirely without AI assistance, aiming to test whether his programming skills have eroded from over-reliance on AI tools. The experiment, documented on his channel alexcloudstar, will involve planning, design, and coding from scratch with no AI help, and he plans to share behind-the-scenes content on X.

A few months ago I caught myself staring at a merge conflict for a solid two minutes, completely stuck, before I remembered that resolving merge conflicts is not actually hard. I had just forgotten how to do it without a suggestion popping up first. That moment stuck with me. I've been a full-stack developer for years, running my own company, shipping products, writing code every single day. And somewhere in the last couple of years, a huge chunk of the actual thinking got outsourced without me really noticing. Autocomplete finishes the sentence. Autocomplete finishes the function. Autocomplete finishes the architecture decision. At some point you're not writing code anymore, you're proofreading someone else's. I'm not writing this to dunk on AI tools. I use them, I ship faster because of them, and I'm not going back to typing every line by hand as some kind of purity test. But I noticed a real erosion happening, and I don't think I'm the only one. Ask a developer under 30 to explain what's actually happening in a piece of generated code and you'll get a shrug more often than you'd expect. Not because they're not smart. Because the muscle never got built in the first place. The Bet So I'm running an experiment on my channel. The premise is simple: build real software, start to finish, with no AI involved anywhere in the process. Not for the writing, not for the planning, not even for a quick "what's this error mean" lookup. Just me, a text editor, and whatever documentation existed before 2022. A few ground rules I'm holding myself to: No AI, full stop. Planning, design, and code all come from scratch. No tooling lock-in. One day it might be Neovim, the next it might be Zed. The point isn't the editor, it's the thinking. Daily-ish uploads, but I'm treating this as a long run, not a sprint. Burnout kills more content series than bad ideas ever do. I'll be posting behind-the-scenes bits of each episode on X, and the full builds will go up on YouTube under the channel alexcloudstar https://www.youtube.com/@alexcloudstar . Why This Matters Beyond Content Here's the thing I actually want to find out, and it's bigger than "can I still code." Programming skill has always been about holding a problem in your head long enough to see its shape. AI tools are incredible at generating solutions, but they're just as good at letting you skip the part where you actually understand the problem. That gap doesn't show up on a good day. It shows up when the tool is wrong, or down, or when you're debugging something novel at 2am with no clean pattern to match against. I don't have a tidy conclusion for this yet. That's kind of the point of running it as a series instead of writing a hot take. I want to see, honestly, whether a few months of deliberate no-AI building sharpens things back up, or whether it turns out the muscle is gone for good and has to be rebuilt from zero. If you're a developer who's felt that same low hum of "wait, could I actually still do this without the assist," I think you'll get something out of watching along. And if you've got a language or a project idea you want to see tackled the hard way, I'm taking requests in the comments. The trailer drops today, with the full series coming soon after. To catch it: Subscribe on YouTube: alexcloudstar https://www.youtube.com/@alexcloudstar Follow me on X https://x.com/alexcloudstar for behind-the-scenes of every episode Let's see how rusty I really am.