I used to make developer videos. Tutorials, walkthroughs, the usual. Then I stopped. Not because I ran out of ideas, but because the landscape changed — and I needed to figure out what was still worth saying.
What Changed
Around 2022, something shifted. The content that used to take hours to produce — step-by-step tutorials, framework introductions, "how to build X" videos — became trivially easy to generate. Not by me. By AI.
ChatGPT and its successors didn't just answer questions. They answered them well enough that the value of a 20-minute video explaining how to set up a Nuxt project dropped to nearly zero. Why watch me do it when you can ask a model and get working code in seconds?
I didn't stop because I was afraid of AI. I stopped because I realised I was competing with something that would always be faster at delivering information. And I had no interest in racing to the bottom.
Information vs Judgement
Here's the distinction I kept coming back to: AI is exceptional at delivering information. It's not good at judgement.
It can tell you how to use Prisma. It can't tell you whether you should, given your specific constraints, team composition, and long-term maintenance burden. It can generate a Nuxt component. It can't tell you whether that component belongs in your architecture at all.
Judgement comes from experience. From building systems that failed. From inheriting codebases that were "best practice" three years ago and are now unmaintainable. From understanding that technical decisions have business consequences that outlast the sprint.
That's the gap. And it's widening.
The Value Gap
I've noticed something in the last two years: the gap between junior developers and senior engineers has become more pronounced, not less.
Juniors can now produce code faster than ever. They can scaffold projects, generate boilerplate, and ship features without fully understanding what they're building. That's not a criticism — it's a structural shift. The barrier to producing output has collapsed.
But the barrier to producing good output hasn't moved. If anything, it's higher. Because now you need to evaluate AI-generated code, understand its assumptions, and know when it's subtly wrong. That requires more experience, not less.
The seniors who can do this are more valuable than ever. The juniors who can't are more replaceable than ever. And the content that helps bridge that gap — content about trade-offs, architecture, and decision-making — is exactly what's missing.
What I'm Coming Back With
In 2026, I'm returning to YouTube. Not with tutorials. Not with "build X in 10 minutes" content. That era is over for me.
Instead, I'm focusing on the things AI can't replicate:
- Architectural decisions and their consequences
- Trade-offs between competing approaches
- Real-world system design, without the client names
- Honest post-mortems on projects that didn't go as planned
- The thinking behind the code, not just the code itself
This is content for people who already know how to build things, but want to build them better. For developers who've shipped production code and now want to understand why some systems age gracefully while others become liabilities.
Who This Is For
If you're looking for someone to teach you React from scratch, this isn't it. There are better resources for that — including, frankly, AI itself.
But if you're a mid-level developer trying to think like a senior, or a senior trying to think like a consultant, or a consultant trying to articulate why your approach matters — I think we'll have something to talk about.
See you in 2026.