Disclosure: This article was drafted with the assistance of AI-powered tools. I guided, edited, and refined the concepts to match my personal philosophy. The truth is simple: AI is a powerful accelerator, and I am not pretending otherwise.
1. Vanilla JS in 2024 and Beyond? Absolutely.
I am a Vanilla JS enthusiast. I build frontend applications using native Web APIs and backend services using Node.js built-ins. I reach for standard platform features first.
Why? Because since 2007, I’ve been reading code—open source, closed source, enterprise systems, and weekend hobby projects. I’ve seen enough dependency deprecations, breaking changes, and supply-chain security vulnerabilities to know when to avoid unnecessary abstractions. AI simply makes me faster at writing clean, standard-compliant vanilla code.
Do I outright ban React, Vue, Angular, or Express? No. I just rarely need them for what I am building today. With modern features like native ESM, Import Maps, fetch, and Node's built-in test runners and watch modes, the native platform has never been stronger.
2. Hobbyist > Professional (For Me, Right Now)
I haven’t been paid to write commercial code since 2019, which completely shifted my perspective.
- Professional Development: Focuses on scale, large team maintainability, rigid SLAs, and established industry frameworks.
- Hobbyist Development: Focuses on speed, personal understanding, and staying close to the metal.
I enjoy being in the trenches with Node's native http.createServer() rather than hiding behind a massive stack of nested dependencies. It might take slightly longer to build complex routing from scratch, but it is incredibly easy to understand, debug, and maintain over years. For my current workflow, that simplicity is invaluable.
3. AI is the PowerPoint of Our Era
We have seen this kind of technological shift before:
- PowerPoint was met with skepticism by those who preferred physical blackboards, hand-drawn slides, and overhead projectors.
- Excel transformed financial modeling, moving it away from manual ledgers and calculators.
AI is the exact same evolutionary step.
Does relying on AI make us "less intelligent"? Perhaps in terms of memorizing boilerplate syntax. But AI is exceptionally good at producing documentation, boilerplate code, unit tests, and initial drafts. What used to take me hours of searching through old documentation now takes seconds.
It is a tool of efficiency. We don't hesitate to use linters, IDE autocomplete, or search engines. AI is simply the next logical iteration of those tools.
4. Navigating the Ethics of AI Code
As we navigate the ethics of AI assistance, we need a framework to keep our work grounded. Advanced tooling is highly capable, but without a human touch, it risks becoming isolated, robotic, and cold.
Before shipping any AI-assisted code or content, I run it through a simple checklist—THINK:
- T = Is it True? (Does the code actually work, or is it an AI hallucination?)
- H = Is it Helpful? (Does this solve a real problem?)
- I = Is it Inspiring? (Does it encourage others to build?)
- N = Is it Necessary? (Are we bloating the codebase or the web?)
- K = Is it Kind? (Is it respectful of the user's experience, performance, and accessibility?)
5. Final Thought: Tools Don't Change the Destination
Whether you navigate a new city using a paper map or a modern GPS system, you still arrive at the exact same destination. The tool doesn't change where you end up; it simply changes how smoothly and efficiently you get there.
Writing code with AI or drafting ideas with standard web APIs doesn't diminish the final product. The application still has to run efficiently, and the architecture still has to make sense.
I am AI-first. Vanilla JS only. A hobbyist at heart.
If this approach works for you, wonderful. If not, the beauty of the web is that there is room for all of us to build.
Brixton Mavu | Harare, ZW
Building with http, fetch, and the power of the open web.
Top comments (0)