Can Vibe Coding Save the Internet?

I am a 90s kid.

I watched the internet take shape in real time. The Flash generation. Cartoon Network and Fox Kids websites filled with strange little Shockwave games. I did not know what any of it meant. I just knew that from my living room I could explore an endless world.

We did not have much, which somehow made it feel even bigger.

A few years later a friend showed me Neopets, and I was hooked. But what really grabbed me was the question underneath it all: how does this work? My dad, a webmaster back when that was the title, had taught me HTML at the kitchen table. Neopets felt different. This was the dynamic web. Things responded. Pages changed. It felt alive.

That question stayed with me.

By the early 2000s I was opening a text editor and writing code. I did not call it software engineering. I was not thinking about careers or money. I just wanted to understand the machinery behind the curtain and build small pieces of that world myself.

Maybe for you it is not code. Maybe it is cooking, woodworking, power-lifting. Whatever it is, you recognize the feeling. The thing that scratches an itch in your brain in a way nothing else does.

The internet used to be awesome.

The Internet Feels Different Now

The dead internet theory has been around for a while, predating AI as we now think of it in 2026.

The theory is simple, that sometime over the last decade or so, non-human activity has outpaced human activity to the point that it's not implausible to spend a full day on the internet and not interact directly with a single human, or even consume content generated by a human.

I don't mean in the sense of 100% organic, grass-fed, human-produced content with no machine (smart or otherwise) involvement at all; but content that is either purely procedurally or inferentially generated with minimal human involvement or oversight for the pure purpose of economy of attention.

I recently stumbled upon a post on LinkedIn, "My internet is dead" by Gregg Bolinger.

His comparison to searches being treasure hunts reminded me of stumbling across a directory listing with no index.html file and just seeing what all they had. I hadn't thought about that in years!

But once the nostalgia fades... I find it harder and harder to find genuine and authentic content. Everything feels like "Hey ChatGPT, write me a blog post about end to end sales!". We've even been conditioned to start scrutinizing common constructs in our language and grammars like the em dash or negative contrast statements. Where do you suppose the AI learned those from?

Vibe Coding: A Quick Note

There is a fierce debate ongoing about what vibe coding means and the comparison and contrast of vibe coding vs agentic coding vs AI assisted coding.

For the sake of this article, pretend none of that matters. Because it's not the point. Pretend it just means "I can build whatever I want because my time and ability are both less constrained now".

We Used to Build Things Just Because

I do not mean that everyone was a hobbyist saint building for the love of art. People were monetizing, freelancing, turning side projects into income. That has always been part of the web.

But it was not the default assumption. What felt different was how normal it was to ship something small or niche without first asking about growth strategy, marketing pipelines, or investors.

A page could exist because someone cared about a niche game mechanic. A forum could exist because five people wanted to argue about theology. A tool could exist because one person was annoyed enough to fix a problem for themselves.

That still exists. The internet largely runs on those tiny projects built by one person who cared enough to solve something.

Some of it was passion. Some of it was pride. The motivations were mixed, but the barrier between idea and execution was low enough that experimentation felt cheap.

Over time, the incentives tightened. Metrics became the default lens. More likes. More engagement. Even sharing ordinary parts of our lives started to feel like product strategy.

It is not clear that we can unwind that shift. The forces behind it are structural.

But when the cost of building drops again, the calculus changes. And that is where this moment starts to get interesting.

I Built a Stupid Thing in 20 Minutes

I was scrolling LinkedIn this morning, and every "Suggested" post was some form of AI-hype.

I thought it would be funny, maybe even a little ironic to create a tool that translates hype posts into "internal monologue". For example:

Hype Because of advancements in AI, we were ale to downsize our development team by 50%.

Reality: We couldn't afford to keep them all on the payroll, but it sounds better to say we're lean than broke.

Hype: Programmers who haven't learned how to think in systems are obsolete.

Reality: Those who can code are a dime a dozen, but only some of them have friends in HR to rewrite their job descriptions.

Hype: AI has gotten so good at writing code that we don't even need to inspect its output anymore.

Reality: We're too lazy to verify, so now 'good enough' is the new standard.

So I did. I would have never sat down and written this by hand, the novelty of the idea would have worn off.

You can play with it here.

But because I was able to just sit down and spend 10 minutes just chatting with ChatGPT about it, turn that into a phased plan, and had it live on Vercel all within 20 minutes.

It's stupid. It doesn't make money. It won't have MRR and ARR and CPCs and CEOS AND RSOADS and EPIDAS.

But it was fun, it cost me virtually nothing but a few minutes of my time, and it inspired the thoughts you are reading now.

So... Could This Actually Change Something?

"Vibe coding", switching now to a definition of AI software written entire by someone who does not have a background in coding or software engineering just writing prompts, probably not. But it might save some corners of it, the weird ones, the personal ones, the ones that are still uniquely human.

I think to build more serious projects, a combination of study, discipline, and experience will always be necessary as it always has, and I'm not sure at that point I'd use the term "vibe coding" as we seem to be centralizing on defining it.

But whatever term we use, for that group that wants to seriously and deeply learn these tools, and either already comes from an engineering background or is willing to learn the software engineering principles that still remain outside of "who wrote the code?" concerns, are significantly empowered right now. Think about the crowd that would have been just hacking nights and weekends in 2005, but with Claude Max. That has the potential to be potent.

I do believe it will help, but the forces that led us to where we are not weren't strictly technical. They were structural, and a better on-ramp for for project idea to production alone won't change human behavior (as a consumer) in a significant way, or shift the focus of algorithms that are designed to drive engagement, time-on-site and advertising revenues at all all other expense.

I genuinely don't know. But building this little Thursday morning funemployment project reminded me of an easier time, when life felt simpler. And maybe that alone warrants having just a little bit of hope.