AI · Industry

Writing Code Is Dead. Engineering Isn't.

Replit, Lovable, and Google AI Studio turned building software into something founders can do themselves. Here's where that works, where it breaks, and why it's the biggest opportunity I've seen in a decade.

Gabriel Higareda · Nov 3, 2026 · 7 min read · AI

I've been writing code for a living for over a decade. I have a CS degree, I've been an engineer at a startup, and I've built a seven-figure freelance business on Upwork one pull request at a time. So believe me when I say this isn't a hot take I'm making for clicks:

Writing code, as a job, is dead.

Engineering isn't. Those used to be the same thing. They aren't anymore, and the gap between them is where the money is going to be for the next ten years.

I Went Through Every Stage of This

AI didn't hit me all at once. It hit me in waves.

First I was scared. I genuinely thought I was going to be replaced. Everything I'd spent my career getting good at was suddenly something a chatbot could do in thirty seconds.

Then I drank the Kool-Aid. I decided that if I wasn't using AI for 100% of everything, I was going to fall behind. Every task, every file, every decision went through a model first.

Then it bit me in the ass. Code that looked right and wasn't. Architecture decisions that made sense in the moment and fell apart a month later. Confident answers that were confidently wrong. So I got cautious.

And then I saw the opportunity. Not in AI replacing people like me, but in what AI does for founders, and what happens to those founders six months later.

What "Writing Code Is Dead" Actually Means

Five years ago, I was writing code full-time as an employee while my brother wrote code for a single client. Our dream was to someday have a team of engineers maintaining five to ten projects.

Today, my brother and our founding engineer manage all ten of our projects between the two of them. And they spend far less of their time actually writing code than either of us did five years ago.

The work didn't disappear. The typing did. What's left is the part that was always the real job: deciding what to build, choosing how to build it, checking that it was built right, and keeping it alive once real people depend on it. That's engineering. The keyboard part was never the valuable part. We just couldn't separate the two until now.

This Is the WordPress Moment All Over Again

When WordPress showed up, it took something that used to require a developer, putting a real website on the internet, and handed it to everybody. It now runs a huge chunk of the web.

The people who won that era weren't the ones who refused to touch it. They were the ones who mastered it. They learned the plugins inside and out, then started building for WordPress: themes, plugins, agencies, hosting. A whole economy grew up around it, almost like Roblox, where the platform is free and the people who get really good at building on it make the money.

Replit, Lovable, and Google AI Studio are that moment for apps. Learning them right now is like learning WordPress when it first came out. A founder with no engineering background can describe a product and have a working version of it by the end of the weekend. That's real, and it's not going away.

But remember the other half of the WordPress story. It also produced a mountain of slow, broken, plugin-stuffed sites that somebody eventually got paid to fix. That's going to happen again, at a much bigger scale. I wrote about why I moved off WordPress already. This is the same pattern, one level up.

These Aren't Demos. They're Making Money.

I'm not speculating about this from the sidelines. These are real products, built with these tools, generating real revenue:

Each of these got to market faster and cheaper than a traditional build would have allowed. Each of them also needed someone who knew what they were looking at to get them from "it works on my screen" to "it works for paying customers."

Where the Tools Get It Wrong

Here's the thing nobody selling these tools will tell you: they're incredibly good at building something, and not very good at knowing whether it's the right something. A few examples from real projects:

None of these mistakes look like mistakes on day one. The app runs. The demo is great. The problems show up when you have users, data, and a reason to change something without breaking everything else.

The tool doesn't know what it doesn't know. And if you're a non-technical founder, neither do you. That's not an insult. It's just the gap.

The Playbook If You're a Founder

Here's my honest advice, and yes, it's advice that sends you away from me at first:

Build your MVP yourself. Use Lovable, Replit, or AI Studio. Get it in front of people. Find out if anyone actually wants it. These tools are going to keep getting more expensive, but for now they're still a lot cheaper than hiring me. Don't spend agency money validating an idea you can validate on a weekend.

Then, once you have customers, bring it to someone who knows how to scale it. You'll know it's time when:

That's the path from idea to money. AI makes the first half of it accessible to almost anyone. The second half is where people like me come in, and it's going to be a gold rush for engineers who know how to walk founders down it.

Why You Still Need an Expert

The question every founder should be asking isn't "can AI build my app?" It can. The question is: how do you know it built it correctly?

You can't ask the AI. It already told you it did. You need someone who has seen enough real systems succeed and fail to recognize the difference.

I've built more than 20 projects over my career, and some of them now make tens of millions of dollars. I'm top 1% Expert-Vetted on Upwork with a 100% Job Success Score. That's not a brag. It's the point. When I look at what a model produced, I know which parts will hold and which parts will fall over at the first thousand users, because I've watched both happen.

That's what engineering is now. Not writing the code. Knowing whether the code is right.

Writing code is dead. Long live engineering.

Built Something With AI? Let's Make It Last.

Bring me what you built with Lovable, Replit, or AI Studio. I'll tell you what's solid, what's going to break, and what it takes to scale.

Book a 15-min call