The Jailbreak

You’ve been here before.

You needed a tool. A workflow. A dashboard. Something that would make your job easier, your team faster, your data accessible. You asked IT. You asked engineering. You submitted the ticket, made the case, followed up.

And they said no.

Or worse — they said yes, and then nothing happened. It went on the roadmap. The roadmap went into a drawer. Your need stayed unmet.

So you did what everyone does. You built a workaround in Excel. You duct-taped something together in Google Sheets. You found a shadow IT solution that sort of worked. You made do.

That was the old world. The gatekeeper controlled who got what built, and most people got nothing.

Now? You can describe what you need in plain English and AI will generate working software in an afternoon. No coding skills required. No permission. No ticket. No roadmap. No waiting.

This is genuinely mindblowing and liberating. And if you’re a non-technical person who has been diving deep into the latest vibe coding capabilities, it’s a taste of real creativity and freedom.

Finally you can build the thing yourself. Solve the problem. Gatekeepers be damned.

But as you go deeper down the disposable software path, there’s something the gatekeeper knew that you’re going to need to learn.

Why This Time Is Different

This isn’t like learning Excel or picking up a no-code tool.

AI-generated software can be real software — connected to APIs, processing real data, running real logic. The gap between ‘I wish I had a tool that…’ and ‘here’s the tool’ has collapsed to a conversation.

Someone who couldn’t write a line of code six months ago (like me) can now describe what they want to an AI in English and get something that actually works. That’s not an incremental improvement. That’s a new reality and it’s game on.

Which means the number of people who can create software just exploded. And so did the amount of software that’s about to exist.

This is where it gets interesting.

What the Gatekeeper Knew

The gatekeeper wasn’t just being a jerk. They weren’t saying no because they enjoyed crushing your dreams or protecting their turf.

Okay, some of them were. But the good ones? They knew something.

They knew that building software is relatively easy. Maintaining software is hard.

Every tool that gets built needs to be secured. Updated. Debugged. Maintained. Understood. Trained. Evolved. It’s like a living system.

The gatekeeper looked at every request and thought: if I say yes to this, who’s going to own it for the next five years?

They didn’t have the capacity to own everything. So they said no to most things.

This was frustrating for you. It was also rational.

The problem wasn’t that they were wrong about the maintenance burden. The problem was that they used that knowledge to block you instead of helping you understand the tradeoff.

Now the gatekeeper is gone. AI lets you build without permission.

Which means you need to understand the tradeoff yourself.

You’re the Owner Now

Here’s the shift: you’re not asking permission anymore. You’re the owner.

Owners don’t ask ‘can I build this?’ They ask ‘am I willing to maintain this?’

It’s easy to screw around. It’s hard to raise a child.

Every piece of software you commit to is a child you’re agreeing to raise. Not for a sprint. For its lifetime. The moment that tool becomes something you or others depend on — the moment it connects to real workflows, touches real data, informs real decisions — you’ve made a commitment.

This isn’t an argument against building. Build away. AI has given you an extraordinary capability and you should use it.

It’s an argument for being honest with yourself about what you’re signing up for.

Before You Commit

Before you deploy, ask:

Is this a throwaway or a dependency? A script you run once to clean up data? Build it, use it, delete it. No commitment required. But a tool your team will rely on every week? That’s a dependency. Dependencies have to be maintained.

What happens when it breaks? And it will break. The API changes. The data format changes. Something upstream shifts. Who fixes it? If the answer is ‘me, at 10pm, when I’m trying to finish something else’ — is that okay with you?

What happens when I leave? Or change roles. Or go on vacation. Can someone else figure this out? Is it documented? Will it become someone else’s mystery to unravel?

Would I rather own this or rent it? Vendors exist to take problems off your plate. When you buy instead of build, you’re not paying for software — you’re paying for someone else to handle the maintenance, security, updates, and support. Sometimes that’s worth it. Sometimes it’s not. But it’s always a choice.

None of these questions lead to ‘don’t build.’ They lead to ‘build with your eyes open.’

We’ve Been Here Before

If you were running technology in the 1980s, you remember what happened.

The PC revolution put computing power on every desk. Suddenly anyone could build their own tools. Finance ran Lotus 1-2-3. Marketing had dBase. Sales kept everything in ACT! or Excel. People finally had the freedom to solve their own problems.

And for a while, it was glorious. No more waiting for IT. No more begging for resources. Just build what you need and get on with it.

Then the chaos set in. Nothing talked to each other. Data lived in silos. Getting a unified view of anything meant printing reports from six different systems and stapling them together.

The companies that thrived weren’t the ones who built the most. They were the ones who figured out what to keep building themselves and what to consolidate. They learned what was worth owning.

We’re about to repeat that cycle. Except the PC put a tool on every desk. AI puts a developer on every desk. The rate of creation isn’t 10x faster — it’s 100x. Which means you’ll learn these lessons 100x faster too.

The Freedom and Responsibility Are Real

I don’t want to undersell what’s happened here. The ability to describe a problem in plain language and have AI generate a working solution — without permission, without budget, without begging — is a genuine shift in power.

Use it.

Build the dashboard you’ve needed for years. Automate the workflow that’s been eating your time. Create the tool that actually fits how you work.

Just know what you’re signing up for. The freedom to create comes with the responsibility to maintain. Congratulations. You’re the gatekeeper now.

P.S. The tell for whether something is worth committing to: Would you be proud to support this tool for the next five years? If yes, build it and own it. If not, maybe it’s a throwaway — or maybe it’s a problem better solved by someone else.