Ten years ago, I lost a close friend in a bike accident. At her death anniversary, I said the nicest things about her, things I wish I’d said when she was alive. It planted an idea: what if we said what people meant to us before it was too late?

I sat on that idea for a decade. More on that here. Then, two years ago, I decided to actually build it. A platform where you send anonymous messages to people who matter, telling them what they mean to you. Write one, read one. No edits, no deletes, nothing performative.

I reached out to a reasonable development company for a quote. They sent back a detailed breakdown: 8 resources (UI dev, admin front-end dev, React dev, API dev, DB designer, server setup, testing, PM), estimated timeline, and a total cost of ₹2,97,500 plus taxes.

I shelved it. Nearly 3 lakhs for a side project with no returns felt a bit too much.

Last week, I built it in 2 hours using Claude Code.

Not a prototype. Not a rough version I’d need developers to “actually build.” The actual product. Database, API, front-end, authentication, deployment. Working. Live at pre-post.com.

Build cost: $20 for Claude Pro.

What Actually Happened

I’m not a developer. I can read code, understand basic logic, fumble through documentation when I need to. But I’ve never built a production web app from scratch.

I opened Claude Code, described what I wanted in plain language, and walked away. When I came back, the foundational structure was done. Database schema, API endpoints, basic UI components.

I tested it. Found things I wanted changed. Told Claude. It fixed them. I wanted a different visual approach for the preview feature. Described it. Claude rewrote the component.

Within two hours, I had a functional product. Not “good enough for now.” Actually good.

The Cost Gap Isn’t the Point

The real shift: the barrier between idea and execution has collapsed.

For decades, if you wanted to build software, you either learned to code (months to years of investment) or hired someone who could (expensive, time-consuming, requires managing people). Most ideas died in that gap. Not because they were bad ideas, but because the friction was too high.

That friction is gone.

I’m not special. I didn’t unlock some hidden technical ability. The tool just works. And it works for anyone willing to learn how to use it, which, critically, is much easier than learning to code.

This changes the equation for:

Creators paying ₹50k–₹2L for custom tools or platforms. You can build these yourself now.

Small business owners quoted ₹5–10L for CRMs, inventory systems, or booking platforms. Those quotes assumed you needed a team of humans.

Consultants and freelancers outsourcing technical work. You can bring it in-house without becoming a developer.

Early-stage founders bootstrapping on tight budgets. Your MVP doesn’t cost ₹15L anymore.

Why This Matters Beyond Builds

Developers have spent the last year watching AI go from “helpful autocomplete” to “does most of the job.”

The reason I could build pre-post.com in 2 hours isn’t because I learned some new skill. It’s because the skill requirement disappeared. Domain knowledge (understanding what you want to build and why) now matters more than technical knowledge (knowing how to build it).

That shift, from “you need technical expertise” to “you need to know what you want,” is coming for most knowledge work.

Legal research. Financial modeling. Market analysis. Content production. Customer support. All of it runs on the same substrate: reading, analyzing, deciding, writing. The same substrate AI is getting exponentially better at.

The gap between what you think AI can do and what it actually can do right now is enormous. The models released last week are unrecognizable from versions even six months ago.

The people who will come out of this well aren’t the ones with the most expertise in their current role. They’re the ones who start using these tools seriously, right now, and figure out how to do their work differently before they’re forced to.

What You Should Actually Do

Start using AI. Not casually. Not as a search engine. Actually use it for real work.

If you’re a creator, try building that tool you’ve been meaning to commission. Fair warning: there’s a learning curve. You’ll need to understand basics like databases integration, hosting, deployment. But it’s learnable in days or weeks, not months or years. In my case, it was 4 hours with couple of good friends who helped me understand how this works. And once you do, you’re not dependent on dev quotes anymore.

If you run a small business, take what you spend the most outsourcing and see if you can do it yourself. If you’re in consulting, find one deliverable you currently spend 8 hours on and see if you can do it in 2.

Don’t assume it can’t do something because it seems too complex. Try it. If it kind of works today, it’ll probably work perfectly in six months.

Here’s the real test: if you’re not using AI daily in your work right now, you’re not going to be competitive in 2–3 years. The people using these tools are moving faster, delivering more, and they’ve unlocked leverage you don’t have yet.

The Uncomfortable Truth

Two years ago, building pre-post.com required ₹3 lakhs and a team of 8. Last week, it required 2 hours. That’s not a prediction. That happened.

If you work in an industry where the core of your job happens on a screen, this affects you. Not eventually. Now.

The tools are available. They cost almost nothing. The question is whether you’ll use them.