AI · 6 min

AI will replace us. Now what?

If a tool writes code in seconds, why hire a development studio? What AI really changed, what stayed the same and what to ask whoever you hire.

It’s the question we get asked the most, and not only by clients: if an AI tool writes code in seconds, why hire a development studio? We’d rather answer it head-on, from inside the craft.

What did change

Real things happened, and denying them would be dishonest:

  • Repetitive work became almost free. Basic structures, forms, connections to well-known services, routine tests. What used to take an afternoon now takes minutes.
  • Prototyping is cheap. You can see an idea working before committing to a large budget.
  • The bar went up. A generic site or a simple tool no longer justifies, on its own, paying for custom development.

What didn’t change

Writing code was never the hard part of a project. The hard parts are still here:

  1. Understanding the real problem. What a client asks for is rarely what they need. Asking well, discarding and prioritizing is human work.
  2. Deciding with judgment. What to build first, what to leave out, which technical debt is acceptable. Every decision has costs that a tool doesn’t carry.
  3. Being accountable for the result. When something breaks on a Friday night, someone has to own it. A tool doesn’t sign a contract or pick up the phone.
  4. Maintaining it over time. Software lives for years: business rules, third-party services and security threats all change.

So, will it replace us?

Our answer is that it will replace certain tasks, not the craft, and that it changes what is worth charging for. Writing lines of code gets cheap; what becomes scarce is judgment, careful review and responsibility.

That has an uncomfortable consequence: unreviewed generated code is hidden debt. It works in the demo and fails with real data, usage spikes or an attack. Someone experienced has to read it, test it and understand it before it reaches production.

Now what?

For us, three things:

  • Use the tools, with a critical eye. Rejecting them on principle is as unprofessional as accepting everything they produce.
  • Put the effort where AI doesn’t reach: talking to the client, designing the solution and verifying quality with explicit tests and reviews.
  • Be transparent about how we work and what we deliver.

And if you’re evaluating whom to hire, these questions are worth more than any promise:

  1. Who reviews what gets delivered, and how?
  2. What happens if something fails after launch?
  3. Is the code mine, and can I maintain it without depending on anyone?
  4. How much of the work is understanding my business, and how much is typing?

If you’d like to talk about your project, write to us. You talk directly to whoever builds it.