I think most indie hackers are building too fast and validating too late

A realization I’ve had after building multiple products:

AI has made building ridiculously easy.

You can go from idea → MVP in days now.
Sometimes even hours.

I’ve done it myself.

Built tools.
Shipped apps.
Experimented a lot.

And honestly, that speed feels addictive.

But I’m starting to notice a hidden problem.

Because building is so easy now, many of us (including me) are using building as a substitute for validation.

We build because we can.

Not because we’ve proven we should.

I’ve caught myself doing this multiple times:

Get excited about an idea
Start imagining the market
Build the MVP
Only then ask: “Wait… who actually wants this?”

That order feels wrong.

Sometimes the real skill isn’t shipping faster.

It’s resisting the urge to build before deeply understanding the problem.

AI has reduced the cost of execution.

But it has not reduced the cost of building the wrong thing.

And that’s where I think many indie hackers silently struggle.

Curious to know:

How do you validate ideas today?

Before building MVP?
After building MVP?
Or do you also learn mostly by shipping?