Vibe Coding Makes Building Easy, But...

Building used to be the hard part. You'd spend months learning a stack, wiring up auth, fighting deploys and that effort acted as a filter. By the time something shipped, you'd at least committed.

That filter is gone. With AI, you can ship a working app in a weekend. Which is amazing — and also a trap. When building is nearly free, it's just as easy to spend that weekend building something nobody wants, then another month polishing it because you're emotionally invested now.

The scarce skill isn't shipping anymore. It's knowing what's worth shipping.

That's why I keep coming back to ads as a validation tool. Not surveys, not "would you use this?" DMs, real strangers, spending their actual attention on a real offer. You write a landing page, point a small amount of paid traffic at it, and watch what happens:

Do they click the ad? (is the promise interesting)
Do they sign up once they land? (is the offer interesting)
That's demand, measured — and you get it in days, before you've sunk weeks into code. It also tests the thing code can't: the positioning. Same product, different message, completely different results. Cheaper to learn that from a $50 traffic test than from a launch that flops.

One honest caveat: a small ad test doesn't prove the whole business. It tests one framing of an idea with one audience. But "early evidence strangers care" beats "I have a good feeling" every time.

In a world where anyone can build anything, the edge isn't building faster. It's building the right thing. Validation used to feel like the boring step before the fun part. Now it kind of is the fun part.

How are you all deciding what's worth building before you build it?