A new way of building software is catching on. People are skipping the traditional engineering process and using tools like Replit, LLM apps, and code agents to go from idea to product in a matter of hours. No sprints, no full specs, just instinct, speed, and iteration.

This is what people are calling vibe coding.

It's not polished. It's not necessarily scalable. But it's working. Products are getting built. People are shipping. And in many cases, they're getting real traction.

Gatekeepers don't like it. And they've been quick to say why. The code is bad. The architecture is wrong. The tools are unreliable. These criticisms aren't totally wrong. But they also miss the point.

They See the Flaws, Not the Possibilities

The thing that makes vibe coding interesting isn't that it's perfect. It's that it lets people build things who couldn't before. A designer, a founder, a generalist with an idea can now sit down and actually make something. Not just a mockup, but a working version.

And it's not the engineers who are leading this wave. It's people outside the walls of traditional software. People who, for years, weren't able to get their ideas built. They're the ones who see the opportunity. While the insiders see mostly the risks.

That contrast is telling.

The tools aren't perfect - but they're improving quickly. AI models are getting better at writing code, not worse. So the objections from today's gatekeepers may not even apply in six months. The flaws won't last. But the shift in who gets to build things will.

What Matters Now Is What to Build

As coding becomes easier, faster, and more automated, the real skill is moving elsewhere. It's not about knowing how to write code. It's about knowing what to build.

And that's always been the highest leverage move in software. Not writing more code, but writing the right code. Solving the right problem. Building the right thing.

Which brings us to taste.

Taste Is Not What You Think It Is

Taste is often misunderstood. People think it's just a personal aesthetic. Something subjective.

But in reality, taste - especially in product work - is about judgment. About making the right calls, over and over. About understanding what people want, and what the moment calls for.

This is the kind of skill we associate with people like Steve Jobs. Rick Rubin. Hans Zimmer. None of them were the most technical in their fields. Jobs wasn't an engineer. Rubin doesn't play most instruments. Zimmer doesn't write classical notation. But they all knew what would work. They had taste.

And while taste has been seen as something you either have or you don't, that's changing too. Because now, taste can be tested. You can build fast, ship fast, and get real feedback. You can learn what lands and what doesn't. You can improve.

That turns taste into something closer to a real, measurable skill - something closer to objective taste.

And once you can measure it, you can build it into teams. You can make it part of the culture. You can make better decisions, earlier, with less waste.

More and Better Software

The outcome of all this isn't just more people building things. It's better alignment between what's built and what people actually need.

If we can make more software, that's a win. But if more of that software starts out closer to the right idea - if it's more useful from day one - that's a bigger win. That's software with more utility. Software that does more for the people using it.

That's good for users. Good for business. And good for society.

So yes, vibe coding looks strange to people who've spent years mastering the rules of software. But the rules are changing. And the sooner we focus on what to build - and get better at deciding that - the better off we'll be.