The taste gap is the new moat

Everyone can make the average thing now

Here’s what changed. A competent landing page, a competent logo, a competent block of copy, all of it is now a few seconds away for anyone with a prompt. The floor rose overnight. And when everyone can reach the floor, standing on it is no longer worth anything. The average output stopped being a differentiator the moment the average became free.

What didn’t become free is knowing which of the ten competent options is actually the right one. That gap is the new moat. Ira Glass famously described the taste gap as the distance between a beginner’s taste and their early work. AI closed the ability side of that gap for everyone at once, which makes the taste side the only part left worth paying for.

Generating is cheap, judging is not

AI is spectacular at producing options. It’s indifferent about which one is good, because good depends on context it doesn’t have: this audience, this brand, this moment, this goal. Taste is the accumulated judgment that lets you look at ten plausible drafts and know, quickly, which one serves the actual situation and which nine just look fine.

• The value is no longer in making a version. It’s in choosing and refining the right version.
• Taste isn’t vibes. It’s pattern recognition earned over years of seeing what worked and what did not.
• The machine widens the funnel of options. A person with judgment is still what narrows it to the right one.

When competent is free, the scarce skill is knowing which competent thing is actually good.

If you have never heard Glass say it himself, the two-minute rendition of that interview is still the best thing ever put in front of a discouraged beginner, and it is worth every one of its millions of shares.

It’s the judgment, not the typing

This is why the prompt is the new spec, describing what good looks like is now the hard part, and why AI stays a great junior and a dangerous senior. The tools raised everyone’s output and, in doing so, made discernment the thing worth paying for. Fourteen years of seeing what works isn’t obsolete in this world. It’s the moat.

What this means for your work

Don’t compete on producing more, the machine wins that outright. Compete on judgment: the ability to look at what it makes and know what’s actually right for the situation in front of you. That’s the skill that compounds and doesn’t commoditize. If you’re trying to work out where human judgment should sit in an AI-assisted process, that’s a conversation I enjoy.