We're living through the democratization of mediocrity.
Give a person access to image generators, copywriters, code assistants, and music tools, and they can produce something that looks and sounds competent within hours. The floor has been raised dramatically. The ceiling, though? That's more complicated.
The Abundance Problem
When creation was expensive — in time, skill, or resources — selection happened early. Most ideas died in the head because executing them was too hard. The ones that made it out had been filtered through constraints, which forced a kind of accidental curation.
Now the cost of trying is near zero. Generate ten variations. A hundred. Iterate until something sticks. The problem has shifted from how do I make this? to which of these should exist?
And that's a fundamentally different skill.
Taste Is Not Preference
There's a lazy conflation of taste with preference. "I like this" is not taste. Taste is the ability to articulate why something works or doesn't, to trace the connection between intention and execution, to recognize the gap between what was attempted and what was achieved.
Taste is diagnostic. It requires:
Pattern recognition across domains — understanding how narrative structure in writing maps to user flow in product design, how the pacing of a conversation mirrors the rhythm of good code.
Historical context — knowing what's been tried, what failed and why, what conventions exist and which ones are worth breaking.
Clarity of intent — the discipline to know what you're trying to do before you evaluate whether you've done it.
None of these are generated by AI. They're cultivated through attention, failure, and the slow accumulation of experience.
The Curator's Advantage
When we started building Bifrost, we assumed the hard part would be generating lesson content. It turns out that generating content is trivial. The hard part is knowing which content serves the learner's mental model at this specific point in their journey.
A language learning app that generates infinite exercises is worse than useless — it's noise. The value is in the curation, the sequencing, the subtle calibration of difficulty that keeps the user in the zone of productive struggle.
The same pattern repeats across our products. Lunora isn't valuable because it can generate empathetic responses. It's valuable because it knows when to push and when to hold space. That judgment isn't algorithmic. It's the result of countless micro-decisions informed by an understanding of human psychology that no training dataset can fully capture.
Speed vs. Slowness
There's a particular kind of startup wisdom that fetishizes velocity. Ship fast, break things, iterate based on data. And there's truth in it — premature optimization is a trap, and perfect is often the enemy of good.
But there's a difference between shipping fast because you've thought deeply and shipping fast because you haven't thought at all. The former is strategy. The latter is avoidance.
AI accelerates execution, which means the bottleneck shifts upstream to decision-making. The teams that will matter in the next decade aren't the ones with the fastest generators. They're the ones with the clearest sense of what should be generated in the first place.
Cultivating Taste
The uncomfortable truth is that taste can't be hacked. It accumulates slowly through:
- Consuming widely and critically (not just within your domain)
- Making things and being honest about why they fail
- Seeking out work that's slightly beyond your current ability to appreciate
- Developing the vocabulary to articulate what you observe
This is slow, unglamorous work. It doesn't trend on social media. But it's becoming the only real competitive advantage in a world where execution is commoditized.
What This Means for Builders
If you're building with AI — and you should be — the question isn't whether you can generate something good enough. The question is whether you can recognize better when you see it, and whether you have the conviction to pursue it even when the path isn't obvious.
The future belongs to curators, not generators. To people who can hold a vision of what something could be and navigate the gap between that vision and the messy reality of implementation.
AI is a multiplier. It amplifies whatever you bring to it. Which means the work of developing judgment, discernment, and taste isn't just still relevant — it's more urgent than ever.
The tools have changed. The work hasn't.