Claim
AI democratizes execution ability to the point where ability is no longer the bottleneck in most creative domains. Taste, the quality of discriminating judgment about what is good, becomes the scarce resource. The complication is that the tools that make execution free also reduce the friction through which taste is developed. Taste builds through deliberate, embodied attention, and tools that automate the making also automate away the deep noticing that making produces.
Mechanism
Taste develops through accumulated attention to outcomes: making something, evaluating whether it is good, adjusting, and repeating. When AI handles more of the making, the evaluation cycle still runs, but adjustments are applied to AI outputs rather than to one's own developing craft judgment. Over time, a practitioner develops skill at prompting and selecting but not the underlying judgment that makes selection meaningful.
"Ability is basically free, so for most people, ability exceeds taste."
In markets where anyone can execute, taste determines who people choose to work with and what actually gets built at the level of quality that earns attention. The scarcity is not skill at production but quality of judgment about what is worth producing.
Conditions
Holds when: execution is genuinely commoditized in a domain and the differentiating variable is quality of judgment about what to make and what not to make. Fails when: execution remains technically constrained or buyers select on output volume and cost similarity rather than quality distinction.
Evidence
Jason Liu published this as an essay on July 5, 2026. It is a practitioner reflection, not an empirical study. Liu writes from experience building AI engineering systems where the bottleneck shifts from "can we build this" to "do we know what is worth building."
Signals
- You can produce more output per day than you can evaluate well.
- The work you are most proud of is the work you edited most heavily, not the work you generated most quickly.
- Your criteria for what counts as good have not meaningfully evolved since you started using AI tools.
Counter-evidence
The "taste is the scarce resource" claim assumes buyers can distinguish taste from execution volume. In many markets, faster-cheaper-good-enough output competes directly with taste-differentiated work, and volume wins. Liu's argument holds in markets where buyers are sophisticated enough to select on quality rather than throughput. In commodity markets, the bottleneck remains price and speed, not taste.
Cross-references
- (none in current corpus)