Human judgements of flat design: Tech pros preaching AI taste often lacked it before AI

Tech professionals are increasingly preaching about the need to develop “taste” when using AI tools, but many of these same voices never demonstrated discernment in their pre-AI work. This hypocrisy reveals that the real issue isn’t AI creating tasteless content—it’s that people who lacked critical judgment before are now producing mediocre work at scale, making their deficiencies more visible than ever.

What taste actually means: In the AI context, taste encompasses four key skills that should have been applied to work all along.

• Contextual appropriateness: Knowing when AI-generated content fits the situation versus when human input is essential.
• Quality recognition: Distinguishing between useful AI output and “slop” through domain knowledge and aesthetic judgment.
• Iterative refinement: Understanding that AI provides a starting point requiring further development and polish.
• Ethical boundaries: Recognizing when AI crosses lines of authenticity, legality, and respect.

The tasteless reality: Many professionals worried about AI-generated mediocrity are guilty of producing low-quality work themselves, long before AI arrived.

• Common examples include copying code without understanding it, sending unproofread communications, and creating cookie-cutter designs that mirror every competitor.
• These behaviors demonstrate a fundamental lack of “critical judgment, discernment, or appreciation of aesthetic quality”—the very definition of taste these critics now claim to champion.
• As Matthew Sanabria, a tech industry observer, notes: “Don’t complain about mediocre work when you’re producing mediocre work yourself.”

Breadth beats depth: While taste can develop through deep domain expertise or broad cross-functional knowledge, breadth proves more valuable when working with AI.

• Depth of taste allows experts to recognize refined versus merely functional AI content within their specialty.
• Breadth of taste enables switching between domains effectively—like software engineers writing documentation or marketers creating designs.
• Those with broad taste “iterate faster because they have opinions about what ‘good enough’ looks like across multiple domains.”

How to develop actual taste: Rather than treating AI taste as a mystical new skill, focus on fundamental practices that were always important.

• Tomorrow: Compare one piece of work you’re proud of against one you’re not, identifying specific differences.
• This week: Study three examples of excellence in your domain, analyzing patterns and creative choices.
• This month: Iterate on something you’ve created multiple times, with each version addressing specific critiques.
• Always: When someone lectures about AI taste, “ask them to show you their work from before AI.”

The bottom line: People succeeding with AI aren’t those who suddenly discovered taste—they’re those who already possessed it and adapted their standards to new tools.