As AI increasingly handles execution and building, the differentiating human skill shifts to taste: the ability to run a mental "virtual machine" that predicts, given an idea, whether a specific ingroup will like it.
The source explicitly questions whether taste is a uniquely human advantage, since it's built through the same feedback/adjustment loop used to train AI models — repeated reps, not innate gift.
"Taste actually means you're able to run a virtual machine in your head where given an idea, you can predict for a certain ingroup whether they're going to like it or not. You just have to do reps. It's almost like training a model."
Related: Agency as the AI-Era Differentiator, Narrow Audience / Small Group Product Focus.
Evan Spiegel rejects the framing of a "standoff" between design, PM, and engineering disciplines as AI reshapes how products get built. AI now lets designers ship code directly, collapsing some of the traditional handoff between design and engineering — but Spiegel argues design's role as a quality "bottleneck" remains essential precisely because AI increases output volume: someone still has to gate that output for a cohesive product experience, and that's design's job, not a discipline being displaced.
Discussing the future of entertainment, Elizabeth Stone (Netflix CPTO) argues formats will keep multiplying (games, live, podcasts, clips) and personalization/immersion will deepen, and AI will materially reshape production and creative tooling — but humans will remain at the core of storytelling. She pairs this with a broader claim that craft excellence in creative disciplines (see Comparative Advantage by Role Persists Amid AI Fluidity) isn't going away even as execution tools change.
Adam Mosseri (Instagram) frames taste as the scarce resource precisely because AI collapses the cost of building: "In a world where it's easier to build things, it's more important to make sure that your time is spent figuring out what you should be building in the first place." As execution becomes cheap and commoditized, the bottleneck shifts to deciding what is worth building. He argues the people who benefit most from AI aren't necessarily the best builders but those who are clear-eyed about AI's current strengths and weaknesses and have a 'nose' for where its capabilities are heading — taste applied to judging the tool itself, not just the output. See ai-strategy-requires-heavy-constraint for the related claim that AI cannot originate strategy, only execute it well once heavily constrained by human judgment.