Elizabeth Stone (Netflix CPTO) argues that even as AI lets anyone touch any part of the stack ("everyone can be everything now" — PMs shipping code, designers writing PRDs, engineers "producting"), the comparative strengths that historically separated roles still hold and should still anchor accountability: PMs own framing "the what," engineers own "the how" (scalability, quality, build/deploy risk), and data scientists own the trustworthiness and interpretation of data.
Stone extends the comparative-advantage argument with a concrete role split: PMs frame "the what" (problem framing), engineers own "the how" (scalability, quality, build/deploy risk), and data scientists own the trustworthiness and interpretation of data. She explicitly rejects the idea that AI-accelerated coding squeezes out design expertise: "I think it would be a mistake to say design and deep design expertise and thinking gets squeezed out just because we can write code faster." Her worry isn't that design disappears but that, without a unifying design system, faster building by non-designers risks "shipping Frankensteins" — inconsistent design languages and interaction patterns across the product. See Design Systems and Templates for the mitigation.
Elizabeth Stone sharpens this distinction: 'writing code in a specific language' is a skill AI is actively eroding the value of, but 'understanding how systems and products work' is not — conflating the two overstates how much of an engineer's comparative advantage AI actually threatens. Losing systems understanding would remove the ability to judge whether AI-generated output is actually good, or diagnose why it's broken — precisely the judgment systems-thinkers-over-specialists and mentorship-based-craft-training are meant to preserve.