AI Can Write the Code, but Someone Has to Know if It Is Any Good
Netflix technology chief Elizabeth Stone says great engineering and design are still scarce, and accountability for quality has not gone away.
Lenny's Podcast: Product | Growth | Career · Why Netflix is betting on systems thinkers—not specialists—in the AI era | Elizabeth Stone (CPTO)Elizabeth Stone, the chief product and technology officer at Netflix, was asked a blunt version of the question hanging over every tech team right now. If AI lets everyone build, prototype, and analyze, do separate crafts like engineering, design, and data science still matter? Stone replied, “There’s still something that is not replaceable when I think about the craft and how they think about what good looks like,” meaning the trained judgment to tell whether work is actually good.
By craft, Stone said she means the judgment that does not come with the tool. Engineers, designers, and data scientists may now speak more languages than they used to because AI helps them stretch across roles, but humans still have to decide whether the team is solving the right problem, whether the solution is right for Netflix members or business partners, and whether the quality holds up. Stone said, “I still find great engineering to be scarce. Great data science to be scarce. Great creativity to be scarce.”
That is why she said mentorship needs just as much investment as before. Junior employees may arrive fluent in new AI tools, and Stone said, “I can guarantee you that earlier career talent is going to be teaching older folks like me many new things, too.” But she argued they still need to be taught what excellence looks like, because the person who submits code or ships a product remains accountable for its quality, its design, and how consumers experience it.
Stone drew a distinction between “being able to write lines of code in a particular language like Python or C++” and “understanding how code, computer systems, products work.” Someone still has to know whether the product is good, whether it is working as expected, and how to diagnose it when it fails. Netflix tries to move fast, take risks, and fail fast and recover fast, she said, and that calls for a grasp of how these systems are working.
Looking at some AI-written code, she admitted, unsettles her. “I have no idea why” it performs better, she said, “and if this thing breaks I’m going to have no idea how to fix it.”
Questions people ask
If AI writes most code, what should a beginner engineer still learn first?
Fundamentals still matter: how software runs, how data moves, how to test, and how to debug. Those skills let you review AI output, spot bad assumptions, and fix failures. Writing from scratch is useful mainly as training for reading and judging code.
Does knowing what good looks like apply outside engineering?
Yes. In design, it means judging whether a product is clear, coherent, and easy to use. In data science, it means judging whether data can be trusted and whether the analysis answers the right question. In each case, taste and verification matter more than speed.
How do teams stay accountable when they use AI tools?
Many teams treat AI output as a draft that needs human review and testing before it reaches customers. Netflix emphasizes that the person who submits the work owns the outcome. Clear ownership, code review, and testing help keep quality from slipping.