Elizabeth Stone runs product, engineering, and design at Netflix. Asked what she hires for now, she did not name a language, a tool, or a framework. She named systems thinking.
In an interview on Lenny's Newsletter, Stone described a shift most companies are living through but few have named. The narrow specialist, brilliant at one slice of the stack, is no longer the most valuable person in the room. The person who sees how the slices connect is.
That is not a soft observation. It is a hiring decision at one of the most operationally serious companies on the planet, and it tells you where the advantage is moving.
Why the Specialist Premium Is Falling
For twenty years, the market paid a premium for depth. You got rewarded for being the best at one thing, the SEO expert, the paid media specialist, the front-end engineer who lived in one framework.
AI is quietly repricing that. A model can now produce a competent draft of the specialist's output in seconds, whether that is ad copy, a SQL query, or a first pass at a component. The floor of any single skill has risen, and the scarce thing is no longer the output.
The scarce thing is judgment about how the outputs fit together. Which of these three AI-generated options actually serves the customer. How a change in pricing ripples into support load. Why a faster funnel might quietly lower retention. That is systems thinking, and no model does it for you, because it requires holding your whole business in your head at once.
I made a version of this argument in agency is the skill that matters now. When execution gets cheap, the value moves to the people who decide what is worth executing.
Excellence as an Operating System
The sharper idea in the interview is how Stone frames quality. She does not treat excellence as an individual trait, the star performer who saves the day. She treats it as an operating system, something built into how teams make decisions, not something a hero supplies on demand.
That distinction matters more in the AI era than before. When a team can generate a huge volume of AI output, the bottleneck is no longer production. It is keeping the signal high while the volume explodes. Netflix, by Stone's account, is actively wrestling with how to handle massive AI output without letting quality slide.
An operating system for excellence is how you solve that. If quality depends on one talented person reviewing everything, it collapses the moment volume triples. If it is built into the process, into shared standards and clear taste, it holds.
Stone also expects AI fluency as a baseline, not a specialist role. Everyone is expected to work with these tools, the same way everyone is expected to write an email. That reframes AI from a department to a literacy, which is exactly the right altitude.
Notice what that does to org design. If AI is a literacy, you do not build a walled-off AI team and wait for it to bless everyone else's work. You raise the floor across the whole company and let judgment concentrate where it belongs, in the people deciding what to do with the output.
It is also why Stone's remit spans engineering, product, and design at once. Systems thinking is hard to demand from teams that are structurally kept apart. Putting the disciplines under one roof is not a tidiness move, it is the org chart admitting that the interesting problems live in the seams between functions.
What This Means for Everyone Not Named Netflix
You do not need Netflix's budget to act on this. The move is available to a five-person agency and a solo founder just as much as a streaming giant.
Stop hiring only for the deepest spike in one skill. Start weighting for people who ask how their piece connects to the rest, who can hold two or three parts of the business in view at once. Those people compound in an AI-heavy team. Pure specialists plateau, because the tool now covers their floor.
Build your standard of quality into a system, not a person. Write down what good looks like. Make taste explicit and shared, so it survives more volume and more turnover. The thing a model cannot copy is a well-defined point of view about what is worth shipping, a point I pushed in taste is the last thing AI cannot copy.
And treat AI as a baseline literacy across the whole team, not a toy for the innovation corner. The gap is no longer between companies that have AI and companies that do not. Everyone has the same models. The gap is between teams that think in systems and teams that still think in silos.
Netflix just told you which side pays. The tool is not the moat. The way your people think is.