Writing
The Evolution of Tech Leadership: How Senior ICs and Managers Are Reshaping Organizations
The old division of labor was clean: managers coordinated people, senior ICs solved technical problems. I don't think that line has been real for a while, and AI is finishing it off.
Most of the highest-leverage work I've done — at Google, at Coupang, now at Roblox — looked suspiciously like management without the reporting chain. Aligning teams that didn't report to me, framing problems so other people could execute, deciding what not to build. I've led teams formally too, and the honest comparison is that the informal version was often harder. Authority you have to earn per-decision is more expensive than authority that comes with the org chart.
AI accelerates this because leverage no longer maps to headcount. A strong IC with agents, context, and credibility can move surfaces that used to need a whole team. Meanwhile a manager without technical judgment degrades into a process router, forwarding decisions they can't evaluate. Both roles have to change, and neither gets to keep its old job description.
Where the roles converge
What managers and senior ICs converge on is orchestration. Managers have always orchestrated people. Senior ICs increasingly orchestrate everything else: systems, agents, platforms, technical direction, cross-team dependencies. In both cases the actual job is the same — take something ambiguous and turn it into something shipped, using whatever resources you can marshal.
Where they still differ
The difference that survives is the shape of accountability. A manager owns staffing, performance, and team health — whether people are growing and whether the team can sustain itself without heroics. A senior IC owns the technical boundaries: the architecture, the leverage points, and usually the first reference implementation — the prototype built deeply enough that the design argument everyone was having simply ends. Good organizations keep these roles complementary instead of forcing every high-leverage engineer into people management, a conversion that wastes a lot of strong engineers and produces a lot of mediocre managers.
In practice, senior IC leverage looks like: defining the system boundary several teams build against, turning fuzzy risk into concrete controls and launch gates, and raising the org's technical taste without becoming the bottleneck every change has to pass through. That last one is the hardest, because review is the most natural form of influence and the least scalable.
The leaders who matter over the next decade will be defined less by org-chart shape and more by whether they can turn ambiguity into shipped, governed systems. Which box your name sits in matters less than it used to. I think that's mostly good news.