Henryk Sarat: managers stay hands‑on

Henryk Sarat’s post landed because it took Brian Chesky’s broad warning about “pure people managers” and translated it into a specific operating choice: stay technically involved even after moving into management. On May 19, Sarat wrote on X that when he was an engineering manager, he “intentionally stayed hands-on technically” and added that “pure people-management engineering folks will have a hard time in an AI-driven world,” according to the post cited in the social briefing. The comment was framed as a reply to a discussion around Airbnb’s management model and Chesky’s recent remarks on how AI is changing the value of managers. (finance.yahoo.com) The broader backdrop came from Chesky’s own comments in early May. Airbnb’s first-quarter 2026 results, published May 7, said the company was continuing to integrate AI across the business, while recent podcast and media coverage attributed to Chesky said AI now produces about 60% of Airbnb’s code and that managers who only run recurring meetings or supervise work without direct involvement will be under pressure. That gave Sarat’s post a clear target: engineering managers who have drifted into coordination, status updates and people administration without preserving technical judgment. (news.airbnb.com) What Sarat appears to be arguing is not that every manager should behave like a full-time senior individual contributor. His point, based on the wording in the post and the thread context, is narrower: managers still need enough hands-on exposure to keep their technical intuition current. In an AI-heavy environment, where code generation, design review, debugging and architecture tradeoffs are changing quickly, that intuition becomes harder to fake and more important to maintain. That is an inference from his post and the surrounding discussion, rather than a separate quoted claim. (finance.yahoo.com) That idea also fits the leadership model now being discussed around Airbnb and elsewhere: the “hybrid” manager, sometimes described as a player-coach or tech-lead-manager. Coverage of Chesky’s remarks said future leaders would need to blend people leadership with direct engagement in the work. Sarat’s contribution stood out because he described having already made that choice in practice, rather than just endorsing the concept in theory. (msn.com) Why the post is getting traction is straightforward. For engineering managers aiming at director roles, the old assumption was often that seniority meant less time in code, systems and design details. Sarat’s comment pushes the opposite view: AI may reduce the value of managers whose main job is mediation, while increasing the value of leaders who can still assess technical quality, challenge assumptions and understand how work is actually done. Several write-ups of Chesky’s remarks made the same distinction between pure people management and hands-on leadership. (sea.peoplemattersglobal.com) Sarat himself is identifiable online as an engineering leader associated with Paxos, where public pages list him among engineering management speakers and team members. That does not by itself validate the management thesis, but it does help explain why the post was read as practitioner advice rather than abstract commentary. (happenstance.ai) The practical takeaway from the thread is narrow and concrete: in AI-shaped engineering organizations, the managers under the most scrutiny are likely to be the ones furthest from the work. Sarat’s post resonated because it offered a simple counter-model — keep managing people, but do not give up the technical edge that lets you judge, guide and unblock the work yourself. (finance.yahoo.com)