As generative AI raises the bar for what good product judgment looks like, Marty Cagan argues the PM population is bifurcating rather than shrinking uniformly. Empowered product managers who bring real Product Sense and judgment are reportedly commanding higher salaries, because AI tooling makes their judgment more valuable, not less. Meanwhile "product owner" roles and PMs on feature teams — those executing a backlog rather than exercising judgment — are predicted to shrink or disappear. Cagan frames this as the same bifurcation already visible among software engineers under AI: it amplifies the leverage of people with strong judgment and erodes demand for those whose job was mostly execution. See also AI Era Bottleneck Shift: From Building to Deciding.
A concrete split within this bifurcation: the standalone 'associate product manager' job title (a junior IC role that functions mostly as project management) is predicted to be eliminated by AI tools like ChatGPT, since that project-management work is exactly what the tools automate. By contrast, well-run APM (Associate Product Manager) Rotational Programs are expected to persist, because they function as a concentrated 'crash course in product sense and coaching' — the part of the role AI doesn't replace.
Apply: when advising on APM-adjacent roles, distinguish the job title (at risk) from the rotational development program (durable) — the deciding factor is whether the role is coaching-dense or task-dense.
Tamar Yehoshua (Glean; ex-Slack CPO) frames the same bifurcation on a 5–10 year timeline: AI will blur the lines between PM, engineer, and designer, but what it mainly eliminates is execution-heavy PM work — writing specs, coordinating handoffs, chasing status — since that work is closer to rote translation than judgment. PMs who add real creative and strategic value keep their jobs and, in her view, keep thriving, because AI 'can't be creative' and can't decide what's worth building in the first place. See also Trying New AI Products as a Standing PM Practice for her recommended way to stay ahead of the shift.
AI is expected to automate rote PM busywork — status updates, Jira hygiene, launch docs — and to surface information ('here's what customers are asking for, here are the problems today'), but not the harder judgment of deciding what to build in response. Tamar Yehoshua frames this as a filter that will 'weed out the good from the bad PMs': PMs whose value was mostly execution and information-gathering lose ground, while PMs whose value is judgment and creative/strategic decision-making keep or gain leverage. Quote: 'the not so good PMs' jobs will go away, the great PMs will still have great jobs.'