The PM Role Isn't Dying — It's Relocating

The Future of Product Managers

TLDR: The classic PM role — sitting upstream, translating customer insights into specs for engineers — is eroding fast. Shipping has gone from expensive to nearly free, which means the filtering function PMs traditionally served is no longer needed in the same way. The question isn't whether the role survives, but what shape it takes next.

Summary:

The piece opens with a striking anecdote: a genuinely excellent product leader with 15 years of experience shipping developer tools can't even get a conversation for a group PM role. The authors, former PMs themselves, treat this not as a market anomaly but as a structural signal. And I think they're right to do so.

Their core argument is clean and hard to argue with. For years, software development worked like a funnel. PMs owned the translation layer between customer insight and engineering effort. That translation layer existed because shipping was expensive — you needed someone deciding what was worth the cost. Now that shipping is close to free, the justification for that layer gets thinner every quarter.

The authors describe two live experiments. At Kilo, a 40-person company runs with one PM and a WAUzer (Weekly Active User) model where engineers each own a product area and stand up weekly with numbers and targets. It worked, producing around 10% week-over-week growth. At Solo, it's just two people — one developer — moving at what would have required a team of ten a few years ago. No PM at all. Both experiments point toward the same conclusion: the spec-to-code handoff is collapsing.

The two paths they lay out are worth sitting with. Path one has PMs shifting left toward go-to-market work — commercial discovery, customer research, saying no to features, finding paying customers before building. Path two is the "long thin layer" where engineers own everything in their product area: customer conversations, support, metrics, roadmap decisions, all of it. The authors are honest that path two requires a rare profile — technically sharp, commercially minded, and customer-facing all at once.

What I find most honest about this piece is the line that PMs were "the original vibe coders" — writing the spec while engineers were their LLMs. That's a sharp reframe. It doesn't flatten the PM's contribution, but it does name the structural analogy clearly. The role abstracted away complexity on one end. AI is now abstracting away complexity on the other end. Something has to give.

Key takeaways:

• The traditional PM funnel (insights to spec to code) is collapsing because shipping is now cheap
• Two viable successor models are emerging: PMs shift left toward commercial strategy, or engineers absorb the product hat entirely
• The "engineer owns their product area" model works at small to mid-size companies but requires unusually broad people
• Senior PMs with deep experience are struggling to find roles that look like the ones they left — the job is changing faster than the job market
• Startups have already moved; large companies will follow over the next five years per the authors' read

Why do I care:

I've watched the frontend ecosystem go through its own version of this. We used to have dedicated CSS people, then dedicated JavaScript people, then the full-stack engineer ate both. Now AI tooling is compressing the gap further. The PM story rhymes with what I've seen in engineering — specialization erodes when the cost of the underlying activity drops. What the authors get right, and what I'd push harder on, is that the shift-left path (PMs doing commercial discovery and strategy) is more durable than the "engineer who does everything" path. The latter is real and works brilliantly in a startup of 10 people, but it doesn't scale to a 200-person org with complex customer relationships. The PMs who survive this transition will be the ones who stop competing with AI on spec-writing and start competing on judgment — on the messy, relationship-heavy work of knowing what's actually worth building before anyone writes a line of code.

The future of Product Managers