Product Management is dead. Long live Product Management.

The product manager is the most doubted job in tech right now. It's also never been more necessary in the age of AI.

Product managers are getting hit from both sides, and leaders are relearning what happens when you build without structure.

Last week a startup founder posted that his company was "currently left with no PM's and PO's." They'd hired a product manager and a product owner with the dream every startup has: ship fast without giving up quality. They tried the classic pipeline, then PM-led squads, then product-owner ownership. None of it stuck. So this year they went bolder. No PM or PO role at all. Founders set the big bets, go-to-market leaders own roadmap priority, senior engineers own delivery. He said clarity, speed, and quality all skyrocketed.

Then he got to the part I was waiting for. "Now I am the bottleneck. It doesn't scale any further than this." Followed by a request for help: he's hiring product people now, folks who are "commercial-minded, great with stakeholders and problem solvers."

He eliminated the product role, moved faster for a while, and rediscovered exactly why it existed, one painful lesson at a time. He's now advertising for the thing he got rid of. He just isn't calling it that yet. I'm not saying this to dunk on him. He was refreshingly open about a hard year, and I respect that more than the usual founder spin. I'm saying it because it's the clearest example I've seen of what's happening to my profession.

When a company says product management "didn't work," I always wonder how much agency they were given. Were they treated as teammates or as outsiders? Was their onboarding real? Were their questions heard as attempts to find the actual problem, or waved off as friction? Because once everyone can build, it gets very easy to build a polished solution to a problem no real customer actually has.

And this isn't one founder having a bad year. It's a pattern. About a third of companies that replaced people with AI have already rehired for roles they cut. Nine in ten that announced AI-driven layoffs went on to underperform the S&P 500 by around 60 percentage points. And PM job openings are back to a three-year high. The market tried pulling out the judgment layer, watched what happened, and started putting it back.

Everyone's a builder, everyone's shipping

Everyone's a builder now. I've written about why that's not necessarily a good thing, and I stand by it. But if anyone can build, don't we still need people to help orchestrate it?

And everyone's supercharged on AI now, shipping faster than ever, merging everything, moving at full speed. Ship, ship, ship. In that mode, the product manager asking "wait, should we?" looks like friction.

Building the thing was never the hard part

The hard part is deciding what to build, for whom, and getting it out the door without lighting the house on fire. When everyone can build, the bottleneck doesn't vanish. It just goes from "can we make it" to "should we, for whom, and what breaks when we do."

And there's something funny about all this bottleneck talk. Weren't we happy enough with last year's pace? Now that AI says we could do 10x more, the judgment in the loop becomes the bottleneck, and nobody stopped to ask whether we should be going 10x faster in the first place. Do customers need 10x more features?

Product's job has only gotten harder, and there are more landmines than there have ever been, because the cost of building the wrong thing quickly has never been lower or more tempting.

So no, the answer isn't to slow everyone down. The answer is to take the increased velocity and point it at the right thing, with someone minding the requirements, the rollout, and the cleanup.

We keep rediscovering product management

There was already a discipline for deciding what to build and why. It's called product management. Remove it to move faster, and you spend the next year rebuilding the exact competencies you got rid of, one rediscovery at a time, convinced you invented it. Every few years the industry kills the discipline, reinvents it from scratch, and acts surprised when it looks familiar.

The product manager title is optional, but the work is not. Somebody has to hold the "what" and the "why" or you'll find out months later when the thing you shipped fast turns out to be the wrong thing, fast.

The process-only PM is cooked

The PM who only wrote tickets and ran standups is in real trouble, and probably should be. If your entire value was moving cards across a board and scheduling the meeting about the cards, a good engineer and a decent AI can absorb that, and they will.

The PM who can build, and who has taste about what's worth building, has never been worth more. That's the version of the job that survives. It was always the real job.

Product management isn't dying. It's finally proving it was always about judgment.

Long live product management.