Why PMs Might Be Best-Placed to Build Alone
The consensus is that AI finally exposes the PM role as unnecessary. Engineers can now build whatever they want because AI handles the tedium. Designers can ship without waiting on eng. PMs are the last holdouts of a role that was always organizational padding, and the tooling is just revealing what was always true.
I think the opposite is closer to true. The interesting question isn’t whether PMs were ever needed. It’s who the solo-builder era newly includes. For years, the answer was engineers. Then designers, once the tooling caught up. The new thing is that PMs, long assumed to lack a craft, always had one. It just didn’t produce an artifact, so nobody called it a craft. And it happens to be the craft this era rewards most.
None of this is an argument against teams. Teams with talent density still outbuild any one person, and probably always will. The claim is narrower: about who the solo-builder lane newly opens to, and what they bring that the people already in it don’t.
Crafts Inherit Their Defaults
When an engineer builds solo with AI, the default is to build a technically interesting thing. When a designer builds solo with AI, the default is to build a beautifully crafted thing. Both can produce great products, but the default is always inherited from the craft you bring. “Is this worth building?” and “what’s the smallest version that would tell me if I’m right?” are also craft questions. They were the PM’s craft questions. They just didn’t show up as code or pixels, so nobody noticed they were craft. And they get harder as AI removes every other bottleneck.
The cost of making the wrong thing drops. The cost of being wrong about what to make stays roughly the same.
This is the craft a good PM spent a decade practicing, badly or well. Reading a market. Sitting with ambiguity. Weighing what users actually need against what they say they want. Understanding the business well enough to know which tradeoffs are real and which are imagined. Deciding what to cut. It was always there. It was just the kind of craft a PRD couldn’t show off.
Jensen Huang said something close to this on a podcast last year. Asked who the smartest person he’d ever met was, he reframed the question: the smartest people sit at the intersection of being technically astute, empathetic, and able to infer the unspoken. The technical part is becoming commodity. The other two are not.
The Take Has No Venue
The reason this doesn’t get said is structural. Each visible craft has tribal infrastructure for celebrating “I built this alone”: indie hacker Twitter, Product Hunt launches, Dribbble shots, GitHub stars. PMs don’t have that infrastructure because their craft never produced a celebratable artifact. The closest thing is a PRD, which is not art. So the take that the PM craft might now be the scarce one has no venue to be said, and the memes fill the vacuum.
Not Every PM
I want to be careful here. Not every PM was actually practicing this craft. Some were translating between teams, some were running ceremonies, some were producing documents that summarized what the engineers had already decided. Those PMs were doing process work, not the craft. The memes were about them.
But the PMs who were actually doing this craft now have every other part of the stack unlocked for them in a way that’s never been true. AI is best at exactly the things PMs historically relied on other people for. What’s left is the craft nobody else was practicing.
The Craft That Was Always There
If this is right, then the hardest thing for a solo-building PM isn’t learning to code or learning to design. It’s recognizing that the craft they already practiced, the one nobody called a craft, is the one this era rewards most. The temptation will be to adopt a more legible craft identity to feel legitimate. The legibility was never what made a craft real.
I’m still figuring out where this breaks. The obvious counter is that some engineers spent a decade practicing their visible craft and this one. They will outcompete a PM who only ever practiced this craft poorly. That’s probably true. So the claim is narrower than it first sounds: it’s not that all PMs are well-placed. It’s that the specific craft a good PM practiced for years is the craft the solo-builder era rewards most and produces least.
The memes will probably still win the discourse for a while. The infrastructure isn’t built yet.