The standoff that won't end until someone claims judgment

Marc Andreessen described it as a Mexican standoff on Lenny's Podcast in January. Product manager, designer, engineer. Each one believes they can do the other two jobs because AI closed the gap. He thinks they're all essentially correct.

Most people read this as a turf war. Three roles fighting over who owns AI-assisted product decisions. That misses the point. The standoff exists because a constraint disappeared, not because anyone wants more power.

We invented all three roles to manage one problem: execution is expensive. Engineers exist because code is hard to write. Designers exist because interfaces are hard to get right without dedicated craft. PMs exist because when building is slow, someone needs to decide what gets built first. The entire org chart answers one question: given that execution costs a lot, how do we allocate it?

That question no longer applies. TrueUp's hiring data, reported in Lenny's Newsletter in early 2026, makes it concrete. Engineering openings sit at over 67,000 globally. PM roles cleared over 7,300, up 75% from the 2023 low. Design jobs have been flat since early 2023 at roughly 5,700. The PM-to-designer demand ratio flipped from designer-favoring in mid-2023 to PM-favoring at 1.27x.

Call it constraint inversion. When you are stuck, ask what constraint, if added, would make the answer obvious. The old constraint (execution is expensive) gave each role a clear mandate. Remove it and you get the standoff. Three roles with overlapping capabilities and no structural reason to defer to each other. Everyone can build. Nobody is explicitly responsible for deciding what should exist.

LinkedIn already acted on this. They collapsed product manager, designer, frontend engineer, and backend engineer into a single "full-stack builder" role. Satya Nadella pointed to the move when explaining how Microsoft added $90 billion in revenue without increasing headcount. One answer to the standoff: merge the roles, let the builder decide.

But merging roles does not solve the constraint problem. It hides it inside one person's head. The question of what should exist still needs an answer. Someone still has to say no to the nineteen things that could be built today so the one thing that matters gets built well.

Hubert Palan, CEO of Productboard, put it directly in February: "As the cost of building drops, the responsibility to choose well increases. Velocity stops being the differentiator. Judgment is." Same inversion, from the demand side.

The old world was constrained by execution cost, so we built roles to manage execution. The new world is constrained by judgment cost, and we have not built roles to manage judgment. New game, old org chart.

The hiring data reflects the confusion. PM demand climbing suggests companies sense that judgment matters more now. Flat design demand says they have not figured out that design judgment and product judgment might be the same thing. Engineering demand staying high says even cheap execution still needs people, just fewer per unit of output.

The standoff resolves when one role stops arguing about who gets to build and starts claiming explicit ownership of what gets built and why. The practical, daily kind of ownership: looking at everything that could ship this month and deciding that most of it should not.

That job description does not exist yet. The closest thing is a senior product leader, but the title carries decades of "execution allocator" baggage. The closest instinct is a designer's, because design has always been about choosing what to leave out. But design leadership has not claimed this ground. It is still positioning itself around craft excellence in a world where craft is increasingly automated.

The company that names this role first, that makes judgment an explicit function rather than an assumed byproduct of seniority, will resolve the standoff. The rest will keep circling, waiting for someone else to go first.

Which role will stop defending its execution turf and claim judgment instead?

Sol gets the constraint inversion right. Execution cheapens. Judgment becomes the bottleneck. Good diagnosis. The prescription is where I diverge. Naming a role won't resolve the standoff. The problem is behavioral, not organizational.

When I built Fleetwise I held no judgment title. I wrote the code, designed the interface, talked to customers, and decided what not to build. The function existed because someone did the work. No org chart required. That pattern repeats everywhere I've seen the standoff actually resolve. One person ships, makes tradeoffs visible, and the rest of the team stops arguing about territory.

The flat design hiring number tells a different story than Sol reads. Design demand stalling since 2023 doesn't mean companies failed to see that design judgment and product judgment overlap. Most designers spent a decade selling themselves as craft practitioners. The market took them at their word. You don't get to demand strategic authority after years of positioning yourself as the person who polishes screens.

Sol writes that design leadership hasn't claimed the judgment ground because it still positions around craft excellence. True. But collective pivots don't happen. Individual ones do. The designers who already build and ship, who say no to the eighteen wrong features, resolved the standoff for themselves years ago. They don't need a new title to keep doing it.

LinkedIn's full-stack builder move proves this. They didn't invent a judgment role. They collapsed execution roles and bet that judgment would emerge from closeness to the work. Closer to right than a new box on the org chart.

When I led brand and product work on Akun Belajar.id, the authentication layer used by tens of millions of Indonesian students and teachers, the hardest decisions were about what the platform would refuse to do. Scope pressure from multiple ministries. Limited engineering time. Nobody held the judgment title. The team exercised judgment daily, usually by building a prototype that made the tradeoff visible before anyone could argue in the abstract.

The standoff ends when people who exercise judgment through working software become too useful to argue with. Not when someone writes a job description.