The Mental Models That Break When Building Gets Cheap

The short version

Most popular mental models, the ones that show up in every "frameworks for better decisions" grid, are economics with the assumptions hidden. The biggest hidden assumption is that building is the expensive, scarce, slow part of any system. For forty years that was true. AI is ending it. When the cost of building collapses, four of these models invert and quietly start giving the wrong answer: Opportunity Cost, Bottleneck, Sunk Cost Fallacy, and Local vs Global Optimum. A few get sharper instead, because they were never about cost: First Principles Thinking, The 5 Whys, and Incentives. The dangerous part is that an inverted model still feels right, because the framework looks unchanged. So the move is to run a Chesterton's Fence check on your own head: find which of your mental models still price building as scarce, and retire the ones that do.

Every few weeks someone posts a grid of mental models and the internet nods along. First Principles. Opportunity Cost. Bottleneck. Sunk Cost. They feel timeless, like physics. They are not. Most of them are economics, and economics has assumptions baked in. The biggest assumption running through the canon of decision frameworks is that building is the expensive, scarce, slow part of any system. That was a safe assumption for a long time. It is no longer safe.

When agents collapse the cost of building, the math underneath these models changes. Some get sharper. A few invert completely and quietly start giving you the wrong answer. The dangerous part is that they still feel right while they do it, because the framework looks the same on the surface. Nobody runs a Chesterton's Fence check on their own mental models. So let me do it.

The ones that break

These four are not wrong as logic. They are wrong about the world they assume. Here is the audit before I walk through each one.

Opportunity Cost

The classic version says every yes is a no to something else, so choose carefully. That logic only holds when building is scarce. When you can stand up five agent-built prototypes over a weekend, the cost of saying yes to a fifth one approaches zero. Serial prioritization, the entire ritual of the roadmap, was a rationing system for a resource that is no longer rationed. The opportunity cost has not disappeared, it has moved. It now lives in attention and judgment, not in engineering hours. The teams still treating "what do we build next" as the hard question are optimizing a constraint that left the building. This is most of why I argue you should kill the roadmap and run an anti-backlog instead.

Bottleneck

The model is correct: the slowest part of a system sets the pace, and improving anything else is wasted effort. The error is in where everyone thinks the bottleneck is. For most of the last two decades the bottleneck was engineering throughput, so we hired, we sprinted, we measured velocity. That bottleneck is dissolving. The new slowest part of the system is taste, problem selection, and the judgment to know what is worth shipping at all. Plenty of orgs are pouring AI into the old constraint and cannot understand why nothing actually got faster. They relieved a bottleneck that had already moved downstream. The shift from operator to editor is the whole subject of the old PM versus the product builder.

“

The bottleneck did not disappear. It moved from engineering throughput to taste and problem selection. Most orgs are still relieving the old one and wondering why nothing got faster.

”

Sunk Cost Fallacy

This one does not break, it gets stronger, and almost nobody has updated their behavior to match. The model says past spend should have no bearing on the next decision. Fine. But when code is nearly free to regenerate, throwing work away stops being the painful option and becomes the cheap one. The emotional weight of "we already built this" was always irrational. Now it is also expensive in a second way, because clinging to last quarter's build keeps you from regenerating something better in an afternoon. The fallacy is the same. The penalty for falling for it just went up. I wrote about the version of this that mattered most to me in my best product decision was a kill, not a launch, and about the cost of hoarding old work in the prototype graveyard.

Local vs Global Optimum

The frame warns that incremental optimization can trap you in a good-enough version while the radically better solution sits unexplored, because exploring it costs too much. Read that last clause again. The whole reason incrementalism won was that exploring the global solution space was prohibitively expensive. Agents make that exploration affordable for the first time. The model survives intact. The excuse for staying local does not. "We will iterate toward it" used to be prudent. Now it is often just a habit dressed up as discipline, and instant prototyping is how you actually go explore.

The ones that hold, and get sharper

Not everything on the grid is collapsing. A few models matter more in this environment, not less, because they were never about the cost of building in the first place.

First Principles Thinking is the entire game now. When building is cheap, the differentiated act is knowing what is actually true about the problem before you generate anything. The cost of being wrong about the fundamentals used to be hidden inside long build cycles. Now you find out fast, at scale, in public.

The 5 Whys gets sharper for the same reason. Agents are extraordinary at giving you a confident answer to the question you literally asked. They are indifferent to whether it was the right question. Root-cause discipline is the human's job, and it is the one that does not delegate.

Incentives stays exactly as true as it ever was, because it was never about technology. People respond to what they are rewarded for, not what they are told to want. No model collapses that. If anything, it is worth pairing with why survivorship bias quietly wrecks PM judgment, because both are reminders that the world keeps lying to you in the same direction.

“

A framework is inherited infrastructure. It carries the assumptions of the era that built it. Run the same Chesterton's Fence check on the models in your own head that you would run on anyone else's rule.

”

The actual takeaway

The point is not that mental models are dead. The point is that a framework is a piece of inherited infrastructure, and infrastructure carries the assumptions of the era that built it. You already know to ask who put a rule there and what problem it was solving before you tear it down. Apply the same test to the frameworks in your own head. Half of them were written for a world where the expensive thing was making the thing. That is no longer the expensive thing, which is the same structural shift the whole AI agent fleet is built around.

The product builder version of decision-making is not "use better mental models." It is "audit which of your mental models still price building as scarce, and retire the ones that do." Run the Chesterton's Fence check on your own thinking. Most people never do, which is precisely why it is worth doing.

Pick one thing to try this week

Take the last big "what should we build next" debate your team had. Ask one question: would this still be a hard decision if building it cost an afternoon instead of a quarter? If the answer is no, you were not making a product call, you were rationing a resource that is no longer scarce. That is one inverted model, caught in the wild. Find the next one.

Sources: the "15x Mental Models" grid from Hustle Badger prompted this audit. First Principles, The 5 Whys, Sunk Cost, and the rest are standard decision-science frameworks; the inversion argument is mine.