Warning: Unlimited Freedom Produces Boring Products
Why Unlimited Freedom Produces Boring Houses, Mediocre Engineers, and Programs That Never Ship
THE BLANK PAGE LIE
Why Unlimited Freedom Produces Boring Houses, Mediocre Engineers, and Programs That Never Ship
There is a mold sitting on a bench in my memory. It is the shape of a part that does not exist yet.
Before I ever wrote a patent, before 32 years of aviation research and development across two companies, before any of it, I learned how to lay composite. And the first thing you learn at the layup bench is humbling in a way that took me years to fully understand. The part is not the genius. The mold is the genius. You can be the most talented fabricator alive, and without a tool to lay against, you have nothing — wet cloth and resin and a puddle on the floor. The carbon fiber becomes a part only because the mold tells it where it is allowed to go. The constraint is not the obstacle to the part. The constraint is the part.
I didn’t have language for that at the time. I called it fabrication. Years later I would learn that the engineers who study creativity systematically had a name for what I was feeling at that bench, and a body of theory behind it. But the bench taught it first, the way a bench always does. The shape lives in the boundary. Remove the boundary and you don’t get freedom. You get a puddle.
Hold onto the mold. We’re going to come back to it.
The Confession
Here is the part I’m not proud of, and the reason I’m writing this instead of something more comfortable.
For a good stretch of my career, I believed the opposite of what that bench was trying to tell me. I believed — the way most of the industry still believes — that creativity is what happens when you remove the restrictions. Give a sharp engineer room. Take the handcuffs off. Clear the schedule, open the budget, hand them a clean sheet of paper, and stand back while genius does its thing.
I believed it because it flatters everyone in the room. It flatters the engineer, who gets to be the unbound genius. It flatters the manager, who gets to be the enlightened patron of genius. It is a very comfortable story and it is almost entirely wrong, and I can tell you the precise moment I started to suspect it.
I watched two teams, in the same era, work two problems. One team had everything. Generous budget. Open-ended timeline. A mandate that amounted to go make something great. The other team was strangled — hard weight limit, fixed cost, a calendar that did not negotiate, and a list of components they were flatly forbidden to touch. By every assumption I held at the time, the first team should have produced the masterpiece and the second team should have produced a compromise.
It went the other way. It always goes the other way, and once you’ve seen it go the other way enough times, you stop calling it a surprise and start calling it a law.
The well-fed team spent months looking for the problem. They generated options nobody would build. They re-opened decisions that were already closed. They mistook motion for progress because the blank sheet gave them nowhere to stop. The strangled team had no such luxury. They knew exactly what they were solving, because the boundaries told them, and inside those boundaries they did the most inventive work I saw in that period. The constraint didn’t crush them. It aimed them.
That is the forensic correction I owe you, because I held the wrong theory out loud for years: the blank page is not a gift. It is a void. And a void does not produce a masterpiece. It produces a search with no exit.
The Teardown: What Actually Happens on the Empty Plot
Let me give you the cleanest version of this, the one I’ve started using whenever someone asks me why their designs cost too much and take too long.
Put a blank sheet of paper in front of an architect. No site. No lot lines. No slope, no trees, no water, nothing fixed. Tell them to design a house. You will get a boring house.
Not because the architect is untalented — because you have handed them an infinite problem with no definition of done, and an infinite problem has no shape to push against.
Now give that same architect a real plot of land. There is a lake on it. A stream cuts through. A ridge rises on the north edge. There is a stand of old trees that, by law and by conscience, cannot be touched. Suddenly the architect is alive. Now the house has to answer something. The water wants a view and a foundation problem. The slope wants to be worked with instead of flattened. The protected trees say not here, which is the most useful instruction a designer can receive, because every not here sharpens the here.
Hand a bounded site to a good architect and they will hand you back a masterpiece — because a masterpiece is, at bottom, an elegant answer to a hard question. Hand them a blank page and they will hand you back a search, an invoice, and a building that could be anywhere, which is the same as being nowhere.
This is the mechanism, and the mechanism is not poetry. It is the geometry of the solution space:
(a) An unbounded problem has an infinite solution space. There is no edge, which means there is no way to know when you’ve arrived, which means you never arrive — you iterate until the money runs out and call wherever you stopped the answer.
(b) A bounded problem collapses that infinite space to a finite set of candidates that actually satisfy the boundary conditions. The constraints do the brutal work of elimination for you, before the creative work even begins.
(c) Inside that finite set, the real invention happens — fast, because the engineer is no longer searching the horizon. They are solving the actual problem in front of them, and a solvable problem is the only kind that ever ships.
The blank page feels like more options. It is actually fewer answers. The bounded site feels like fewer options. It is actually the only path to an answer at all.
The Masters Were Never Geniuses. They Were Constraint Workers.
We tell the story wrong, and telling it wrong is costing the industry a fortune.
We hold up Steve Jobs, Frank Lloyd Wright, and Elon Musk as visionaries — men who saw what others could not, who pulled the future out of the air by force of imagination. It is a thrilling story and it teaches you nothing useful, because it tells you the ingredient was genius, and you cannot buy genius, hire genius, or schedule genius. So the story leaves you helpless, which is the surest sign a story is wrong.
Look at what they actually did. Not the myth. The work.
Frank Lloyd Wright did not conjure Fallingwater from nothing. The boulder was there. The waterfall was there. The hillside was there, and the client insisted the house sit on the falls rather than politely across from them. Wright did not fight a single one of those constraints. He worked through them — one cantilever, one structural problem, one stubborn variable at a time — until the only building those constraints would permit turned out to be one of the most beautiful structures ever built. The waterfall is not the scenery behind Fallingwater. It is a load-bearing constraint, and it dictated every decision that made the house immortal.
Steve Jobs did not invent the portable music player; portable music players existed when he came back to the company. What he did was bound the problem until it had no choice but to be elegant. A thousand songs in your pocket. That is not a vision statement — that is a hard, testable constraint with a pass/fail line. Every prototype his people brought him, he said no. Too big. No. Too few songs. No. Too many buttons. No. People mistake those refusals for perfectionism. They were constraint injections, each one collapsing the solution space a little further, and by the time he finally said yes the design was already fully solved. It only remained to be built out of electronics. The elegance was not inspired. It was enforced.
Elon Musk did not invent the rocket or the electric car. He nailed a number to the wall — cost per pound to orbit, battery economics, range — a number so unforgiving that every materials choice, every manufacturing decision, every line of the supply chain had to bend around it or get cut. He did not ask his teams to be creative. He asked them to solve a savagely bounded problem and work every issue until the numbers complied. The creativity fell out of the constraint, the way it always does.
Three men. Three industries. Three eras. One method. Not one of them came up with a good idea and rode it to glory. Each of them picked up a hard problem with one hand tied behind his back, refused to put it down, and worked every issue to perfection until the answer became inevitable. Strip the genius mythology off and what’s left is something you can actually use: take the constraint, work it completely, don’t let go until it’s solved.
The Forensic Signature
I spend a good part of my working life now reading the record of things that broke — 8-plus years analyzing failures for attorneys under my Professional Engineer’s license, which is a long way of saying I get paid to point at what actually happened instead of what everyone wishes had happened. And the record on this is not ambiguous. I’m not offering you an opinion. I’m pointing at the cause-and-effect chain that shows up every time.
The chain runs like this. Unbounded brief, which produces an infinite solution space, which produces endless option-generation, which produces schedule slip, which produces cost overrun, which produces a late, expensive, forgettable result that a survey of customers would have rated as perfectly fine. Every link pulls the next. None of them is the villain alone. The villain is the missing boundary at the front, and everything downstream is just the void doing what voids do.
Now run the other chain. Bounded brief, which collapses the solution space, which forces concentration, which produces fast convergence, which produces lower cost, which produces a result with an actual point of view — the kind of thing that gets cited, licensed, remembered. Same engineers. Same talent. The only variable that changed was whether someone had the discipline to draw the lines before the work began.
Which lets me say the thing that gets me in trouble in design reviews, and the line I’d build this whole piece around if I had to keep one sentence: being a hero leads to a zero.
The culture worships the engineer who tears it all down and rebuilds from a clean sheet, who questions every assumption, who refuses to be limited by what came before. In the record, that engineer is expensive. The heroic from-scratch rebuild burns the budget, re-introduces untested risk into a program that was standing on solved ground, stretches the calendar, and lands — almost every time — on a result no better than what already worked. It is usually ego wearing the costume of ambition. The actual hero is the engineer who knows precisely what not to touch, who treats the solved problem as solved and pours every bit of capability into the bounded problem that’s actually open. That engineer files patents. That engineer ships. That engineer is invisible in the mythology and indispensable in the record.
The Aftermath, and the TRIZ Prediction
So here is what to do with this on Monday morning, because a teardown that doesn’t change your behavior is just entertainment.
Before any design effort starts, ask two questions and refuse to start until both are answered.
First: What is already solved that we can lock down? Find every component, subsystem, geometry, and interface that has already been proven, tested, and validated — every part that already has five years of paid-for research amortized into it. Declare it fixed. Take it out of the solution space entirely. Re-solving it is not innovation, it is the most expensive form of waste there is, because you pay twice for the same answer and inherit a fresh set of unknown failure modes for the privilege.
Second: What are the hard boundaries of the part that’s actually open? Define the playing field with no mercy and no fuzz — weight, cost, thermal envelope, interfaces, schedule. Make the constraints explicit, specific, and non-negotiable. Then hand that bounded problem to your engineers and get out of the way. You will not get wandering. You will get an answer, faster and cheaper than the blank sheet ever produced, and very often more elegant — because elegance, as the bench taught me, lives in the boundary.
Now let me ponder it, because I don’t trust an answer that arrives this clean.
The satisfying conclusion is constraints always help. That is not true, and pretending it is would set you up to fail in a more sophisticated way. For this answer to be wrong, only one thing has to be true: the constraint has to be the wrong one. A badly chosen boundary does not produce a diamond. It produces a fractured mess, on schedule and under budget, which is arguably worse than a slow mess because nobody catches it until it’s shipped. Carbon under the right constraint is a diamond. Carbon under the wrong constraint is still just graphite — useful, sure, you can write with it, but nobody builds a life around it. The lesson is not that adding pressure makes things valuable. The lesson is that the constraint has to be the true one — drawn from the physics of the problem, not from someone’s calendar anxiety or someone’s need to look decisive.
Which is the TRIZ-forward prediction, and the assumption I’m now holding that I wasn’t holding when I started writing this. The next real discipline in engineering management is not constraint application — we’ve half-learned that. It’s constraint selection: the craft of choosing which boundaries to impose, at which level, at which moment in the program, so that the solution space collapses toward the masterpiece instead of toward the fracture. Almost nobody teaches it. Almost everybody gets it wrong, usually by importing constraints from the budget meeting instead of deriving them from the problem. The organizations that learn to choose their constraints the way Wright read his hillside — that’s who builds the next frontier. Not the ones with the cleanest sheet of paper. The ones who know exactly which boulder to leave in the middle of the lot, and then refuse to move it.
Give your engineers the boulder. Then watch what a real constraint worker does with one hand tied behind their back.
Herbert Roberts, P.E. is a licensed professional engineer with 32 years in aviation research and development across two companies.