Now imagine that was the end. The story ends there. No deductions, no culprit, no account of what happened. Just fog, elegantly tolerated, while a man who was supposed to do a job admires the difficulty of it instead.

We recognize this, I hope, as absurd because it inverts the whole point of the detective. And yet some version of that sentence is said, without irony, in design reviews, leadership off-sites, and posted job descriptions every week. “Embrace ambiguity”, “Get comfortable in the gray”, “Learn to sit with not knowing”. These things are offered as wisdom, framed as a sign of seniority, the thing that separates the mature practitioner from the junior one who still wants to be told what to build. I want to argue that this is, at best, a half-truth wearing the costume of a deep one. And the fastest way to see why is to look at the real Holmes, because he is not the target of this essay. He is the standard it is built on.

The realist in the hunter's cap

It could be argued that the real Holmes is the most ambiguity tolerant character in English fiction.  He sits in uncertainty longer than anyone around him can stand to, refuses to leap to any conclusion. He withholds the verdict until the evidence forces it, and his famous line about eliminating the impossible comes at the end of the work, never the beginning.  If "comfort with ambiguity" means anything honest, he has it in abundance.

What must be acknowledged is that does not embrace ambiguity the way the slogan means. His goal is not to permit ambiguity to tag along as long as possible. He is embracing it in the way a realist embraces any fact. The mud is a fact. The locked door is a fact. The missing hour in the suspect's account is a fact. Most importantly the fog of the genuinely not-yet-known, is also a fact, right there on the table with all the others, and deserving no special reverence. He notes the shape of the uncertainty, fixes his mind on it, and then he burns off every part of it that deduction can reach, without exception, without sentiment, and accepts only what survives, even the residue of uncertainty.

The first cut: timing

There is a kernel of truth in the slogan worth saving before I take it apart. Serious work does begin in genuine uncertainty. You do not yet know the real user, the true constraints, or the goals hiding underneath the stated ones. Someone who panics at that, who demands false certainty on day one, is no use either. The capacity to keep working before the answers arrive is a real skill, and Holmes has it in abundance. All respect to Holmes.

But notice when the slogan applies its blessing. It blesses the fog at the front of the work, as a posture you put on before lifting a finger. The realist does the opposite. Acceptance of uncertainty earned, clearly communicated at the end, and only for the part of the fog that would not burn away. Front-loaded embrace is a costume. Back-loaded acceptance is a result. The honest question I am asking here is why so much fog survives through the front of the work in the first place, when the discipline for clearing it is no secret. The answer, most of the time, is not philosophical. Specifically it is the calendar.

Every unknown has an expiration stage, a point in the work past which it should no longer exist. Ambiguity around product strategy should be dead before requirements are written. Requirements ambiguity should be dead before the interface is drawn. The work is a sequence of stages, and each stage has a kind of fog that is supposed to be burned off inside it and never carried across the line into the next one. I submit that the real discipline is reducing each one on time.

The business timeline is what frequently breaks this. The need for an MVP at all costs by a date that was set before anyone understood the problem does not make the reduction happen faster. It means it never happens at all. The fog crosses its expiration stage still fully intact, smuggled into a stage where it has no business surviving, and the “embrace ambiguity” mantra stands at the border and waves it through. It stops being an epistemic posture and becomes the customs officer, stamping unfinished work as mature so that shipping it feels like seniority instead of what it is.

I want to be careful here, because the deadline itself is not the true villain of the story. A deadline is a fact, like any other, and the realist already works fast. Speed is part of the discipline, not the enemy of it. The failure is not having a timeline. The failure is letting the business clock quietly replace the uncertainty reduction schedule, meaning that the answer to the question, "is this fog of uncertainty cleared yet" inevitably becomes, "we are out of time”.

When time genuinely does run out, there is still an honest move and a dishonest one, and on the screen they look identical. The honest move is to ship the residue labeled, with the unresolved question named and the risk visible, eyes open. The dishonest move is to ship the same residue disguised as a decision, the fog dressed up as a choice somebody made. One is an error bar. The other is a lie about how much you know. The first respects whoever inherits the work. The second sets them up for failure, and unfortunately that path is often too tempting for those making the decisions. 

This is the moment the stowaway becomes the saboteur. A little tolerated uncertainty, riding along while the work begins, is harmless and even necessary. The deadline is what forces it past the stop where it should have been forced off the train, and the passenger you meant to drop three stages ago is now in the walls of the product. It feels, from inside the team, like self-sabotage, and it doesn't stay limited to the self for long.

The second cut: the cult status of ambiguity

The timing problem is significant, but it is actually the smaller of the two, and sets up the real one.

The deeper failure of the cult of “embracing ambiguity" is that it grants ambiguity a status it hasn’t earned. It treats fog as a special object, something you are meant to have a sophisticated relationship with, to commune with, to prove your maturity by tolerating. It asks you to revere the unknown.

Like Holmes, I am a realist and so I refuse ambiguity that dignity. Ambiguity is not a higher-order thing that demands a higher-order feeling. It is a condition of the room, no more sacred than a locked door or a missing witness. You do not admire the locked door. You do not develop a nuanced relationship with the suspect’s missing hour. You note them and you handle them. The whole posture of reverence is the error, because reverence is what grants amnesty to uncertainty, and that is what leaves the fog standing.

This is the load-bearing point of the entire essay, so let me put it plainly. You cannot eliminate every shred of unnecessary ambiguity until you have first refused to worship any of it. The designer who finds fog interesting, who lingers in the rich problem space because the difficulty flatters them, will always leave some of it standing, because some of it felt too meaningful to cut. The realist has no such weakness. Ambiguity gets no reverence, so it gets no exceptions. It is fact, not virtue, and that demotion is what makes its elimination possible.

Ambiguity should never be a designer’s ally.

What the carving actually produces

Say you accept this. Say you point the full discipline at the fog and reduce, without sentiment, everything that can be reduced. What do you get at the end?

It is tempting to say: the perfect solution, revealed. Carve away everything unnecessary and the right answer emerges from the marble, the way the sculptor's conceit says the statue was always inside the stone, waiting.

It’s poetic, but I do not believe that, and a realist cannot. A perfect solution waiting to be uncovered is a faith claim; it is the one thing on the table the realist has no fact to ground. Worse, it smuggles back in exactly the idealism we just threw out. If reduction reveals perfection, then nothing irreducible is ever left at the end, and the whole realist temperament, the willingness to accept the residue that would not burn, has nothing to stand on.

So here is what this process of hunting and carving up ambiguity actually produces. Not the perfect answer, but a perfectly bounded space. It produces a region with clean edges, where every reducible unknown has been reduced, the constraints are real and named, the user is real and known, and all the tradeoffs are on the table. The definition of the problem is perfect not because the space is empty of uncertainty, but because it is honest about the edges of its certainty. Inside a space defined that well, a real solution can finally be built, or often found, because the constraints have quietly implied it. But the perfection lives in the boundary, never in a guarantee about the contents.

This is an achievable thing, rather than simply being poetic, and I would submit science as the proof. Among other things, it does this with simple math. A serious result does not arrive claiming the unknown has been abolished. It arrives with error bars and confidence intervals. It drives with a formal and quantified statement of exactly how much ambiguity could not be reduced. That is the most honest thing a discipline can do: not pretend to a clean answer, but reduce all that can be reduced and then clearly communicate the size of the remainder. The detective accepts the residue by temperament. The scientist measures it and prints the number. Either way, the move is the same, and it is the opposite of both panic and pretense.

Why any of this is an ethical matter

Everything so far has been about craft, but this is where it enters the heavier realm of ethics, and it centers on a single property of ambiguity that does not ever seem to be said out loud. Ambiguity does not evaporate, it is conserved. When a team fails to resolve it, it moves on, or becomes entrenched, It might even spread, but It never disappears on It's own.

If the strategy is ambiguous, the requirements that descend from it will also be ambiguous. If the requirements are ambiguous, the interface built from them is ambiguous. If the interface is ambiguous, the person on the other side of the glass inherits every question the organization failed to answer. The fog does not dissipate as it travels downhill. It accumulates, and it pools at the lowest point, which is always the user, because the user is the one party with no power to send it any further down the line. They cannot escalate it, reassign it, or defer it to the next sprint. They simply absorb it, painfully, one confused session at a time.

This is the line I would carve over the door: any ambiguity the team does not resolve is eventually outsourced to the user. A designer can carry that fog for a while; that is part of the job, but the end-user should never have to, and yet they are the ones who almost always do.

It’s worth asking who chose the end-user for this role anyway No user has ever opened a confusing product and thought, "what a rich, ambiguous problem space this is." No one has ever felt grateful to a checkout flow for keeping them "comfortable in the ambiguity" of not knowing if their checkout is going smoothly or not. The thing we praise ourselves for tolerating in the conference room, or on LinkedIn, is also something we end up quietly shipping to people who never agreed to tolerate anything. Their confusion is our unfinished work, conserved and forwarded, arriving as cognitive friction, as a small daily tax on attention, as the slow erosion of trust that we then describe in a deck as an engagement problem to be handled in a future sprint.

One test, three verdicts

It helps to stop treating ambiguity as one thing. Not by sorting it into a tidy catalog of types, but by asking every patch of fog the same single question, the one Holmes asks of every fact in the room: has this survived honest investigation? The answer comes back in one of three ways.

Sometimes the fog is genuinely irreducible. You investigated in good faith, you reduced everything that could be reduced, and what is left is the real weather of the problem: a true unknown, a tension with no clean resolution, a future you cannot see from here. This you accept, the game Is afoot. You dress for the hunt, you track and contain it, you name it honestly inside the space. This is the residual uncertainty Holmes keeps as a map of the problem space, and the real thing the slogan was clumsily reaching for the whole time.

Sometimes, however, the fog is reducible and simply wasn't, for whatever reason. Nobody clarified the goal, named the owner, defined the metric, decided the priority, or agreed what a word meant. This is not the romance of an open problem. It is organizational debt wearing the costume of creative complexity, and it does not ask to be embraced. It asks to be cleared, with exactly the unglamorous clarifying question that a culture of comfort exists to keep anyone from asking. 

It must also be mentioned that sometimes the fog survives investigation because it was never meant to be investigated. Someone is maintaining the uncertainty, because the confusion is doing work for them. That is no longer a craft problem. It is a power problem, I have seen It, and it is a different essay than this one. I name it here only so the map stays honest: not all fog is an accident, and the same discipline that sorts the first two will eventually walk you right into the third.

The designer as realist

So no, I do not think the designer's job is to embrace ambiguity. I think the job is to hunt it down, Identify It, and reduce It until Is no longer a problem.

You track it through the language, where different people use the same word for different things and never notice. You track it through the requirements, where a passive sentence hides the fact that no one actually decided something critical. Through the interface, where every question left open upstream has surfaced as one more thing the user now has to work out alone. You reduce everything that reduces, without exception and without amnesty, including the fog that happens to be interesting, or flattering, or convenient to leave alone. You bound the space until its edges are clean and its residue is named. You accept what genuinely will not burn, without flinching and without dressing it up as something nobler than it is.

Most importantly, you refuse, as a matter of plain responsibility, to let the user become the storage container for the unresolved ambiguity the team was too comfortable to ask questions about.

The realist embraces ambiguity as fact, never as virtue, and reduces every reducible unknown, without exception. Honors the residual uncertainty that survives al investigation, and is honest about its size and nature. The realist remembers, at every step, that whatever fog is left unbounded does not disappear. It rolls downhill, and comes to rest on whoever has the least power to refuse it.

The value of the design realist is not "comfort with ambiguity". It is something far more valuable It is refusing to let your comfort with inadequately resolved uncertainty, with ‘ambiguity”, become someone else's problem to deal with.