The Importance of Being Imminent
Quantum Computing, Probability, Possibility, Unicorns, and the Industrial Production of Bullshit
The Importance of Being Imminent
Quantum Computing, Probability, Possibility, Unicorns, and the Industrial Production of Bullshit
Quantum computing is the only industry in which the absence of the promised computer is routinely presented as evidence that the computer is progressing beautifully.
For several decades now, we have been assured that quantum computers are coming.
They are always coming.
They have developed the punctuality of an aristocrat and the evidentiary standards of a séance.
One is forever told that the revolution is near, that the breakthrough is imminent, that the latest machine has changed everything, and that anyone insufficiently impressed has failed to understand the future. The future, for its part, has shown admirable discipline in refusing to arrive.
What we actually have are elaborate experimental devices, lavishly refrigerated, ceremonially photographed, and christened “computers” by people who understand perfectly well that the word does most of the commercial labour.
This is rather like placing a saddle on a laboratory rat and announcing progress toward cavalry.
The rat is real.
The cavalry is not.
Yet somehow there is already a consultancy practice for equine transformation.
The Computer That Must Not Be Asked to Compute
There is something wonderfully modern about a machine that must be praised for what it may someday do rather than judged by what it does now.
A normal computer is subjected to vulgar questions.
What can it calculate?
How fast?
At what cost?
With what reliability?
Can I use it?
Can anyone else use it?
Does it solve a problem I actually possess?
Such questions are considered almost indecent in quantum computing.
There, one is encouraged to admire the architecture, contemplate the possibility, appreciate the roadmap, and never spoil dinner by asking whether the object has yet become the thing whose name appears on the brochure.
The machine is not judged as a machine.
It is judged as a prophecy.
Prophecies enjoy excellent margins.
Possibility: The Most Abused Word in Technology
There is a small logical distinction that has caused immense financial difficulty for anyone trying to sell the future.
A thing may be possible without being probable.
A thing may be probable without being imminent.
A thing may be imminent without being useful.
A thing may be useful without being economic.
And a thing may be economic in one absurdly narrow demonstration without being a general revolution in anything whatsoever.
This is not advanced mathematics.
It is grammar with consequences.
But in quantum marketing, possibility is dragged into a dark room, given several billion dollars, and emerges dressed as inevitability.
“A sufficiently capable fault-tolerant quantum machine could, in principle, perform certain classes of computation differently from classical machines.”
Very good.
That is a proposition.
Then the marketing department arrives.
“Quantum computing will transform industry.”
A promotion has occurred.
Then the consultant arrives.
“Quantum computing is transforming industry.”
A miracle has occurred.
Then the conference organiser arrives.
“Your board needs a quantum strategy now.”
A robbery has occurred.
The Unicorn with a Cryostat
The quantum computer is frequently defended with the observation that the underlying physics is real.
Of course it is.
So is refrigeration.
So are lasers.
So are probability amplitudes.
So are superconducting circuits, trapped ions, photonic systems, noise, decoherence, measurement error, and all the other magnificent inconveniences that reality places between a mathematical possibility and a useful machine.
The question was never whether physics exists.
The question is whether the object being sold under the magnificent title “quantum computer” deserves the noun in anything like the ordinary commercial sense people are invited to imagine.
A unicorn may contain perfectly real carbon atoms.
This does not settle the unicorn question.
Yet quantum marketing regularly performs exactly this trick.
The components are real; therefore the advertised beast is treated as real.
By that standard, I own a cathedral because I possess several bricks.
Dorian Gray’s Roadmap
If Dorian Gray had founded a quantum-computing company, his portrait would have aged horribly in the attic while the roadmap remained forever three to five years from perfection.
This is the great aesthetic achievement of the field.
The calendar changes.
The promise does not.
Every few years there is a new threshold, a new architecture, a new fidelity figure, a new qubit count, a new error-correction result, a new “advantage,” a new benchmark, a new chip name, and a fresh photograph of an object resembling a chandelier designed by a committee of physicists and Bond villains.
And always the same conclusion:
We are closer than ever.
This statement has the useful property of being true even if one moves forward by an inch during a journey of a thousand miles.
It is therefore perfect for investor relations.
One can be closer than ever indefinitely.
Indeed, whole careers may be built on it.
A Benchmark Is Not a Business
One of the more charming customs of quantum promotion is the ceremonial benchmark.
A machine is given a highly specific task, usually one chosen with intimate knowledge of what makes that machine look impressive, and then compared with a classical approach under conditions sufficiently specialised to require a paper, three appendices, and a public-relations team.
The result is announced as progress toward supremacy, advantage, utility, or some other noun that once had an ordinary meaning.
This is scientific theatre of a very high order.
It may even contain excellent science.
But the commercial inference is often magnificent nonsense.
A greyhound can defeat a submarine across a meadow.
This does not establish the end of naval power.
A toaster can outperform a supercomputer at making toast.
One hesitates to call this computational advantage.
Yet give the toaster a cryogenic enclosure, a Greek letter, and a venture round, and someone will be willing to keynote it.
The Qubit: A Number in Search of Meaning
Marketing adores quantities because quantities can be made larger.
A larger number is almost always superior to an explanation.
Thus we have been treated to qubit counts with the solemnity once reserved for battleship tonnage.
More qubits.
More progress.
More future.
More opportunity.
More slides.
The difficulty is that a qubit is not a magic bean.
A thousand unreliable things do not automatically become one reliable thing by being counted enthusiastically.
Noise matters.
Connectivity matters.
Error rates matter.
Control matters.
Logical operations matter.
Error correction matters.
Scale matters.
The actual algorithm matters.
The cost of obtaining a robust logical computation matters.
And, rather awkwardly, whether the calculation is useful to anyone outside the laboratory matters.
These qualifications are terribly hostile to typography.
“1,000 QUBITS” looks excellent in forty-eight-point font.
“The operational meaning of this number is architecture-dependent and should not be confused with useful fault-tolerant logical computation” tends to spoil the brochure.
So the brochure wins.
It usually does.
Error Correction and Other Socially Inconvenient Details
The quantum-computing story becomes particularly entertaining when one reaches error correction.
The promised machine must perform delicate operations on information that has the manners of a nervous duchess: touch it incorrectly and the evening is ruined.
Noise arrives.
Errors arrive.
Decoherence arrives.
Control imperfections arrive.
Reality, unlike venture capital, is relentlessly non-promotional.
So one needs error correction.
And error correction needs redundancy, control, decoding, overhead, engineering, and all the dreary physical detail that separates “could exist” from “does exist.”
This is where the unicorn develops an accountant.
Suddenly the magical beast needs infrastructure.
It needs thousands or perhaps vastly more physical resources to produce useful protected logical behaviour, depending on architecture and target.
It needs operations accurate enough to survive long computations.
It needs control systems able to scale without turning the miracle into a statistical accident.
It needs to do all this repeatedly.
Reliably.
Economically.
Preferably before the classical competitor improves again.
This is less romantic than “quantum changes everything.”
It is therefore discussed after lunch.
The Classical Computer, Rudely Refusing to Die
Quantum prophecy suffers from an especially irritating enemy: the ordinary computer keeps getting better.
This is extremely poor form.
The classical machine was supposed to wait politely at the side of history while its quantum successor approached in ceremonial dress.
Instead, classical hardware improves.
Accelerators improve.
Algorithms improve.
Approximation improves.
Distributed systems improve.
Specialised chips improve.
Memory systems improve.
Compilers improve.
Software improves.
And mathematicians, those dreadful people who refuse to respect product roadmaps, occasionally discover that a supposedly impossible problem can be handled rather well by thinking harder.
The comparison therefore changes while the quantum machine is still being promised.
This is the part of the race nobody puts on the poster.
The target moves.
The competitor learns.
The “intractable” becomes less intractable.
The theoretical advantage remains theoretical while the practical baseline quietly becomes inconvenient.
There is nothing so destructive to a technological prophecy as a rival technology that insists on having engineers.
Probability Is Not Optimism with Mathematics Nearby
Consider the structure of the promise.
For the advertised revolution to occur, many difficult things must happen.
The hardware must scale.
Error rates must fall.
Control must remain precise.
Fabrication must become repeatable.
Logical qubits must become useful.
Error-corrected operations must become deep enough for valuable algorithms.
The algorithms must matter.
The classical alternatives must not erase the advantage.
The cost must be tolerable.
The machine must be operable.
The outputs must be reliable.
Someone must possess a problem worth paying to solve.
And, most offensively of all, the entire chain must work outside a press release.
If each of these is uncertain, then the probability of the final promised outcome is not obtained by selecting the rosiest assumption and multiplying it by enthusiasm.
A chain of uncertain events is still a chain of uncertain events.
In polite technological society, however, joint probability is treated like adultery: everyone suspects it exists, but mentioning it at dinner is considered vulgar.
Possibility Has Become a Financial Instrument
The deeper joke is that possibility itself has become monetisable.
One no longer needs to build the future.
One can securitise the expectation of it.
A sufficiently dramatic scientific possibility can support investment decks, consulting practices, national strategies, policy announcements, university centres, keynote circuits, “readiness” programmes, and corporate transformation initiatives long before the underlying machine is capable of performing the advertised economic role.
The future has discovered recurring revenue.
This is a considerable advance.
The machine may be uncertain.
The invoice is fault tolerant.
The Consultant’s Quantum Computer Already Works
There is, however, one quantum computer that unquestionably works today.
The consultant’s quantum computer.
It requires no cryogenics.
It has perfect uptime.
It can operate simultaneously in banking, shipping, defence, pharmaceuticals, insurance, retail, mining, manufacturing, agriculture, government, and luxury hospitality.
It scales instantly.
Its error rate is zero because no output is falsifiable.
Its core algorithm is breathtakingly simple:
IF executive fears missing future THEN sell quantum strategy.
Its secondary routine is equally elegant:
IF executive asks for commercial timetable THEN show roadmap with arrow pointing upward and right.
This machine is already profitable.
It may be the first genuinely useful quantum product.
Quantum Readiness: Preparing for a Guest Who Never Rings the Bell
Corporations are now encouraged to become “quantum ready.”
This phrase deserves preservation in a museum.
It is perfect.
It sells preparation for an object whose arrival date is unknown, whose economically useful form is uncertain, whose competitive applications remain contingent, and whose technical architecture may change before the strategy document has finished circulating through procurement.
It is preparedness without an event.
One imagines Victorian households being sold dragon-readiness audits.
“Madam, your estate is dangerously unprepared for large reptiles.”
“Have large reptiles been observed?”
“That is precisely why you must act now.”
The quantum-readiness business has improved on insurance by dispensing with the inconvenience of an insurable event.
“Could” Is Doing Heroic Work
Whenever reading about quantum computing, watch the word could.
It is the hardest-working employee in the entire sector.
Quantum machines could transform chemistry.
They could transform optimisation.
They could transform finance.
They could transform logistics.
They could transform materials science.
They could transform security.
They could transform medicine.
At this rate they may soon transform interior design.
“Could” is ideal because it contains no timetable, no probability, no cost model, no engineering closure, no commercial proof, and no obligation to survive contact with the future.
It is grammatically innocent and commercially magnificent.
The trick is then to place “could” in paragraph three and “will” in the headline.
By paragraph seven, “is” has usually appeared.
A tense change has become a technological revolution.
The Museum of Things That Were Five Years Away
Technology has a splendid museum containing things that were five years away.
Some eventually arrived.
Some arrived in forms nobody predicted.
Some became useful only in niches.
Some were overtaken by alternatives.
Some remained demonstrations.
Some died quietly while the keynote speakers moved on to the next inevitability.
The lesson should have been humility.
Instead, we invented larger fonts.
Quantum computing has benefited from this historical amnesia because each new generation of promises is treated as though the previous generation had never made any.
The forecast is perpetually reborn without original sin.
Yesterday’s “imminent” becomes today’s “early days.”
Yesterday’s “breakthrough” becomes today’s “foundational milestone.”
Yesterday’s revolution becomes today’s research programme.
Nothing fails.
It merely acquires a longer roadmap.
The Great Semantic Bailout
This is why the language matters.
If the machine under discussion does not yet possess the characteristics that ordinary people reasonably associate with the promised computer, the easiest solution is not to build the missing capabilities.
It is to redefine success.
A laboratory apparatus becomes a computer.
A demonstration becomes utility.
A benchmark becomes advantage.
A roadmap becomes delivery.
A research milestone becomes an industry milestone.
A possibility becomes a market.
And a market forecast becomes evidence that the technology must be real, because surely nobody would forecast billions of dollars for something insufficiently mature.
This is circular reasoning with excellent catering.
Bullshit Is Not the Same as Lying
The most sophisticated quantum hype does not require lies.
Lies are crude.
Bullshit is more elegant.
One simply assembles true statements in an order designed to produce a false impression.
Quantum mechanics is real.
True.
Quantum phenomena can be manipulated experimentally.
True.
Specialised devices can execute quantum circuits.
True.
Researchers have demonstrated technically interesting results.
True.
Therefore the world has quantum computers in the sense implied by decades of commercial promises.
No.
That last step entered wearing someone else’s invitation.
The argument resembles a gentleman who proves that because bricks exist, Notre-Dame has already been rebuilt.
One admires the confidence.
One should not hand him the restoration budget.
The Burden of Proof Has Been Cryogenically Reversed
Ordinarily, the person making the extraordinary commercial claim is expected to prove it.
Quantum computing has achieved something far more advanced.
The sceptic is required to disprove the future.
Ask whether the promised machine exists, and one is told that one does not understand the science.
Ask whether a demonstration scales, and one is told that scaling is precisely what the roadmap addresses.
Ask whether the roadmap will work, and one is told that the field is making rapid progress.
Ask whether rapid progress implies arrival, and one is shown another roadmap.
This is not an argument.
It is a Möbius strip with branding.
The Five Questions That Ruin a Quantum Dinner Party
The next time someone announces a quantum breakthrough, one need ask only five impolite questions.-
What useful thing can this machine do now that a serious user actually needs?
-
Does it do that thing better, cheaper, or faster than the best practical classical alternative?
-
Can it do so repeatedly, reliably, and at economically meaningful scale?
-
How much of the claimed result depends on carefully chosen benchmarks, post-processing, mitigation, filtering, or experimental conditions that do not generalise?
-
If the answer is “not yet,” why are we calling the promised object present tense?
These questions have a terrible defect.
They are comprehensible.
Nothing kills mystique faster than a question that can be answered in English.
The Importance of Being Earnestly Sceptical
There is no shame in research.
There is no shame in experiments.
There is no shame in exploring whether quantum phenomena can someday support machines of genuine computational value.
The shame lies in pretending that the journey and the destination are the same object because the distinction interferes with fundraising.
Research is allowed to be uncertain.
Engineering is allowed to be unfinished.
A hypothesis is allowed to fail.
A roadmap is allowed to slip.
A machine is allowed not to exist yet.
These are ordinary conditions of serious work.
The absurdity begins only when uncertainty is marketed as inevitability and delay is rebranded as progress toward an ever-receding commercial dawn.
The field does not need less science.
It needs less theatre.
It needs fewer unicorns.
And it could survive with considerably fewer men on stages explaining that physics has already disrupted an industry which has not noticed.
A Possibility Is Not a Probability
This is the whole trick, once the smoke has cleared.
Possibility is not probability.
Probability is not imminence.
Imminence is not utility.
Utility is not economic superiority.
A laboratory effect is not a computer.
A collection of qubits is not a revolution.
A benchmark is not a market.
A roadmap is not delivery.
A press release is not evidence of arrival.
And a machine that has been “nearly here” for decades should perhaps be judged with the same standards applied to every other creature said to exist but never quite available for inspection in the promised form.
The unicorn community has, at least, shown greater restraint.
No unicorn has ever demanded a national strategy.
No unicorn consultancy has advised the board.
No unicorn startup has announced a trillion-dollar total addressable market.
No government has yet warned of a strategic unicorn gap.
And no serious adult has pointed to a horse, attached a horn, and demanded that sceptics celebrate the arrival of mythical biology.
Quantum computing might one day become something extraordinary.
That possibility is precisely why it should be discussed honestly.
Until then, the most advanced quantum system in existence remains the marketing department: simultaneously certain and uncertain, profitable whether right or wrong, and capable of occupying two contradictory states until someone foolishly asks for a working machine.
At which point, naturally, the wavefunction collapses into a roadmap.