The New Profile Razor Unit: Why Engineers Need Mark E. Smith
- David Turner
- Jul 23
- 6 min read
Updated: Aug 25
The Razor Unit: Why Engineers need Mark E. Smith
"You won't find anything more ridiculous than this new profile razor unit, made with the highest British attention to the wrong detail, become obsolete units surrounded by hail." Mark E Smith, The Classical (Hex Enduction Hour - 1982)
The pattern in the room
Anyone who has sat through a program review recognises the shape of it immediately. A team defends a tolerance three microns tighter than the design requires while the launch date slips past the point the business case still works. A supplier gets re-tooled because a surface finish looks wrong under studio lighting nobody outside the building will ever recreate. The costing sheet turns red and the meeting still spends forty minutes on a chamfer. None of this is laziness. It's usually the opposite: real skill, applied with total sincerity, to the wrong problem. You can't seriously tell someone to care less about their work. You can only ask them to check, without bias, whether the thing they are perfecting is the thing that was ever asked for.
Forty years on, the line still works because the pattern hasn't gone anywhere. It's there in organisations who would rather obsess over something invisible to the customer than solve the problem in front of them. It's there in the unspoken instinct that treats being profitable, reliable and repairable as somehow unsporting, as if pragmatism were a moral failure and refusing to compromise a virtue, even when nobody asked for the sacrifice.
Two disciplines, mistaken for one
The razor unit's failure isn't really an engineering or design problem. It's a category error. Herbert Simon drew the distinction in The Sciences of the Artificial (1969): natural science describes what already exists, while design is a science of what ought to be, dealing in judgements that satisfy rather than calculations that optimise. Donald Schön went further in The Reflective Practitioner (1983) and named the failure mode directly: "technical rationality," the assumption that professional problems are just theory and technique applied correctly to a well-defined question. His argument was that this collapses exactly where uncertainty and lived judgement matter most, and that good practitioners instead work by "reflection-in-action" - tacit, embodied, adjusted in real time. Michael Polanyi had already given the idea its cleanest line in The Tacit Dimension (1966): we know more than we can tell.
Engineering, at its best, is repeatable, testable, transferable between people because it's written down as a rule. That's exactly what the razor unit's makers had. What they didn't have was design judgement - the thing rooted in observation rather than specification, that knows when the rule has stopped serving the purpose it was built for. They applied technical rationality to a problem that needed reflection instead, and the result was an object with immaculate tolerances and no reason to exist.
Could a machine write Hex Enduction Hour in 2032?
Large language models are prediction engines working over text that already exists. Ask whether one could produce anything like the razor line, and the answer is no. Not because the wordplay is hard to imitate, but because the observation isn't secondhand. Smith didn't arrive at it by reading criticism of British manufacturing. He arrived at it by living in a city that was visibly collapsing around him, watching things fail in real markets, in real time, for real reasons, with nothing at a safe remove.
That's the piece an AI model can't supply: the exposure. This is Polanyi's tacit knowledge at its most extreme; knowledge that resists being told because it was never built out of things that could be told in the first place. It isn't a data problem, and it isn't waiting on more parameters. It's what a human mind is, dropped into interesting enough circumstances to say something true about them.
Lines of flight
Deleuze and Guattari, in A Thousand Plateaus (1980), had a term for how a system breaks new ground. Not replacement from outside, but a 'ligne de fuite', a line of flight: an escape along some unpredictable vector that destabilises the system's own coded structure from within. Their word for a system's pull toward hierarchy and fixed order was 'territorialisation'. A line of flight is what deterritorialises it, and crucially the escape isn't only negative. It produces something new in the act of fleeing.
Smith fits the description almost exactly. Working class, Prestwich, autodidact, no formal training, openly hostile to the polished, media-shaped assemblage that culture was supposed to run through. The refusal wasn't only posture. It produced a body of work with its own coherence, one that couldn't have been arrived at by staying inside the system and doing it better.
The harder problem isn't AI
Here's the part that matters more though. Even setting AI aside entirely, the conditions that produce someone like Smith are being dismantled.
He needed density: proximity to a wide, unchosen cross-section of people, repeated and unmediated exposure to a place changing under him. Streets, football grounds, the docks at Salford, the brutalist concrete of 20th-century Manchester. Those spaces are closing, and the people who most need the accidental collisions they generate are priced out of the cities that still have them, ending up instead on new-build estates staring at their phone.
The phone doesn't replace the exposure. It replaces it with something that looks like input but functions more like an algorithm than a street corner: curated, narrowed, optimised for engagement rather than truth, feeding you more of what you already responded to rather than what you didn't choose to see. Model collapse, in the real world. Smith himself, asked about mobile phones, called them worse than heroin: no let-up, no rest, no contemplation. A little cockroach in the brain, he said, that goes in, lays an egg, and comes out the other side¹. He landed on that image long before anyone had the vocabulary for the attention economy.
So it isn't robots replacing human insight. It's the human conditions for insight disappearing regardless, with AI as one more accelerant amongst several.
Why any of this matters
Because renewal needs destruction, and destruction needs someone willing to do it without sentiment. That's how a chamfer argument gets interrupted before it eats the launch date. That's how a consensus that everyone privately doubts gets said out loud.
This isn't a new idea. Joseph Schumpeter named the economic version of it "creative destruction" in Capitalism, Socialism and Democracy (1942). Capitalism's essential fact, in his account, is the constant demolition of old economic structures to make room for new ones. Trace it back even further through different traditions - the Hindu cosmology of Shiva, Buddhism's concept of impermanence, and Japanese aesthetics giving it physical form through wabi-sabi and kintsugi. Different traditions, arriving independently at more or less the same conclusion: nothing stays worth keeping by being protected from destruction. Smith naturally had his own, more succinct version, paraphrasing Disraeli: "What Manchester destroys today, London does tomorrow."²
A system that loses the capacity, or the willingness, to destroy what no longer works doesn't reach some calm equilibrium. It accretes. Bad decisions get built on top of older bad decisions, and consensus goes unquestioned because questioning it costs something nobody wants to pay. Nothing collapses dramatically. It just gets quieter, smoother, more agreeable, and slowly airless - sterile rather than oppressive, which is the more frightening version, because there's no obvious force to work against. Just a long subtraction of friction until there's nothing left to push against. This is why engineers need Mark E Smith.
The counterweight is smaller than any of this
It only works if it's been earned. There's a version of contrarianism that's just as borrowed as the consensus it opposes, rejection as its own performance, its own comfortable tribe. That isn't what Smith was doing with the razor, or the phones, or Manchester. The value was never in the negation itself; it was about arriving at the position from his own observation, rather than adopting it because it had become sanctioned to hold.
That's the scarce thing, and it's as scarce in engineering and product teams as it is anywhere else. Not the capacity to disagree - plenty of people manage that, often loudly for effect. The scarce thing is disagreeing from your own ground, whether the target is a new profile razor unit, a chamfer three microns too tight, a smartphone, or the next enthusiastic claim that AI or whichever other au courant tech is about to change everything.
Postscript
A non-negotiable point before this goes anywhere: Mark E Smith would have hated, or at least laughed at, every word of the above. Turning a throwaway line about a razor into a tour through Schumpeter, Deleuze, Guattari and Polyani is precisely the kind of overworked, self-important intellectualising he spent a career cutting down - quite rightly too. He'd have had a much better way of putting all of this into a couple of acerbic lines, which in a roundabout way proves the point entirely.
David Turner is the founder of Kói, an independent technology consultancy advising investors, founders, and boards on technology assessment and engineering judgement in high-stakes decision-making.
You can reach him at: enquiries@dkoi.design
© Kói Holdings Ltd 2026. All Rights Reserved.
