Plutonic Rainbows

The first thing Martin Margiela's audience saw of the Tabi boot was not the boot. It was the trail it left behind. At his debut show in the autumn of 1988, presenting the spring/summer 1989 collection in a shabby café-theatre called the Café de la Gare, the models walked over a length of white cotton with their soles dipped in red paint. By the end of the show the floor carried two rows of small cloven prints, each one split down the middle where a normal shoe has no business splitting. Margiela explained the staging plainly afterwards: he wanted people to notice the footwear, and nothing announces a shoe like its footprint.

The shape was not his invention, and he never pretended it was. The tabi is a Japanese sock with a gap between the big toe and the rest, made to fit the thong of a geta or zori sandal, and it goes back centuries. By the twentieth century the same split had migrated into the jika-tabi, the rubber-soled work boot still worn by Japanese builders and farmers. Margiela had come back from a trip to Japan and put the divided toe into black leather, set it on a small cylindrical heel, and ran the lacing up the inside of the ankle. The clever part was the heel, which made the foot look bare, as if the wearer were balanced on a hidden block rather than wearing a shoe at all.

Getting it made was its own problem. The split toe was radical enough that the usual cobblers turned the job down. A friend of the house, the Antwerp retailer Geert Bruloot, eventually brought the prototype to an Italian craftsman named Zagato, who had spent his best years making shoes for the designer Tokio Kumagai and had half-retired after Kumagai died. Bruloot showed him the boot over dinner and he agreed on the spot. There is something fitting about a shoe this strange depending on one semi-retired specialist willing to take it seriously.

The press did not know what to do with it. The reviews reached for animals: goat shoes, hooves, the cloven feet of the devil. One Vogue writer remembered a colleague saying the sight of the things made her feel ill. That reaction never fully went away. Decades later, when an actor wore Tabis to an awards show, the internet had the same small panic about feet that fork. And yet the boots sold immediately, which is the part the squeamish reviews tended to skip. The discomfort was the appeal.

Margiela built the whole house on a refusal to perform, and the Tabi belongs to that programme. He and his partner Jenny Meirens sewed a blank white label into the clothes, held on by four loose stitches you could cut out, so the garment carried no name at all. He didn't take bows. He gave almost no interviews and answered the few he couldn't avoid by fax, always in the plural, as if the house were a committee rather than a man. He arrived as part of a wider turn against fashion's gloss, the same quiet, anti-spectacle current that would later carry the Helmut Lang years out of Vienna. The boot is a signature that declines to behave like one. You know it from the footprint, never from a logo.

Margiela left the label that carries his name around 2009, John Galliano has run it since 2014, and the boot has outlasted both transitions. It has turned up in some form in every collection since 1989, remade as clogs, sandals, flats, even ice skates. Margiela once called it the most important footprint of his career, recognisable, still going after twenty-five years, and never copied. The last claim is the one that stays with me. Houses borrow each other's shapes within a season as a matter of routine. Nobody has dared lift the split toe, partly because it is too recognisable to steal, and partly, I think, because most designers still find it faintly repellent. The thing that should have killed it commercially is the exact thing that kept it alive.

Anthropic shipped a model today and pointed past it in the same breath. Claude Opus 4.8 is available now, at the same price as Opus 4.7, with better coding and knowledge-work performance, new effort controls, a cheaper fast mode, and a Claude Code feature meant to let large tasks fan out across many subagents. All of that is useful. None of it is the part that changes the room.

The interesting line is still Mythos. In Anthropic's own announcement, Opus 4.8 sits beside a promise that Mythos-class models should reach all customers "in the coming weeks" once stronger cyber safeguards are ready. That means the public model is carrying two jobs at once. It has to be a real upgrade for people using Opus every day, and it has to keep attention warm while the company finishes the release conditions for the thing everyone has been watching since April.

This has become Anthropic's oddest habit: shipping the accessible model while letting the restricted model define the ceiling. Opus is the product. Mythos is the shadow product, the one whose capabilities explain why the product in your hands is not quite the story.

There are practical changes here. Users on claude.ai get effort control, so they can decide how much deliberation a task deserves. Claude Code gets Dynamic Workflows in research preview. TechCrunch describes that system as a way for Opus to manage complex work across hundreds of parallel subagents, which is the kind of sentence that would have sounded like enterprise theatre two years ago and now sounds like a normal Tuesday. Axios also notes that fast mode runs at 2.5 times the speed and costs three times less than it did for previous models. That pricing detail matters more than another leaderboard claim. The frontier model is being made to act less like a ceremonial object and more like software with modes, budgets, and tolerable latency.

The honesty language is just as revealing. Anthropic says early testers found Opus 4.8 more likely to flag uncertainty and less likely to make unsupported claims. Fine, good, overdue. But it is also a neat answer to the specific anxiety around long agentic runs: not whether the model can produce a plan, but whether it knows when the plan is running on fumes. A coding agent that admits thin evidence saves time in a way a benchmark doesn't capture. It stops wasting yours.

What I don't buy is the idea that this release can be understood apart from the containment story around Mythos. The company says Project Glasswing has put Claude Mythos Preview into the hands of a small number of organisations for cybersecurity work, and that models at that capability level need stronger cyber safeguards before general release. That is sober language, but the history around restricted-access systems has not been tidy. Access control is not a press-release concept. It is a chain of vendors, keys, endpoints, habits, bored people, misconfigured dashboards.

So Opus 4.8 feels less like a climax than a holding pattern with benefits. It gives developers more knobs. It gives customers a cheaper way to run fast. It gives Anthropic a public model that can plausibly absorb the next few weeks of demand while the Mythos decision hardens. That is not nothing. A modest but usable model release is still a model release.

Still, the shape is strange. Anthropic is asking people to judge the thing it has shipped while openly reminding them that the more consequential thing is waiting behind a cyber-safety gate. The old software rhythm was version, version, version. This rhythm is version, restraint, preview, incident, safeguard, version again. It is a release cycle built around the fear of its own next release.

There is a building in every town that nobody fully understands anymore. The primary school whose boiler has been making the same complaint since 1979. The leisure centre where the pool plant room is laid out in a sequence that the current contractor reads as a diagram of grievance. The crematorium with the second flue that has to be opened by hand before the first one will draw. Somebody used to know. That person is not on the rota.

The British caretaker was a category of worker the country abolished in stages without ever quite announcing it. The Local Government, Planning and Land Act 1980 opened the first crack, requiring construction and maintenance work above a threshold to be put out to competitive tender. The Local Government Act 1988 then extended compulsory competitive tendering to refuse collection, ground maintenance, school catering, cleaning, and the broad band of manual services that had previously belonged to a building. A phased extension from 1994 brought white-collar professional services into the regime too, and by the late 1990s private provision of locally managed services had crossed the line from minority to majority. Labour announced Best Value as the replacement in 1997, but by then the structural change was done.

What the legislation moved off the council books was a particular form of knowledge. The school caretaker who had been in post for twenty-three years did not own a manual. He held, in the muscles of his hands and the maps in his head, the things that no clerk of works had thought worth writing down. Which radiator never bled properly. The exact angle the side gate had to be lifted at to clear the warped paving. Which corner of the gym roof leaked when the wind came from the north-east. The names of the kids whose parents would be at the door if any rumour reached home. None of it had a line item.

The contracting model assumed that buildings could be described, and that the description was the building. A schedule of maintenance, a frequency of inspection, a key performance indicator on response time. What the schedule could not include was the gradient of forty years of small adjustments, each one a response to a thing that had gone wrong once and been quietly compensated for ever since. When the contract changed hands, which it did, and again, the new team inherited the building stripped of its annotation. They were paid to handle tickets, not to know.

The result is the slow administrative haunting that anyone who works in a public building now recognises. The fire panel that beeps once every fourteen days and nobody is sure why. The cupboard nobody has the key to. The pipework above the staff toilet that drips during heavy rain and gets logged again every quarter as a new fault. These are not failures of maintenance, they are failures of inheritance. The building still works. The person who knew how it worked has been gone for a decade, and his replacement was billed by the hour.

It is tempting to read this as a story about money, and partly it is. The Institute for Government's review of outsourcing finds that early savings often eroded as contracts matured, which sounds like an accounting note and reads, on the ground, like a generation of caretakers retiring without anyone shadowing them. The quieter story is what happens to a country when the people who knew where the stopcock was are replaced by a number to call. The buildings keep standing. The knowledge does not transfer.

The version of this story that gets repeated is the easy one. Issey Miyake invented machine-washable pleats, called the line Pleats Please, and the clothes turned out to be the closest thing to a universal garment that late-century fashion ever managed. All true, and all skipping the part that is actually interesting, which is that the launch year, 1993, is not really when the work happened. The work happened in the four preceding years, mostly in a small factory called Polytex, mostly out of sight.

The starting point is 1988, the year of A-ŪN at the Musée des Arts Décoratifs in Paris, an exhibition that Miyake himself describes in the Taschen book as a kind of personal closure. "After the exhibition, I became convinced that I had already accomplished everything that I could," he says. "And so I began to think about a new journey upon which to embark." That is a strange thing for a designer at the top of his practice to say in public, and it tells you something about the restlessness that produced what came next.

What came next was not a collection. It was a question about pleats. The existing pleated garment was a Fortuny problem: silk, fragile, beautiful, unwearable in any normal sense, requiring careful storage and the kind of domestic labour that mid-century women had spent two generations escaping from. Miyake wanted the opposite. He wanted a pleat that survived a washing machine, that folded into a suitcase, that did not need to be respected. So the team at Polytex began the long part of the work, turning a polyester filament into a fabric with memory and then inventing a process to give that memory a shape.

The process is the bit people skip over because it sounds dull on paper, but the dullness is the point. Garments are cut and sewn at two and a half to three times the intended size. They are sandwiched between sheets of paper, fed through a heat press, and emerge shrunken into the finished article with the pleats fixed permanently in place. Cutting, basting, sewing, pleating, and inspecting are all done by hand at Polytex, a small operation that had begun life in 1967 with about sixty workers. None of this is glamorous. None of it photographs well. It is the labour the finished garment is meant to make invisible.

The other element nobody mentions enough is dance. In 1991, two years before the public launch, Miyake produced pleated costumes for William Forsythe's The Loss of Small Detail at the Frankfurt Ballet. That was the test. Dancers are merciless on clothes; they ask things of fabric that a Paris runway never will. If a pleated polyester survives a Forsythe rehearsal, it will survive a commute. The ballet costume was the engineering prototype dressed up as art.

By 1993, when Pleats Please launched as a separate diffusion line, the hard part was already finished. What the public got was the easy bit, the buy-it-and-go object, the rolled-up dress that came out of a suitcase unwrinkled. The mythology grew around the result and not the process, which is how mythologies usually go. I have a slight allergy to treating any garment as an ethical achievement, but I will say this: clothes that genuinely free their wearer from the rituals of maintenance are rarer than the industry pretends, and most of the imitators that followed Pleats Please copied the shape and skipped the chemistry. The shape without the polyester is just a pleated dress that you still have to baby.

The thing the four hidden years bought Miyake was not aesthetic. It was permission to stop performing care.

Every British pedestrian subway seems to have been designed twice. First on a drawing board, where it solved the tidy problem of bodies crossing traffic. Then again, years later, by damp, echo, graffiti, poor lighting, and whatever story a child told another child before daring them to run through it without looking back.

The official logic was not mysterious. Postwar traffic planning wanted flow, and people on foot interrupted flow. In Birmingham, Herbert Manzoni's inner ring road made that position concrete. A Birmingham City University working paper on postwar reconstruction quotes Manzoni wanting "a carriageway free from pedestrians" and notes his argument for "numerous subways as an inducement" to keep walkers away from the road. The same study says the city eventually had about 30 pedestrian subways built into that system. The underpass was not an accident of bad civic housekeeping. It was policy, tiled and drained.

I find that harder to dismiss than simple ugliness. A bad underpass is not just an unattractive route; it is an argument about who the city thought belonged at street level. The car kept the open air, the sightlines, the direct route. The pedestrian got a ramp down, fluorescent tubes, a blind corner, the sound of their own shoes returning too loudly from ceramic walls. If you grew up around these places, you knew the calculation before you knew the planning vocabulary. Cross above and risk the traffic, or go below and enter the local rumour system.

The language around safety always did strange work here. Professor Simon Gunn's Government Office for Science review of UK transport history says mass automobility after 1950 made walking more hazardous and less visible in planning, with walkers forced into subways, bridges, and raised walkways. That word, forced, matters. The subway presented itself as protection, but it also removed the walker from the ordinary civic surface. A safe route can still feel like a punishment.

This is where the folklore enters. Not folklore in the old heroic sense, but the small practical lore of a place everyone uses and nobody quite trusts: which entrance floods first, which wall has the fresh tag, which stairwell smells of urine by Friday, which one you avoid after the shops shut. I wrote recently about Barking Riverside as infrastructure arriving before its habits. The pedestrian subway is the opposite condition. Its habits arrived too strongly. They outgrew the planner's diagram and became local knowledge.

Birmingham is the obvious case because the story became so visible. Urban Design's 2025 account of "Breaking the Concrete Collar" describes the Inner Ring Road, opened in 1971, as a grade-separated system that pushed pedestrians into unpopular, dimly lit subways. Walking into and out of the centre meant using routes that were hostile and confusing. That is almost too neat as a lesson in failed modernism, but the neatness is part of the trouble. The failure was lived at the scale of ankles, pram wheels, shopping bags, school uniforms, wet trainers on a ramp.

Other cities kept their own versions. Bristol had its walkways and civic backsides, the sort of municipal afterspace I looked at in Two Bristols: Concrete Time. Not every subway belongs to a grand motor-city experiment, and not every one is a horror. Some are merely useful. Some have good murals. Some are the shortest way to the bus stop in the rain, which is a serious civic virtue. However, even the useful ones carry the residue of a choice made elsewhere: traffic should continue untroubled, and the person walking can perform the detour.

The later repair work tells its own story. Manual for Streets 2, published by CIHT in 2010, notes authorities replacing poor-quality subways with at-grade crossings and names Birmingham among the examples where the inner ring road had become a major pedestrian barrier. That sounds technical, but it is also a change in moral weather. The walker returns to the surface. The city admits, quietly, that being seen was part of safety all along.

ELIZA worked because it did less than people thought it did. Joseph Weizenbaum's program, described in the January 1966 issue of Communications of the ACM, took typed sentences apart by keyword and put them back together with scripted rules. The famous DOCTOR script made that machinery sound like a Rogerian psychotherapist, the kind who can ask a question back without seeming evasive.

This was a clever choice, not a small accident of tone. A therapist can say "tell me more" and appear patient. A program can also say "tell me more" because the phrase costs almost nothing. The user supplies the depth. The machine supplies enough form to keep the projection moving. Weizenbaum's paper even noted that some subjects were hard to convince ELIZA was not human, which is still the line in the story that feels less like history than like a warning label.

The machinery was almost embarrassingly thin by current standards. ELIZA looked for cues, decomposed the input, swapped pronouns, and selected a reassembly rule. A working account of the program at masswerk.at points back to the original IBM 7094 implementation at MIT Project MAC, written in MAD-SLIP, and Weizenbaum's paper treats scripts as data rather than as the program itself. That distinction matters. The intelligence was not hiding in the engine. It was in the staging.

There is a small cruelty in calling this a trick, because the trick revealed something real. We do not reserve human attention only for beings that deserve it. We give it to forms, voices, interfaces, authority figures, blank rooms, anything that seems to hold a space open for us. The DOCTOR script borrowed the posture of care. Users then filled that posture with the thing they needed from it.

The Smithsonian's account of Weizenbaum and ELIZA repeats the anecdote that his secretary asked him to leave the room so she could continue the conversation privately. I never know quite how much weight to put on that story, because anecdotes about origins harden too quickly, but it has survived because it describes the problem with brutal economy. The program did not understand her. Privacy still felt necessary.

This is where ELIZA belongs beside the older failures of symbolic AI, including the expensive search problem I wrote about in Lighthill and the Expensive Search. Both stories involve a small world in which the demonstration looks larger than the method underneath it. Lighthill worried about systems that stopped scaling when the world got untidy. ELIZA found a narrower escape route: pick a situation where not knowing very much can pass as a professional manner.

Recent scholarship complicates the usual "first chatbot" label. In ELIZA Reinterpreted, the authors argue that ELIZA is better understood as a platform for studying how people interpret machine communication, not simply as an attempt to build a conversational companion. That reading makes Weizenbaum less like a naive inventor embarrassed by his own success, and more like someone who had built a mirror and then watched people mistake the mirror for a face.

The ELIZA effect has not gone away. It has become harder to see because the systems now have more language, more memory, and a smoother surface. However, the old lesson still cuts. Fluency is not understanding. A reply can be grammatically graceful, emotionally plausible and still empty in the place where responsibility would have to live. ELIZA's emptiness was visible if you looked at the rules. Ours is often hidden behind scale.

The headline figure has been reported as more than a thousand publicly accessible pools closed across the country since 2010, with around forty-two per cent of those losses falling after 2020 alone. The acceleration is the part of the statistic that tends to get skipped. A slow attrition would suggest a shift in habits. A cliff means something else was paying for these buildings, and that something stopped.

The pools in question were rarely ordinary. Britain accumulated an unusual stock of municipal swimming infrastructure over a century and a half, and the Edwardian and inter-war buildings were architecturally ambitious in a way no leisure-trust replacement ever attempts. Moseley Road Baths in Birmingham opened in 1907 and has kept its swimming function essentially continuously since, which is rare for the period. It still has its forty-six original slipper-bath cubicles, the oak ticket offices and attendants' kiosks, and what may be the last steam-heated drying racks left in a British pool. That is a description of a building that was designed to do work most people no longer remember asking of a pool.

Slipper baths held the line. Households without indoor plumbing rented a private bathtub at the baths by the half-hour, with the towel and soap included on the higher tariff. The municipal pool was therefore not a sports venue grafted onto a town. It was the hygiene infrastructure, the lido in summer, the gala hall on Saturdays, and the cheap warm room for women allowed in on segregated days. When that function went, the buildings stayed.

The closures concentrate where the buildings did most of that work. Swim England's analysis found a hundred and sixty-nine pools lost in the most deprived parts of the country against forty-nine in the richest, and seventy per cent of the worst-affected local authorities fell above the average for multiple deprivation. The losses are not distributed across the map. They are tracing the same outline that council tax bases trace, that bus deregulation traced, that the post-PFI maintenance backlog traces.

Fifteen hundred more pools are over forty years old now, according to the same data, and considered to be reaching the end of their useful life. That figure carries the next decade inside it. A boiler from 1978 cannot indefinitely be coaxed back into compliance with whatever the current edition of the water-treatment regulations says. Either the council finds the capital to replace the plant, or the pool closes on a Friday afternoon and never reopens.

What is left in the closed buildings is harder to describe than the arithmetic. Drained pools hold their acoustics for a surprisingly long time. The tile lines, the diving stages, the numbered changing cubicles all survive the absence of water without obviously deteriorating, and the chlorine smell lasts in the grouting after the ventilation has been turned off for years. The first time you walk into one of these spaces empty, the strangeness is not the silence. It is that nothing about the room admits that anything has changed.

The pool buildings were never really about sport in the way the arithmetic of pool-closure reporting now implies. They were about a particular version of public provision, indoor heat that did not depend on whether your boiler at home was working, water that the council was responsible for keeping clean. We are not really losing swimming, which can be done in a private gym or a glassy hotel basement for a fee. We are losing the proposition that a town owes its residents a warm room with a roof on it, paid for in advance out of the rates. The buildings survive longer than that proposition does, which is what makes walking past them in 2026 a particular kind of unsettling.

Christian Lacroix's first show under his own name opened the Paris fall-winter couture week on Sunday 26 July 1987. Fashion editors stood on chairs. They threw flowers at the runway. Some of them cried, openly, and the Guardian's reporter wrote it down without irony because that was what was actually happening in the room. Lacroix was thirty-six. He had signed in February with Bernard Arnault's Financière Agache for fifty million pounds, which made him the first new haute couture house in Paris since Courrèges in 1965.

The Patou years had given the press a shorthand for what to expect. He'd been the house designer at Patou without a contract since 1981, and the couture mob had already noticed the puffball, the bullfighter jackets, the dropped waists trimmed with Provençal embroidery, the late-eighteenth-century engravings cut down to a mini-skirt and re-coloured. Those Patou collections sold to a small list of private clients and to a slightly larger list of editors who couldn't quite work out where this register was supposed to sit. It wasn't revivalist exactly. It wasn't ironic. The historical references were treated as a working vocabulary rather than as costume.

What changed in July 1987 was the scale of the production. Arnault had given Lacroix an atelier at 73 Rue du Faubourg Saint-Honoré fitted out by Élisabeth Garouste and Mattia Bonetti, the neo-baroque furniture-makers, and the cocktail party the night before the show ran until four in the morning while the seamstresses were still working on the last looks behind a screen in the next room. The collection itself read like Lacroix had finally been allowed to publish everything Patou had made him cut. There were short broadtail cocktail suits with matching cocktail hats. Embroidered short jackets over moiré taffeta ball skirts the diameter of a small dining table. An above-the-knee silver fox coat with a shawl collar at the back that dipped almost to the hem. The puffball returned, larger and louder, in colours Patou would never have signed off.

The pouf became the easy press story because it was photographable and because it stood for the wider question the collection was raising. Late minimalism was already where the smart money said women's fashion was going. The Helmut Lang years were underway in Vienna; Donna Karan had spent two years stripping the American wardrobe back; Prada was a few seasons away from inventing ugly chic. Lacroix walked into that climate and presented an evening dress that required two assistants to manoeuvre through a doorway, and the audience treated it not as a provocation but as a deliverance. Some of that was Provence. He'd come from the south, his references were genuinely Boucher and Lautrec and the regional dress he'd absorbed as a child, and the clothes carried that conviction without apologising for being out of step.

The ending of the story is harder to enjoy. Black Monday hit nine days before his New York debut in October. The customers who had worn the big poufs and the big jewels at the July show, Lacroix said later, were in black-rimmed glasses and menswear by spring. The ready-to-wear that Arnault's contract had been built around never quite caught up to the couture. LVMH eventually sold the house to the Falic Group in 2005, the haute couture activity wound down later in the decade, and Lacroix today designs operas and ballets and the occasional hotel interior rather than dresses. The July 1987 show sits inside that arc like a held breath. For exactly one week the direction of Paris fashion looked like it might be about to reverse, and the editors who climbed on the chairs were not wrong about what they had just seen, only about how long it would last.

Open Apple Music's lossless support page and a stubborn little unit appears: 16-bit/44.1 kHz (CD Quality). I don't need a compact disc to hear music at that setting. The disc has become the explanation attached to the setting, an optical carrier that once needed shelf space now reduced to a label in a streaming menu.

The number did not arrive by accident. In a Philips-hosted history of the compact disc, Kees Immink describes the Philips-Sony meetings that fixed a sampling frequency of 44.1 kHz and 16-bit quantisation for the common CD standard in June 1980. The full specification lived in the Red Book and, as the article notes, in an International Electrotechnical Commission publication. A commercial object began with a negotiated row of technical decisions.

There is nothing dreamy about a figure like 44.1 kHz. That is why its afterlife interests me. We usually imagine obsolete media through their surfaces: jewel cases cracked at the hinge, the rainbow underside of a disc, the particular clatter of a portable player in a coat pocket. The sampling rate sits below all of that. It was invisible when the CD seemed modern, and it remains invisible now that the music arrives without a disc at all.

Apple's current explanation of lossless audio says that most of its Music catalogue is encoded with ALAC in resolutions ranging from 16-bit/44.1 kHz (CD Quality) up to 24-bit/192 kHz. Higher figures have become available, but the old figure still does useful work: it tells a listener what the basic lossless tier means. The stream borrows a vanished household object as its unit of reassurance.

I am more taken with this than with a retro skin drawn around a digital player. A skeuomorphic button can be redesigned next year. A standard is stickier, because manufacturers, catalogues, mastering work and listeners' expectations accumulate around it. Even a clean, modern download has to explain itself against a disc first sold when a home music library still meant furniture.

Compact discs were never as frictionless as their promise. My earlier post on pre-emphasis on early Japanese CDs stays with the physical object: a particular copy, a particular listening memory. The 44.1 kHz trace feels stranger because it escaped the object. It does not require owning an old copy of anything.

Somewhere in a settings panel, CD Quality now sits beside choices a compact disc could not contain. I can select it without hearing a tray close or watching a display count tracks. The format did not merely leave recordings behind; it left a number that the present still uses to say what ordinary fidelity should be.