Plutonic Rainbows

The bandage dress is almost always credited to Hervé Léger, and almost everyone who repeats that has it wrong. Léger's version, the painfully tight strapless thing that became a 2000s red carpet uniform, did not arrive on a runway until the early 1990s. The silhouette he is famous for had already been on working models in Paris for half a decade by then, knitted in soft stretchy viscose by a small Tunisian man who was a foot shorter than most of his clients and who pinned dresses directly onto their bodies in the small hours of the morning. Azzedine Alaïa's spring/summer 1986 ready-to-wear was the moment the form arrived complete: a black satin-knit halter bodysuit, open at the back, married to a high-waisted bandage pencil skirt with diagonal seams and a side lace-up. ReSee's archive listing of the original S/S 1986 bandage ensemble is one of the rare places you can still see the construction up close.

What I find interesting is what the dress was answering. The mid-eighties was the high water mark of power dressing, a silhouette of padded shoulders and structured jackets borrowed from menswear and inflated, designed to claim space in rooms that had recently begun to admit women. Thierry Mugler had been building those rooms architecturally, in patent leather and chrome. Alaïa pulled in the other direction. His answer was a body that was not armoured but held; not enlarged but articulated. The Egyptian-mummy references he gave to interviewers were not idle, the wrappings were meant to evoke preservation and protection rather than constraint. The French philosopher Michel Tournier later wrote that the dresses allowed women to be held as tightly as possible while remaining free, a sentence that sounds like marketing copy and is in fact closer to the truth of how the material behaved. The Alaïa knit gave back when you breathed. Léger's bandage, when it eventually arrived, did not.

By October 1985 the trajectory was already legible to anyone paying attention. Earlier that year he had been named creator of the year at the Oscars de la Mode, France's short-lived Jack Lang-era state fashion prize, and had walked onto the stage with Grace Jones, who was wearing one of his fuchsia bandage dresses and a dazzling smile. The press had already christened the look the sexy mummy. Bill Cunningham, reviewing the spring 1986 Paris shows for Details, would write that the season had been defined by the major houses shamelessly copying the hourglass silhouette of Azzedine Alaïa. He was, by any working measure, the most imitated designer in Paris that year, and the imitation continued long after he had moved on, eventually attaching to a different name entirely.

There is a small footnote here about how authorship gets allocated in fashion. Alaïa never operated like a brand. He showed off-calendar by the late 1980s, refused the press machine entirely, hand-stitched garments directly onto Naomi Campbell and Linda Evangelista in his Marais studio, and let buyers and editors travel to him on his schedule rather than theirs. Léger by contrast was a brand from the start, marketed aggressively, quickly licensed, and by the late nineties the bandage dress belonged to him in the public mind, the way the wheel belongs to whoever sells the most cars. The historian Olivier Saillard, who curates the Fondation Azzedine Alaïa archive, has called the appropriation a strange contradiction; I think it is simply what happens when one party treats the work as the point and the other treats the work as the asset.

The original ensemble still sits in ReSee's vintage inventory at six and a half thousand euros, hand-selected for the archive by Pieter Mulier, the designer who now runs the Alaïa house. The satin knit, forty years on, is reportedly still soft.

Alexander Liberman, the editorial director of Condé Nast through most of the second half of the twentieth century, asked Peter Lindbergh a question. Lindbergh had told him, in plain terms, that he could not stand the kind of woman American Vogue kept putting in front of him. Over-styled, heavily made up, surrounded by limousines and small dogs, supported by a rich husband. Liberman, who had heard this complaint before but never from someone he wanted to hire, asked Lindbergh to show him what he meant instead.

Lindbergh picked the stylist Carlyne Cerf de Dudzeele, who liked gold Rolexes and serious jewellery and was therefore the wrong person for the job on paper. He told her not to worry about the clothes. Just bring some white shirts. He picked six models he liked at the moment: Estelle Lefébure, Karen Alexander, Rachel Williams, Linda Evangelista, Tatjana Patitz, and Christy Turlington. None of them was famous yet. He flew the whole group to Los Angeles and put them on the beach at Santa Monica.

What he produced was almost nothing. Six women in oversized men's shirts, mostly unbuttoned, almost no makeup, hair left to do what the wind did. Bare feet. Underwear, where you could see it. Black and white. They were laughing in some frames and looking past the camera in others. There were no clothes to sell because the only clothes were borrowed shirts.

He brought the prints back to New York and laid them out on a table in front of Liberman and the then-editor Grace Mirabella. In his telling years later, the two of them looked at each other and said little. The pictures were not what anyone had asked for, because the brief had been so vague it could not have been. They ran in the August 1988 issue of American Vogue. Within eighteen months, Lindbergh was working with the same group in different combinations, sometimes adding Naomi Campbell, sometimes Cindy Crawford, until in January 1990 he shot the British Vogue cover that everybody now reaches for when they want to date the start of the supermodel era. The Versace finale in Milan eighteen months after that was the same idea staged with music.

The Santa Monica picture is the one that gets less credit because it ran before anyone knew what they were looking at. There is no front-page narrative attached to it, no music video, no runway moment. It is a magazine spread that happened to contain six women whose names did not yet matter, in a setting that contained nothing else, photographed by someone who had spent his career arguing that the clothes were a pretext and the woman was the point. The thesis was the picture; the picture was the thesis. Everything that followed in the next six years, the big-haired Versace catwalks, the George Michael Freedom! '90 video, the perfume contracts, the trademarked first names, started here and worked outward from a shoreline.

What is strange in hindsight is how thin the apparatus around it was. No production designer, no celebrity, no set, almost no clothes. A photographer who had argued his way into a magazine he didn't like and a stylist who agreed, against her own instinct, to leave the jewellery in the case. The picture they made together did not announce itself. The supermodel decade walked out of it slowly, like the tide going back.

Anthropic and the Gates Foundation announced a $200 million, four-year partnership earlier today, aimed at global health, life sciences, education and economic mobility. The framing is familiar: AI for everyone, not just for the markets that can already pay for it. The structure of the commitment is the part worth slowing down for.

Half of the $200 million is the Gates Foundation putting in cash, program design and the operational expertise it has built up running large bilateral health programmes. The other half, Anthropic's half, is Claude usage credits and time from its technical staff. That distinction matters because the two halves do not have the same cost structure. Grants are dollars out. Credits are inventory, the marginal cost of running tokens through Anthropic's existing compute footprint, and they hit Anthropic's books somewhere between full price and zero depending on how the accounting is done.

I do not mean that as a debunking. I mean it as a clarification of what is on offer. Anthropic is committing a substantial fraction of its model capacity, for four years, to programmes whose research agenda will not be steered by paying customers. That is a real concession. It is just not the same shape as a hundred million dollars handed across a table.

The precedent is OpenAI's $50 million deal with the same foundation in January, aimed at supporting roughly a thousand clinics across sub-Saharan Africa by 2028. Four months later, the asking price of the equivalent kind of announcement appears to be four times higher. Whether that is genuine inflation in scope, or two labs leapfrogging each other on a public-relations metric where the dollar number is the only thing that travels in a headline, is impossible to know from the press release. The Gates Foundation tends to know what it is buying. The labs, by contrast, are in the middle of a brand argument with the Trump administration about whether AI is a force for harm or a force for good. A four-year commitment announced now is a brand argument with a planning horizon attached.

The substantive programmes themselves read sensibly. Accelerating vaccine and therapy research is a workload large language models are genuinely useful for, particularly the literature-synthesis and protocol-drafting parts. Establishing benchmarks for healthcare tasks is a known gap. The US programmes around portable skill credentials and employment outcome measurement are less novel as ideas, more about whether the data infrastructure can actually be built. I find the global health side more interesting because the marginal use of one more model call in Seattle is small, and the marginal use of one more model call in clinics in low-income regions is potentially enormous.

What I will be watching for is the public reporting. The partnerships that outlast their press releases are the ones that leave reusable artifacts behind, datasets, evaluation suites, methods papers, working tools that subsequent projects inherit. If those land, the four-year window will have produced something durable. If they do not, the announcement will read in 2030 the way most billion-dollar philanthropic launches read in retrospect, as a number that sat well in a press release and then drifted out of view while the actual work, whatever it was, went on quietly elsewhere.

The UK's AI Security Institute published a short blog post yesterday afternoon that I think is the most consequential thing anyone has written about frontier models this month. It is not an announcement. It is a number, and a story about why the number keeps moving.

In November 2025, AISI estimated that the length of cyber tasks frontier AI models could autonomously complete was doubling every eight months. That was already, by anyone's standard, fast. Three months later, in February, the institute reran the analysis with fresher data and revised the doubling time to 4.7 months. The acceleration was itself accelerating. Yesterday's update is that even 4.7 months now looks too generous. Claude Mythos Preview and OpenAI's GPT-5.5 have both substantially exceeded the trend the institute was tracking, and AISI is openly unsure whether what it is watching is an isolated jump or the start of a new, faster trajectory.

The concrete data sits underneath the abstraction. AISI runs a cyber range called The Last Ones, a 32-step simulated corporate network attack that, until this round of testing, no model had made meaningful progress on. Claude Mythos Preview solved it on six of ten attempts. GPT-5.5 solved it on three of ten. Mythos also returned the first non-zero score on a second range called Cooling Tower, succeeding three times in ten. Two evaluations that were near-zero a few months ago are now in the high single digits, and the institute's own framing of the trend has shifted from "fast" to "we are not sure what we are looking at any more."

What I keep coming back to is the shape of the announcement. AISI was set up to be a calm, slow, scientific counterweight to the marketing claims of the labs. It runs evaluations, it publishes priors, it revises them when the data demands. Twice in six months it has had to revise its own central estimate sharply downward. That is not a failure on AISI's part. It is the institute working as designed. It is also, however, a signal that the institution best positioned to tell a government when to slow down is now telling that government, repeatedly, that the thing it is measuring is faster than the last time it was measured.

Palo Alto Networks, working under Anthropic's Project Glasswing partnership, reached similar conclusions through its own testing, which CyberScoop covered yesterday. The two reports landed on the same day, from different methodologies, with the same finding. Frontier models are now extraordinarily capable at finding vulnerabilities and turning them into exploit paths in close to real time. The wording varies. The line on the graph does not.

What I notice is the asymmetry of who can act on the number. AISI can publish. The labs can choose to deploy or hold. The British government can negotiate access. None of them, working alone, can slow the curve. The curve is the product of a hundred decisions being made in parallel by people who do not report to AISI and who do not particularly care whether the doubling time is eight months or 4.7 months or, the next time the institute reruns the analysis, something shorter still. The useful question is not what the doubling rate is. It is what anyone is going to do once the rate stops being information and becomes the thing the institute is reporting on for its own sake.

Butlin's closed its Filey and Clacton camps in 1983, and that is the year the British holiday camp stopped being a national default and became a thing that people remembered. Billy Butlin had opened the first camp at Skegness in 1936 with the slogan Our true intent is all for your delight. At its peak Pontins ran around thirty camps, Butlin's ran the largest sites, Warner's operated a second-tier circuit, and Ladbrokes had bought up older sites including Caister-on-Sea, which had been operating since 1906 and is sometimes credited as the first camp of all. The whole apparatus assumed that an industrial family with a fortnight off in August would drive to a fenced site on the east or north-west coast and stay inside the fence for the whole fortnight, eating in the dining hall, queueing for the heated outdoor pool, listening to the Redcoats announce the children's talent contest on the tannoy at half past two.

What killed it was Spain. The package-holiday industry had been gestating since the 1950s, but the things that made it mass-market arrived together in the late 1960s and early 1970s: bigger and faster charter aircraft, the easing of the post-war currency-export limits, Franco's industrial-scale construction of resort hotels along the Costa Brava and Costa del Sol, and the price war that Intasun fought with Thomson through the 1980s. By 1983 a fortnight in a half-finished Benidorm tower with full board cost the same as the equivalent fortnight in a Filey chalet with full board, except in Benidorm the sun showed up. Butlin's closing two of its largest sites in the same year was an admission. The camp model required the British weather to be acceptable. It was no longer competitive against a model that didn't.

What the camps left behind, in the places they didn't redevelop for housing, is a very specific kind of ruin. A holiday camp is not a building, it is a system of buildings: a row of identical chalets in three colour-coded lines, a central pavilion with a ballroom and a stage, an outdoor pool with a diving tower, a boating lake, a row of arcade buildings, and a perimeter fence with a single gatehouse. Most of those components survive demolition unevenly. The chalets go first because they are cheap to flatten. The ballrooms last longest because they are unusually large clear-span structures with concrete floors and nothing easy to convert them into. You can still walk past a Pontins ballroom on the north Wales coast and recognise it immediately from the pitched copper roof and the absence of windows along the lower walls. Inside, the parquet is still down. The stage curtain has been pulled across and left. The spot where the resident band's drum kit lived has a slightly darker patch where the carpet didn't fade.

The surviving operators rebranded out of the word camp in the 1990s because the word had become a liability. Holiday village, holiday park, holiday centre. Center Parcs opened in the UK in 1987 and proved that British families would still pay to be enclosed somewhere with their entertainment provided, as long as the enclosure was made of pine trees and didn't have a row of identical peach-coloured chalets visible from the access road. Butlin's narrowed to Skegness, Bognor Regis and Minehead, rebuilt all three around indoor water complexes, and stopped using the chalet form altogether. Pontins was sold to Britannia Hotels in 2011 and continues a thin existence at a handful of sites that look, in the photographs, like they have been preserved on purpose at the exact moment they stopped making money.

The thing the holiday camp had, and that no current British holiday product has, was the assumption that a stranger through a megaphone could organise your week. The Redcoat and the Bluecoat were not service staff. They were closer to parish priests in shorts, structuring the day with sports days, knobbly-knees contests, glamorous-grandmother heats, afternoon bingo, and the children's club so the parents could have an hour. That structure is what the surviving ruins remember. The pool is empty, the stage is empty, the tannoy brackets are still on the wall, and the calendar of forced collective fun has dispersed into seven hundred individual streaming subscriptions watched alone in seven hundred caravans on the same coast.

EGLLZAZX is the kind of address every flight plan filed into a London control room still travels through, eight characters, fixed width, an ICAO location indicator stitched to a three-letter facility code. The format was designed in the 1950s for landline teleprinters that punched their output onto paper tape and clacked it out again at the other end. Every flight plan, every NOTAM, every METAR update, every distress and urgency message between airfields, the same eight characters, the same circuit grammar. Aircraft navigate by GPS. Their paperwork still arrives by telex.

The Aeronautical Fixed Telecommunication Network has been the ground-to-ground backbone of global aviation since the early 1950s, and the bones have not been replaced. Messages stay text-only, capped at roughly 1800 characters, ASCII with the radioteletype shift codes the original infrastructure required. The Wikipedia entry notes, drily, that "the message format betrays the extensive use of radioteletype links in the past." It still does. The end-of-message sequence is NNNN, followed by twelve letter-shift signals that exist only because some terminal somewhere might still be reading them.

You can read the inheritance directly off the facility codes that sit in the second half of the address. YNYX is a NOTAM office. YMYX is a local met office. ZTZX is a control tower. ZQZX is an area control centre. Indian Meteorology Department training material from 2020 lists these as the live address grammar of the international weather feed, and that document is not a history paper, it is operational. The address scheme has outlived the machines that justified it, and every aviation authority in the world is still typing into it.

The official line is that AFTN is "obsolescent". The Aeronautical Message Handling System, based on X.400 over IP, is supposed to replace it. AMHS handles binary attachments, digital signatures, XML weather messages (IWXXM, the new format, contains characters AFTN cannot physically transmit). ICAO encourages all member states to complete the transition as part of the Global Air Navigation Plan. The reality is gateway converters. Most regions run both networks in parallel and let an AFTN/AMHS gateway translate between them, because the cost and certification burden of cutting AFTN entirely is more than any single authority wants to carry. The old network keeps running because the new one cannot quite finish replacing it, and because the old one, for the messages it was designed to carry, still works.

The haunting is structural, not nostalgic. A character limit set by 1950s tape readers now constrains the format of weather products designed for satellite-era forecasting. An eight-letter address scheme now governs how a Boeing 787, a machine that knows its own position to within a metre, announces its intentions to the world. The grammar of a punched-tape machine sits between every modern aircraft and the controllers who route it. Nobody chose that arrangement in 2026. It just survived everything that was supposed to replace it.

The same logic keeps the fax machine humming in radiology and oncology, and the same logic keeps POCSAG running through hospital walls that block every protocol designed later. There is a pattern. The oldest layer of the stack is usually the one nobody can afford to remove, because every replacement carries failure modes the original did not have. AFTN was built for messages that absolutely must arrive, in a format that absolutely must be parsed by the next station along the chain, with no negotiation. Modernity keeps trying to negotiate. The 1950s spec keeps arriving.

Some cities present themselves neatly. Others only reveal themselves from the side, in the service road, under the bridge, behind the civic building where nobody was meant to linger. Bristol, perhaps more than most, belongs to the second category. Its public face is easy enough to identify: Clifton, the bridge, the harbour, Georgian terraces, student cafes, hills, postcard drops of light over the Avon Gorge. But there is another Bristol too, harder, stranger, more provisional. A Bristol of underpasses, concrete walkways, civic backsides, rain-stained walls and fluorescent windows. These two photographs seem to hold both versions of the city in suspension.

The first image, taken behind the old police station in 1984, is almost aggressively unpicturesque. It is not trying to charm anyone. There is no harbour glimmer, no terrace elegance, no "historic Bristol" being offered up for approval. Instead, we are in one of those back-of-city spaces that British urban planning produced in abundance after the war: a narrow road hemmed in by brick, concrete, office windows, walkways and shop units. The city not as spectacle, but as infrastructure.

Overhead, enclosed bridges cross the street like institutional arteries. They do not invite you to look up so much as remind you that the real business of the place is happening elsewhere, behind glass, above street level, in corridors and offices. The street below feels like a leftover channel, somewhere you pass through rather than arrive in. Even the shops have the air of temporary tenants in someone else's plan.

And yet the photograph is compelling precisely because of this ordinariness. The mundane details now become the emotional charge: the bin in the road, the dull double yellow lines, the small shop signs, the pedestrians walking away, the faint glow of interior lights in an office block. Nobody in the scene is posing for history. Nobody seems aware that this ordinary Bristol afternoon will one day become almost unrecoverable.

That is the peculiar ache of such images. The past is rarely arranged like a memory while we are living through it. It is mostly made of errands, side streets, dead time, unremarkable buildings and moments spent going somewhere else. Later, these overlooked places become portals. The unattractive becomes precious because it was not preserved for us. It survived only by accident.

The old Bridewell site has a longer and messier history than the photograph needs to explain. The Island's own account says the Central Police Station complex was built in 1928 and opened as a police station in November 1930, on Nelson Street, near the site of an earlier station. By the time this picture was taken, the building was already moving toward its afterlife. Avon and Somerset Constabulary had been created in 1974, headquarters functions started moving out the following year, and the remaining CID offices left in 1986 when the New Bridewell Police Station was completed across the road.

So the photograph catches the place in a particular middle condition. Not historic yet. Not dead. Not quite modern anymore.

The image has the mood of municipal modernism after the optimism has drained out of it. Those concrete walkways once belonged to a future: efficient, elevated, planned, rational. By 1984, they already look tired. The rain has found them. The concrete has darkened. The dream of circulation and civic order has become a damp canyon behind a police station.

It is easy to mock this kind of architecture, and often it deserves it. But the photograph does something subtler. It shows the melancholy dignity of a failed future. These buildings were not ancient, not yet ruins, not still new. They occupied that strange middle age of the urban landscape, when yesterday's progress has become today's background.

If that first photograph is Bristol as back corridor, the second is Bristol as apparition. The Clifton Suspension Bridge, seen from the bottom of the Avon Gorge in February 1992, is familiar and unfamiliar at once. The bridge itself is instantly recognisable, but the angle changes everything. We are not up at Clifton Observatory, not on the elegant side of the city, not looking at a tourist view. We are down below, at river level, where the gorge is steep and shadowed, where the road runs tight against the cliff and the Avon withdraws into mud at low tide.

From here the bridge appears almost impossible: a thin white line drawn across blue air, suspended between two enormous masses of stone. It is more idea than object. The road, cars, tunnel mouth, lamp post and warning sign pull the scene back into the early 1990s, but the bridge itself seems to exist outside the photograph's date. It belongs to another scale of time.

That contrast is what makes the image work. The gorge is geological time. The bridge is Victorian time. The road is twentieth-century time. The cars are 1992. The photograph is now memory. All these layers sit together in one frame, and none of them fully cancels the others.

The light helps. February light can be cruelly beautiful: pale, cold, clarifying. Here it gives the scene a kind of blue distance. The trees are bare, the river is low, and the gorge recedes into haze. It does not feel cosy or picturesque. It feels like a threshold, a place of departure, arrival, passing through. Bristol as a gap between rock and sky.

The bridge is an obvious icon, but the official history is stranger than the iconography. The Clifton Suspension Bridge Trust describes it as a bridge completed after Brunel's death by John Hawkshaw and William Henry Barlow, opened on 8 December 1864, and still maintained through the toll system. That helps explain why the photograph feels layered rather than merely scenic. The bridge is a Victorian engineering object that still carries daily traffic, a tourist image that remains a working piece of infrastructure. It is not just looked at. It is used.

The two images seem, at first, to have little in common. One is cramped, grey, urban, almost claustrophobic. The other is open, blue, monumental. One shows the city's discarded modernity; the other shows its enduring icon. One belongs to the service entrance, the other to the civic myth. But they share something essential: both are images of things that continue without us.

This may be why photographs of places can feel more unsettling than photographs of people. A face declares its vulnerability. A building, a bridge, a road, a cliff, these simply remain. They do not mourn the versions of us who passed through them. The city carries on, altering itself, erasing itself, repainting itself, demolishing one decade and marketing another.

The first photograph may show a Bristol that has largely vanished: shops renamed, buildings altered or demolished, walkways removed, traffic systems changed. The second shows a Bristol that still appears to exist, because the bridge remains. Even there the continuity is deceptive. The exact February air is gone. The particular cars are gone. The quietness of that road, the colour of that film stock, the early 1990s atmosphere, all gone. The permanent landmark only makes the vanished details more visible.

We think we want the monument, but it is often the incidental that hurts. A lamp post. A shop sign. A road marking. The shape of a bin. The colour of fluorescent light in an office window. The precise shade of a winter sky in 1992. The old city does not return as a grand historical narrative. It returns as texture.

There is also a faintly hauntological quality to both images. Not because they show ghosts, but because they show futures that failed to remain future. The 1984 street is haunted by the post-war promise of planned urban life. The 1992 gorge is haunted by the Victorian sublime, by the idea that engineering could draw a perfect line across a natural void. In both cases, the image contains a vanished confidence.

And yet neither photograph is simply nostalgic. Nostalgia often tidies the past. These images do not. The concrete street is bleak. The gorge road is not romantic in any soft sense. The river mud is exposed, the shadows are hard, the urban fabric is compromised. What makes them moving is not that the past was better. It is that the past was real, physically, stubbornly real, and yet is now unreachable.

You could go to Bristol now and stand near these places. You could find the bridge, walk the gorge, search for the old police station area, compare street views, identify what changed and what survived. But you could not step into either photograph. You could not recover the smell of the street behind the police station in 1984, the sound of those cars, the exact rhythm of that weekday. You could not re-enter February 1992, with that blue air and those shadows and that particular version of the Avon below the bridge. The laws of physics, annoyingly, remain firm on this point, and time refuses planning permission.

What remains, then, is the photograph: not a doorway, but evidence. Evidence that these arrangements of matter and light once existed. Evidence that the city had these moods. Evidence that people passed through them, mostly unaware that they were moving through what would later become history.

Together, the two images form a small portrait of Bristol's split personality. Bristol the concrete back passage; Bristol the sublime gorge. Bristol the failed precinct; Bristol the impossible bridge. Bristol as damp municipal afterthought; Bristol as suspended dream. The real city is the tension between them, the place where a stained concrete walkway and a world-famous suspension bridge can belong to the same emotional geography.

Cities store time unevenly. Some things endure. Some things are erased. Some survive but no longer mean what they meant. And some, by being photographed almost casually, become more vivid after they have disappeared than they ever were when they stood before us.

He almost quit before the show. Gucci was hemorrhaging money, the previous collection had landed without a review anyone could remember, and Tom Ford later said there was a moment when nobody was looking at anything he did. He could have sent anything down the runway. He sent the Fall/Winter 1995 ready-to-wear instead, in Milan, and the line between "almost quit" and "biggest designer of the decade" turned out to be that one show.

The clothes were not complicated. Jewel-tone satin shirts in emerald, sapphire, and aubergine, unbuttoned down past where a shirt is normally unbuttoned. Velvet hip-huggers cut so low they read as a sneer at the high-waisted minimalism the rest of Milan was selling that week. Horsebit loafers with a patent finish so glossy the photographers complained about the bounce. A handful of sharp wool suits. That was most of it. The collection was short, the silhouette was singular, and the styling was about seventy-five per cent of the message.

Ford killed the back-light. The Milan runway convention in 1995 was a long thrust stage with the front row facing each other across it, lit so that the buyers and editors could be seen as much as the clothes. Ford had the back lights dropped and ran the show with a hard spot down the centre, which meant the audience couldn't see each other and had no choice but to watch the clothes. Kate Moss walked. Shalom Harlow walked. Amber Valletta walked. The room got quieter than Milan rooms usually get.

The thing nobody could quite explain at the time was the register. Late minimalism was the dominant note that season, Helmut Lang's slip dresses, Prada's reduced palette, the careful restraint that the late nineties would later be remembered for. Ford's collection was the opposite of all of it, sexual in a way that was almost confrontational, glossy, expensive-looking, with a seventies disco floor underneath it. Amy Spindler wrote in the New York Times that he had brought cool back to luxury at a moment when street fashion had been the only thing that felt cool for years. That was the polite version. The less polite version was that the show looked like sex, and the buyers reacted to it accordingly.

Madonna wore the key look to the MTV Video Music Awards that September, a satin shirt unbuttoned to the navel and the velvet trousers, and a collection that had been hanging in Milanese showrooms became, in a single night, the thing half-famous people in Los Angeles wanted to be photographed in. By the following autumn, Calvin Klein, Versace, and Chanel had all shown their own hip-huggers. Donna Karan put a note on her skirt tags explaining where they were meant to sit on the body, because customers were complaining about waistbands that no longer hit at the waist.

The Chanel show that same Milan-and-Paris fashion month, at the Cirque d'Hiver in March 1995, was Karl Lagerfeld doing his own version of the supermodel spectacle, and the two collections shared a casting roster and a sensibility about what a runway could look like in the mid-nineties. Both shows leaned on the same handful of women. Both treated the audience as a secondary consideration. The difference was that Lagerfeld was already at the top of the mountain and Ford was hauling Gucci up it from base camp.

What Ford did at Gucci over the next nine years has been written about so many times the receipts blur, the 1,200 per cent sales growth, the Pinault acquisition, the eventual YSL appointment, the fall-out in 2004. The thing that gets written about less is how narrow the moment was. The same designer six months earlier had been weeks from firing. The same house a year before had been weeks from administration. The collection that turned everything was not a relaunch built with a year of preparation and a hundred million in marketing. It was one short show in Milan, with a spotlight off the audience and a row of jewel-toned satin moving through it.

On Monday, Mira Murati's Thinking Machines Lab previewed a product it calls an interaction model. The pitch, summarised by the Business Insider piece that ran today, is a system that handles speech the way two people actually handle it: listening and talking at the same time, taking interruptions in stride, translating between languages on the fly. The wider launch is promised for later this year. What was shown on Monday is a preview.

The architectural claim is the interesting bit, not the demo reel. Every voice assistant most people have used, including the current crop of LLM front-ends, is fundamentally turn-based. You speak, it waits, it transcribes, it thinks, it responds. Even when the latency drops to under a second the structure is still strictly sequential. An interaction model, by contrast, runs the listening loop and the speaking loop in parallel, so the foreground stays responsive while the heavier reasoning happens underneath. The reported latency is around 0.4 seconds with internal micro-turns of about 200 milliseconds, which is roughly the point at which human conversation stops feeling like a walkie-talkie call.

Whether this generalises beyond a demo is a separate question. Full-duplex audio is not a new idea in research, it's been sitting around in conversational systems work for years, and shipping it as a product is mostly an engineering exercise in keeping a generative model coherent while it is being talked over. The hard part, historically, has been preventing the model from collapsing into either a babbling overlap or a panicked silence the moment the input pattern departs from the training distribution. Real interruptions are messy. People trail off, backtrack, change their minds mid-clause. You can build a duplex system that handles a scripted interview beautifully and falls apart on a phone call to a plumber.

The other half of the story, in the same Business Insider report, is that Thinking Machines has now lost roughly a third of its founding team to OpenAI, Meta and xAI, with the one-year cliff vesting being one of the levers. I wrote about an earlier wave of those departures in January. The pattern has not changed since. A lab founded by a charismatic ex-incumbent raises a fortune at a valuation that prices in superintelligence, and the talent it gathered to justify that valuation gets bought back, individually, by the labs with even more compute and even larger compensation envelopes. The product gets shipped or it doesn't. The founders mostly end up somewhere else.

So I'm reading the interaction-model preview with two different kinds of attention. As a technical demonstration, it's a genuinely fresh framing of what an LLM-fronted voice product can be, and it points at a future where the dominant mode is duplex rather than turn-based. As a corporate signal, it is the kind of thing a lab puts out when it needs to show the market that the engineering core still functions even while the org chart is being rewritten in real time. Both readings can be true. They usually are.

The thing I want to know is whether the underlying model is small enough to run on a phone, or whether the 0.4-second latency depends on a hyperscaler-grade GPU sitting on the other end of a private fibre. The press release does not say, which is itself a signal. Watch for the API price when the wider launch happens. That will tell you which of those two worlds we're in.