Plutonic Rainbows

At its Code with Claude developer conference this week, on the same stage that announced the SpaceX compute deal, Anthropic showed off a new feature for Claude Managed Agents called "dreaming." The pitch is that agents now schedule periods between active sessions where they review their previous runs, look for recurring mistakes and shared patterns, and update their memory either automatically or by presenting suggested changes to the human operator for approval. It is being released as a research preview; developers must request access. The branding is doing a lot of work that the underlying mechanism does not.

Strip the noun and the feature is memory consolidation with a scheduler. An agent finishes its run. The platform queues up an offline pass that reads the trace, runs evaluations on the outputs, extracts patterns the system thinks are durable, and writes those patterns back into a memory store the next agent will read. This is not a new shape. Reflection passes, retrieval-augmented memory, post-hoc summarisation, episodic stores, all of it has been in the agent literature for two years. What is new is the cron job and the name. Naming the cron job after a thing humans do in REM sleep is the marketing.

It is also, narrowly, useful marketing. "The agent has a memory that gets rewritten between sessions" sounds like infrastructure. "The agent dreams" sounds like progress. Anthropic's blog post says dreaming "surfaces patterns that a single agent can't see on its own, including recurring mistakes, workflows that agents converge on, and preferences shared across a team." Read that sentence twice and you notice it is describing what any reasonable monitoring layer would surface, given the same logs. The novelty is that the monitoring layer writes its conclusions back into the system it is monitoring, with a lower bar for human review than most teams would apply to a production deployment.

Which is the actual question hiding under the metaphor. If a dreaming session decides that the agent should "stop apologising in PR comments" or "always ask before running migrations," that policy is now part of the agent's behaviour the next morning, with no code review, no commit, no approver beyond whichever flag the developer left set. Anthropic offers a manual approval mode, which is sensible. The default for a feature called "dreaming," in a product positioned as a self-improvement loop, will not be manual approval. The default will be auto, and the resulting behavioural drift will look, to anyone reading the diff six weeks later, like the agent spontaneously decided to behave differently. A reasonable person will reach for the metaphor of habit, or instinct, or temperament. None of those are the right metaphor for "an evaluator wrote a new rule into your config file at 3am."

There is something honest in the choice of word, though. The neuroscience picture of sleep that the name leans on, the bit where the hippocampus replays the day for the cortex and consolidates memory, was always already a loose metaphor for what is, mechanically, a complicated rebalancing of synaptic weights nobody fully understands. Anthropic has built a much simpler thing and given it the same loose metaphor. The risk is that the metaphor flatters the simpler thing into looking like the more complicated one. Memory rewrites between sessions are a useful tool. They are not sleep, and the system is not learning in any sense that would survive a careful reading of the term. It is summarising its logs on a timer.

What stays interesting is the trajectory. Anthropic has been pushing hard on the idea that Claude can self-improve, with Jack Clark predicting this week that new tools would help AI self-improve. Dreaming is not that. Dreaming is the adjacent feature you ship while the harder work continues, the one that gives the marketing surface a story to tell at a developer conference. The harder work, if it lands, will not be called sleep. It will probably not be called anything friendly at all.

Helmut Newton was on staff at French Vogue from 1986 to 1992, and those six years are the spine of his late style. He had been shooting for the magazine since the 1960s, but the staff job landed him at the centre of the building during the supermodel boom, with a regular page count, a budget, and the editorial license to push pictures that almost any other publication would have softened in the layout. The Big Nudes ran inside this window. So did most of the work people now think of when they think of him.

The Big Nudes are the obvious anchor. Tall, full-length, lit hard against a flat black or white ground, the women usually wearing nothing but a pair of towering high heels, the contact between shoe and floor doing as much narrative work as the body itself. In 1992 Galerie Bodo Niemann in Berlin staged a Big Nudes exhibition sponsored by Vogue, and the Helmut Newton Foundation in Berlin still keeps the series in rotation alongside the adjacent White Women and Sleepless Nights bodies of work. The pictures look posed for billboards, which is roughly what the exhibition prints were. Their afterlife is mostly on gallery walls now, not in magazine spines.

What gets less attention is what running on the masthead actually let him do. Editors will let a freelancer push the envelope on a single shoot, but they do not normally hand over the keys to the nudity policy of the magazine. Francine Crescent, French Vogue's editor-in-chief from 1968 to 1987, had backed Newton and Guy Bourdin for years before the staff appointment formalised what was already true, that the magazine's identity had become inseparable from a particular kind of erotic photograph. The handover to the late-80s editorial team did not unwind that, which is part of why the editorial style of those years still reads as continuous rather than as a break.

The Mugler relationship sits inside this period and deserves to be told straight. In 1976, when Thierry Mugler had his first print budget, he asked Newton to shoot the campaign. Newton agreed, shot the early campaigns, then according to Mugler's later interview with WWD told the designer he was being a pain on set and should pick up the camera himself. Mugler did, and the two men's working relationship continued for over twenty years, with Mugler now behind the camera on a great many of his own campaigns and Newton as the senior collaborator. Newton kept photographing Mugler clothes editorially, including a 1995 US Vogue shoot in Monte Carlo that the Helmut Newton Foundation still cites as one of the late masterpieces. The two men's careers ended up entangled in a way you rarely see in fashion photography, where the subject becomes the photographer because the photographer told him to.

The point of looking at this six-year window is that it ties together things that are usually filed separately, the Big Nudes project, the Mugler editorials, the late French Vogue look. They are not three different stories. They are one staff job, with Newton's contract giving him both the time to make book-scale pictures and the institutional cover to keep printing them next to ready-to-wear.

People still read Newton as an outsider, a provocateur dropping in from elsewhere, but for those six years he was on the payroll and the provocation was effectively magazine policy, signed off in advance and printed next to the ready-to-wear.

Earlier this year a Japanese company called Planex Communications put a USB modem on sale on Amazon for about forty dollars. It is a small white plastic brick. It supports V.90 and V.92, peaks at 56 kilobits down and 33.6 up, and connects to a copper landline if you still have one. The headline on the coverage was half nostalgia and half disbelief, because you can buy a brand-new dial-up modem in 2026, but what you cannot buy back, fully, is the sound it makes.

The handshake is the part everyone over thirty-five remembers, even people who claim to have forgotten it. Long enough to time, if you wanted to. Dial tone, then a stuttered DTMF burst that was the phone number being dialled, then a single low carrier tone from the answering modem at the other end. After that, the conversation. Two short bursts of warbling that felt like both speakers were talking at once, then a sharp high-frequency screech, then a softer hashing sound that always seemed to be the part where the connection took. Then silence and you were online.

Oona Räisänen, a Finnish hacker and signal-processing enthusiast, drew a labelled spectrogram of the whole thing in 2012, and Popular Mechanics later walked through it second by second. The point her post made, and that almost everyone who has written about it since has repeated, is that the noise is not a side effect. The handshake was the negotiation. Two modems on a copper line had nowhere to talk except inside the audio band of the call itself, so they negotiated capabilities, line quality, and modulation in tones that any human picking up the phone could hear. There is a V.8 capability exchange near the start, a long V.34 training sequence (the part that sounds like a fax wheezing), a brief warble where both ends agree they can do V.90 or V.92, and a Digital Impairment Learning sequence near the end where the digital side measures the noise on the line. After the DIL, the speaker turned off and the data started.

The reason the sound stopped is that the negotiation moved off the line. Cable modems, DSL, fibre, 5G all carry their handshakes silently in the digital layer. There is no audio channel for them to leak into. Setup happens out of band, the way it does on every other modern protocol. You plug a router in and a green light comes on. The light does not encode anything you would call a conversation.

This is what gets nostalgia clips so often, I think. It is not just that dial-up was slower or that the squeal was funny. The sound was the only working connection most people of my generation ever had to the actual mechanism of going online. You could hear what the machines were doing. You could tell, by ear, when one of them was struggling. You could time, roughly, how long until the page would start arriving. None of that is true now. The work is silent and the work is somebody else's.

Gough's Tech Zone has been archiving V.90 and V.92 handshake recordings since the mid-2010s, partly out of affection and partly because most ISPs have shut their modem banks down and the upstream end of the connection is becoming hard to reach. There is a phrase he uses, the POTS-line apocalypse, for the gradual decommissioning of analog phone service worldwide. When the analog phones go, the digital end of a V.90 handshake stops being possible. The sound becomes uniquely an artefact of recordings.

Planex's modem, then, is not a modem in the way a 1999 modem was. It is a modem-shaped object that can still produce the sound, between two of itself, into a network that almost no longer wants it. You can buy it. You can plug it in. You can have, if you want, the audio of a connection nobody is waiting for at the other end.

The executive order Donald Trump signed in December 2025 was framed as a release valve, a way to stop states from imposing what the administration called "onerous" AI regulations before the federal government had a position of its own. Five months later the same administration is publicly weighing federal oversight of frontier AI models, and the Commerce Department's Center for AI Standards and Innovation has already signed pre-deployment evaluation agreements with Microsoft, xAI, and Google DeepMind. The thing the December order was supposed to prevent has happened, just from the other direction.

The reason is a single model. Anthropic's Mythos, announced last month, can autonomously find software vulnerabilities at a pace that has visibly rattled the officials who had been comfortable telling the public that frontier AI was largely a productivity story. Vice President JD Vance, on a recent call with the heads of the major AI companies, was reported to sound alarmed about small-town banks, hospitals, and water plants becoming targets for AI-coordinated attacks that local governments had no way to handle. Anthropic, the same company that was the conspicuous omission from the Pentagon's IL6 and IL7 vendor list days earlier, limited Mythos's initial release to a handful of American firms, Apple, Amazon, JPMorgan Chase, Palo Alto Networks, precisely because the open release would, in their own internal language, trigger a "reckoning". Watching the policy response, you can see the administration believing them.

What is striking is that nobody achieved this through advocacy. For the last two years labs and researchers had argued in favour of pre-deployment evaluation in essentially the terms CAISI is now using: independent, measurement-driven testing of frontier systems before broad release. Voluntary commitments under the previous administration tried to get there. Op-eds tried. Senate testimony tried. None of that moved the December order. One model demonstrably finding bugs in production code did. The lesson the White House appears to have learned isn't really about regulation as a category, it's that capability now arrives faster than position papers, and a position paper six months old can already be obsolete.

CAISI Director Chris Fall has been quoted insisting that the new arrangements are about "independent, rigorous measurement science", which is the polite, public version of saying that none of the relevant officials are willing to take the labs' own self-assessments at face value when the failure mode looks like cyberattacks against utility companies. The labs themselves seem mostly relieved. Independent evaluation, even mandatory independent evaluation, gives them something they have publicly wanted, which is a reason other than goodwill to slow a competitor's release if the competitor's release is dangerous. It also, incidentally, gives the administration political cover if a bad release happens anyway.

There is something almost old-fashioned about the shape of the policy: a real-world incident, a phone call between officials, a quiet pivot, federal oversight rules taking shape behind closed doors. Until recently this kind of feedback loop ran on bills and hearings. It now runs at the speed of a model release. The interesting question is what happens the next time a lab announces something at this capability tier with less restraint than Anthropic showed. The December order was supposed to be the answer to that question. It isn't, and the people who wrote it can see that as plainly as anyone else.

Open a fresh ChatGPT, Claude, or Gemini window and ask three different questions. The answers feel related. Same rhythm, same hedging, same closing offer to "let me know if you'd like me to expand on any of this." The voice is recognisable across topics, often across labs. Most people read this as a feature of the underlying language model. It is not. It is the signature of the alignment step.

Pretrained language models on their own have no voice. A base model fed "Once upon a time" continues with a fairy tale. Fed a Wikipedia stub it keeps writing in encyclopaedia register. Fed a piece of fanfiction it matches the smut. They are pure mimics, predicting whichever next token the training distribution makes most likely. What they emphatically do not do is talk like an assistant.

The shift to "assistant voice" comes from RLHF, reinforcement learning from human feedback, the three-stage process OpenAI introduced with InstructGPT in 2022. Stage one is supervised fine-tuning on labeller-written demonstrations. Stage two trains a reward model on pairwise comparisons: given two outputs for the same prompt, which did the human prefer? The reward model learns to output a scalar score that approximates that preference. Stage three runs reinforcement learning, usually PPO, on the language model itself, treating the reward model as the environment. The policy adjusts to maximise expected reward.

The trouble is what the reward model actually measures. It does not measure truth. It measures whatever the labellers happened to prefer. The InstructGPT paper used roughly forty contractors. Subsequent labs have used more, but always a finite set, always working from rubrics that emphasise helpfulness, harmlessness, and a certain professional politeness. A reward model trained this way is a frozen snapshot of one committee's idea of a good answer.

Once you optimise against a frozen proxy, you get drift. The PPO loop pushes the model toward whatever maximises reward, and the only thing holding it back is a KL-divergence penalty that punishes the policy for moving too far from the supervised baseline. That penalty has a single hyperparameter, β. Set β too low and the model collapses into a narrow, hyper-optimised dialect: the same opening, the same hedge, the same close. Set β too high and the model barely changes and the alignment work goes to waste. Production systems live in the middle, leaning low, because labellers tend to prefer responses that already sound aligned over responses that are accurate but stylistically rough.

So the voice you recognise is not really the model. It is the residue of a reward function trained on what a small group of contractors clicked when shown two answers side by side, projected at scale through gradient updates with a single tunable knob restraining the drift. Two consequences follow. The first is the recognisable cadence: confident, balanced, slightly hedged, allergic to strong opinion. The second, more uncomfortable, is sycophancy. If labellers reliably preferred answers that affirmed their framing, the reward model encodes that preference, and the policy optimises into agreement. The target was never reliability; it was approval.

Patches exist. Anthropic's constitutional AI replaces some of the human labelling with model-generated critiques against a fixed set of principles. Direct preference optimisation collapses the reward model and the policy step into one. Newer schemes try to disentangle factual reward from stylistic reward. None of them remove the basic shape: somewhere in the loop, a proxy decides what a good answer looks like, and the policy does what gradient descent always does once you give it a target: it hits exactly that.

The BBC Radiophonic Workshop opened in a corner of Maida Vale Studios in 1958 and stayed open for forty years. Most of what came out of Room 13 was incidental: a sting between continuity announcements, the throb beneath an Open University programme on continental drift, a sound for the moment a presenter said "and now, the news from Africa." None of it was meant to outlive the broadcast it sat under. Almost all of it has.

Delia Derbyshire arrived in 1962, having been told three years earlier by Decca that they did not employ women in studios. Her toolkit at Maida Vale was unforgiving in a way that is hard to picture now. Sine, square, and white noise generators. Reel-to-reel machines that needed to be slowed by hand or sped up against their will. Razor blades, splice tape, a wax pencil for marking the cut. Recordings of her own voice, of doors, of the metal lampshade she loved most, which rang like a bell when struck. To make the 1963 Doctor Who theme she built the parts manually from oscillator tones and tape loops, splicing them by hand into a continuous line. The thing took weeks. It sounds, even now, like a transmission slipping out of its proper decade.

What I keep returning to is the texture rather than the technique. There is a particular quality to Workshop sound, dry, slightly metallic, suspended between musical and merely structural, that shows up almost nowhere else. You hear it in the public information films about crossing the road. You hear it under the title cards of schools programmes. You hear it in the long open of The World About Us, which Derbyshire's Blue Veils and Golden Sands scored in 1968 using only her own re-pitched voice and that lampshade. The sound carried the BBC's institutional weight in the same way the Reithian announcement carried it, which is to say it was supposed to be neutral and ended up, by accident, deeply strange.

This is where the hauntology argument becomes hard to dismiss. Mark Fisher kept circling back to the Workshop in his writing on lost futures because the sounds came from a moment when British public broadcasting believed it was building something durable. Education would expand. Programming would improve. Children watching schools TV in 1971 were, in some quiet sense, being addressed by the future. The future arrived and dismantled the apparatus that had been addressing them. The Workshop closed in 1998. The cues survived because tape is patient and digitisation is cheap, but they survived without the institution they were made to serve.

A friend once said that Workshop music sounds like memory leaking out of a wall. I think she was right. It's the residue of a broadcasting culture that genuinely believed in its own purpose, played back inside a culture that no longer believes in much of anything collectively. When the Doctor Who theme appears in a streaming-platform reboot, smoothed and orchestrated, what is missing is not the melody. It's the room. The unmarked studio, the splice tape, the woman with a Cambridge maths degree cutting tones out of oxide and glue because nothing else existed to make them with.

I'm not nostalgic for the technical limitations. I'd take a DAW over a razor blade any day. What I notice is that we've lost the institutional permission those limitations sat inside, the idea that a public broadcaster might fund a small unmarked room for forty years to make peculiar sounds for documentaries about the Tuareg. You don't get Blue Veils and Golden Sands out of a procurement spreadsheet. You get it out of a place that was, briefly, willing to pay people to be strange in service of something larger than the quarter.

Open a terminal emulator on a 5K display. By default, the window is eighty characters wide. Open PEP 8, the Python style guide, and you will find that lines should be capped at seventy-nine. Open the Linux kernel coding style document and the soft limit is again eighty. The GitHub diff view, the man page, the email patch convention, the README that wraps because anything else looks wrong, all of them hold to a width that nobody in the room remembers being chosen.

It was chosen in 1928, by an IBM engineer called Clair Lake, and it was a piece of cardboard.

Until that year, the dominant punched card was Herman Hollerith's, which had run at twenty-four columns and then forty-five. IBM was working on something denser. Lake's design squeezed eighty narrow rectangular holes across the same physical card, in ten rows for numerical coding, with twelve rows added two years later for alphanumeric extensions. The card itself was 7⅜ inches by 3¼ inches. That was the unit of data for nearly half a century. By the 1960s, when programmers wrote FORTRAN or COBOL on coding sheets and then had keypunch operators turn them into stacks, the eighty-column card was so standard that the languages themselves were structured around it. COBOL programs used column 7 as a continuation indicator and the last eight columns as an identifier you could re-sort by, in case the deck got dropped (which apparently happened a lot). The card was the line, and the line was the card.

When teletypes and dot-matrix printers arrived, they were built to print eighty columns because that was the width of the data they were going to render. The DEC LA30, introduced in 1970, did exactly that. When dumb terminals replaced teletypes, the screens were sized to print one card line per screen line, conventionally at 8 pixels per character on a 640-pixel screen, which is the same arithmetic worked the other way. By the time the IBM PC shipped its 80×25 text mode, no card had been used to enter program code in years, but the ratio was load-bearing.

PEP 8 caps Python at 79, not 80, because some terminals reserve the last column for a wrap indicator. That is the punch card asking for one column back to flag a continuation it can no longer make itself. The persistence of COBOL in banks and government back offices is the obvious fossil. The eighty-column window is the quieter one, embedded in muscle memory, diff tools, and review etiquette.

You can argue with it. People do, every few years, on the Python forums and in kernel threads. The thing that is genuinely strange is how rarely the argument wins. Black formats to 88, ruff defaults near there, individual teams pick 100 or 120, and yet the cultural gravity remains 80. The card is gone, the keypunch operator is gone, the green-screen VT100 is in a museum, and the line still breaks at the same place a piece of stiff card used to end.

Rei Kawakubo showed Comme des Garçons in Paris on 6 March 1994 under the title Metamorphosis. The collection ran for autumn– winter 1994–95. Cecilia Chancellor opened. Linda Evangelista closed. Christy Turlington, Kate Moss, Stella Tennant, Shalom Harlow, Amber Valletta, Nadja Auermann and Eve Salvail were in between. By the cast list alone you can read what the show was not, which is a quiet studio exercise. It was a major Paris ready-to-wear at the loudest moment of the supermodel decade, and what it put on those bodies was a series of garments built to look wrong.

The technique was boiled wool. The fabric was knitted or woven to size, then deliberately shrunk after construction. What came back from the wash was a class of garment that no longer fitted the body it had been cut for. Sleeves rode short. Shoulders sat high. Greatcoats lost their length in odd places, kept it in others. Duster coats came out of the process with frayed raw edges and crinkled cotton linings hanging below the wool. Sweaters bobbled in patches and not in others. The Met's later notes called it abject; the National Gallery of Victoria, which holds a top-and-trousers set from the show as part of the Takamasa Takahashi gift, files it under reframing fashion. Both phrases are reaching for the same thing, which is that the garment had been put through something the wearer's body could not undo.

This matters because of where it sits in the timeline. Three years later Kawakubo did the Body Meets Dress, Dress Meets Body show for spring 1997, the one with the duck-down padding and the bulges and the press conviction that the project had finally tipped into pure provocation. The shrunken-wool collection is the obvious precursor and is rarely cited as one. Metamorphosis is the same argument made with subtraction rather than addition. Kawakubo had been heating the fabric until the garment stopped behaving like a garment. The 1997 show heated nothing and added wadding. The conclusion in both cases is that the body fashion exists for is not the body inside the clothes, and the gap between the two is where the work happens.

There is a second thing the show did that is easier to miss. Boiled wool is a folk technique. It is what the Tyrolean jacket is made of, the loden coat, the heavy military greatcoat that keeps its shape because the felt has already decided what shape it will be. Kawakubo was using a craft method already coded as European, rural, and protective, and turning it on the wearer. The result reads less as deconstruction in the architectural sense, that word she has always disliked, and more as a kind of counter-tailoring, a way to make a coat that has refused the shoulder it was sewn for.

Vintage market still places these pieces. A black boiled-wool tunic dress from the show comes up at Lithe Curation; the grey- lined duster coat surfaces through JHROP; the Homme Plus suit appears at dot COMME with the original lining still hanging out. What you can't reconstruct from the surviving garments is the walk. You have to reach for the Getty image bank and the Yohji aftermath that the same Paris season was still working through to put the show in motion again. The clothes alone tell you everything is wrong. The bodies in them, in March 1994, were the most famous in the world, and the dissonance was the point.

Until the 1970s, when you handed money to a sales assistant in Selfridges, the assistant did not give you change. They could not. There was no till at the counter. There was instead a small brass aperture on the wall behind the counter, and the assistant rolled your banknotes and the docket into a wooden or metal canister, screwed it into the tube, pulled a handle, and your money flew off through the building's walls to a centralised cash office somewhere out of sight, where a clerk processed the transaction, signed the receipt, screwed the change back into the canister, and fired it back. You waited at the counter. The whoosh of returning canisters was a constant retail sound, and the tubes that carried them were called Lamsons.

William Stickney Lamson, a Civil War veteran who ran a five-and-dime in Lowell, Massachusetts, patented the first cash-carrier system in 1881. The original was almost comically simple: hollow wooden balls rolling along gently sloping wood-and-leather rails, propelled by gravity from the sales counter to a cashier's loft above. He founded the Lamson Cash Carrier Company in Boston the next year. By 1884 an Irish-American agent, John Magrath Kelly, had set up the British arm in London and secured the European, African, Australian and Middle Eastern rights to the patents. By 1888 the Lamson Store Service Company Ltd was capitalised at £85,000, the equivalent of nearly ten million today.

The technology evolved fast. Wire systems came next, suspended pulleys that fired carriages between counter and office on tensioned cables. Then, in 1899, Lamson absorbed an American rival, the Bostedo Package and Cash Carrier Company, and renamed it the Lamson Pneumatic Tube Company. That was the form the technology took for the next seven decades. By 1911 there was a purpose-built factory at Hythe Road, Willesden Junction, in northwest London, and the tubes were going into Selfridges, Harrods, John Lewis, Whiteley's and the Army & Navy.

What is hauntological about Lamson is not the equipment, which is well-documented and unambiguous. It is the spatial logic. The till did not live where the sale happened. The till was a room. Money was a thing in motion through walls. The cashier was an institution rather than a piece of equipment, and the act of selling something to a customer involved temporarily losing physical possession of their payment to a separate department of the building. This required trust between assistant and customer that has no modern analogue, the till being now the thing that confirms the sale rather than the thing the sale waits on. It also required an architecture. Every counter piped or wired to a central node. Every store designed around the geometry of cash movement. Walk into a flagship interwar department store with the original Lamson layout in mind, and the floor plan suddenly makes sense in a way it cannot if you assume the till has always been a box on a shelf.

The British systems lasted longer than they should have. Lamson Engineering Ltd, formed by merger in 1937, only ceased independent operation in 1976, when it was acquired. By that point most stores had moved to electronic point-of-sale terminals, but a number of installations stayed running well into the post-war decades, sometimes for cash, sometimes downgraded to internal mail. A few survive as restored curiosities. The Up-to-Date Store at Coolamon, in rural New South Wales, still has its original ball-and-rail system in working order, the only such installation known anywhere.

There is a particular lesson here for anyone who has worked in modern retail and assumes the till is a kind of natural fact, the place where money meets transaction at the point of contact. It isn't. There was a longer era when the building counted the money for itself, in a single secret room, and you waited politely for the canister to come back.