designerlistings.org Miniblog

Runway Patina: Why Fashion Keeps Borrowing Old Rooms

In 2026, fashion’s newest flex may be old furniture: not as décor, but as borrowed history.

Jonathan Anderson opened Dior’s fall/winter 2026 womenswear show in March at the Jardin des Tuileries with miniature green Sénat chairs as invitations, turning the 1923 public-garden seat into the collection’s emblem. In January, Pharrell Williams and Not a Hotel built Drophaus, a midcentury-leaning domestic set, for Louis Vuitton Men’s fall/winter 2026 runway. In February, Ralph Lauren staged fall 2026 among vintage rugs and antique seating inside New York’s beaux arts Clock Tower building. In December 2025, Matthieu Blazy unveiled his first Métiers d’Art collection for Chanel beside a vintage train in an abandoned downtown Manhattan subway station.

The move reflects more than nostalgia. In a culture flattened by screens, archived icons offer what clothing alone can’t instantly generate: place, memory, rigor, and the feeling that a world existed before the look arrived. Jonathan Olivares argues that design now gives fashion architecture around the body, while long-developed 20th-century furniture carries stability and conviction that fast seasonal churn struggles to project.

Chris Black notes that brands borrow prestige from already-canonical places and objects; Dior did not create the Sénat chair’s meaning. Adam Charlap Hyman sees vintage interiors as evidence of life, not just styling. Vejas Kruszewski points to antique looms preserved for their irregular texture. Flynn McGarry, lately buying vintage pine, values the unrepeatable tone of aged wood and the afterlife of rediscovered things.

Posted on 1 July 2026

Ikea PS 2026 Turns Practical Furniture Into Play

Ikea has unveiled Ikea PS 2026, a 44-piece collection built around what it calls playful functionality: furniture and objects that do useful things, then wink at you a bit.

The PS line began in the 1990s as Ikea’s experimental workshop for domestic design. The idea was rather bold and very Swedish: if Scandinavian design was being copied and sold at absurd prices, why not make the real spirit of it available to people without heroic wallets? Over three decades, that ambition has moved through themes including The Democratic Home, Inspired by History, Designed for Today, and now this latest chapter.

At the centre of PS 2026 is a simple proposition: plain design needn’t be dull. Maria O’Brian, Ikea’s creative leader, frames the collection as an exercise in discovery, where pared-back forms hide extra functions and small surprises.

That mood runs through the range. There’s a rocking bench, an inflatable armchair, wall masks, and adjustable pieces intended to invite use rather than reverence. It’s furniture you’re meant to prod, shift, sit in, and generally get on with.

Designer Mikael Axelsson’s inflatable chair captures the point neatly. Instead of treating design as a sacred relic, PS 2026 pushes it toward curiosity, interaction, and everyday delight.

Posted on 27 June 2026

Architecture’s New Overeager Intern Is Already in the Design Meeting

At Gensler, AI moved from curiosity to infrastructure fast. Three years ago, the firm opened an internal AI sandbox to test generative tools before the architecture world fully settled on whether to trust them. Now AI touches most of the roughly 3,000 projects Gensler handles each year.

The payoff is not robotic authorship so much as amplified imagination. Designers use AI to study daylight, acoustics, circulation, occupancy, ventilation, and energy use before a building exists in concrete, steel, or expensive regret. On Under Armour’s new headquarters in Baltimore, Gensler used AI to model how employees might move through the building over a day and how that behavior would affect environmental performance.

The firm also adopted RunDiffusion about six months ago; within four days, more than 3,000 staff were trained. The result has been a burst of rapid concepting and a new strength in presentation: AI-generated visuals turned into cinematic narratives helped Gensler pitch a mixed-use district near a future stadium and imagine Baghdad Sustainable Forests, the largest ecological development in Iraq.

Elsewhere, caution remains sturdy. A 2025 American Institute of Architects study found about 90% of architects surveyed worried about inaccuracies, security, and separating AI-made from human-made work. Pratt’s Jason Vigneri-Beane and UNC Charlotte’s Sabri Gokmen see AI as especially useful early on, with Gokmen also developing a model that can generate 3D tower forms from prompts or images.

Posted on 26 June 2026

Architect Labs Wants to Make Custom Chips Less Like Alchemy

Custom chip design has long been the technological equivalent of trying to assemble a dragon from IKEA instructions: expensive, slow, and reserved for a tiny priesthood of specialists. Palo Alto-based Architect Labs wants to smash that bottleneck.

The startup has emerged from stealth with a $24 million seed round led by Kindred Ventures, with backing from TQ Ventures, Race Capital, Together Fund, and investors including Srinivas Narayanan, Lukasz Kaiser, Aravind Srinivas, Kunle Olukotun, Trevor Blackwell, Dr. Alex Wissner-Gross, Shaad Khan, plus executives from NVIDIA, Google, and OpenAI. Kindred founder Steve Jang has joined the board.

Founded by Stanford dropouts Ebrahim Hussain and Aaditya Subedi, Architect Labs is building an AI system that designs and verifies chips from end to end, then extends into compilers, runtimes, system software, and eventually AI-model co-optimization. The aim is to let companies, labs, and even nations build custom silicon without becoming chip companies themselves.

That matters because AI infrastructure is shifting beyond generic GPU-CPU-memory setups toward specialized hardware for datacenters, robotics, autonomous systems, spatial computing, defense, personal devices, and wearables.

Hussain previously worked on custom chips at Apple and Tesla after entering college at 15. Subedi researched AI-based code verification at Harvard. Their team includes veterans from Anthropic, DeepMind, xAI, Meta, Google, Intel, and others, with 80-plus tape-outs and leadership across major silicon and AI programs. The new funding will support compute, research, and early production partnerships.

Posted on 20 June 2026

Adidas’s 2026 World Cup Ball Tries to Be Less Chaos, More Physics

Soccer finally did what your ex never could: committed to consistency.

Wind-tunnel testing suggests the 2026 FIFA World Cup ball, Adidas’s Trionda, should fly more reliably than some of its chaotic ancestors. Physics researcher John Eric Goff and collaborators in Japan found that the new four-panel ball keeps a steadier drag coefficient down to about 12 meters per second, avoiding the ugly mid-flight aerodynamic surprise that made the 2010 Jabulani infamous with goalkeepers.

That earlier ball hit a drag crisis around 20 m/sec, meaning its resistance changed abruptly as it slowed. For a keeper, that’s less sport and more jump-scare.

The Trionda uses the fewest panels ever in a World Cup ball, but Adidas appears to have dodged the usual smooth-ball curse by roughening the surface. Each panel has three deep grooves, plus a winding seam pattern and thermal bonding. Tests showed the ball is slightly rougher than recent predecessors, which should stabilize flight. The tradeoff: a bit more drag, likely trimming long kicks by a few meters, while also helping players generate more spin.

Spin still complicates everything; the tests used non-rotating balls, so match play may reveal extra Magnus-effect drama.

The ball also carries a 500Hz IMU sensor embedded in a panel, with balancing weights elsewhere. Working with Kinexon Sports, Adidas uses that live data for VAR on offsides and touch detection.

Posted on 16 June 2026

The Aesthetics of Not Offending Anyone

There is a peculiar hush settling over designed things, as if whole industries have agreed that the highest aesthetic good is not beauty or wit or even memorability, but the avoidance of complaint. So the built world turns into a repetition of glass slabs; brands sand themselves down into interchangeable sans-serif names; cars get stream-smoothed into near-identical capsules; phones, speakers, and laptops arrive as the same softened rectangle in approved tones of black, white, or grey.

This gets filed under minimalism, which gives it an air of discipline. But much of what passes for refinement is really fear with excellent lighting.

Simplicity is not the culprit. Plenty of great design has been simple. The issue is that simple has hardened into a template. Instead of treating design as a chance to ask what a thing might become, companies increasingly use it to ask what has already proved safe.

The logic is financial before it is aesthetic. Investors dislike buildings that might unsettle. Buyers can punish strange-looking cars. A logo that departs too far from the familiar risks instant online recoil. The rational corporate response is a product that has been pre-blessed by precedent: recognizable, frictionless, forgettable.

Then the algorithms arrive to complete the enclosure. They amplify what is already popular, and popularity in design usually means conventionality. So sameness becomes visibility, visibility becomes imitation, and imitation becomes the look of the age.

Posted on 9 June 2026

World Cup 2026 Branding Scores an Own Goal

The FIFA World Cup 2026 logo is a remarkable achievement in modern design, if the design brief was "open Photoshop, drag a PNG into the middle, and call it a day."

Look at it. Really look at it. This is supposedly the visual identity for the largest sporting event on Earth. Billions of viewers. Three host nations. Decades of anticipation. And the final result is what appears to be the World Cup trophy floating awkwardly in front of a giant "26" like it accidentally wandered into the frame during a PowerPoint presentation.

The central problem is impossible to ignore: the trophy doesn't feel integrated into the design at all. It genuinely looks like someone searched "World Cup Trophy transparent PNG," downloaded the first result, and plonked it in the middle. To be fair, they did at least cut it out quite neatly. The edges are clean. That's something. Somewhere, a graphic designer spent a solid three minutes making sure the selection tool didn't leave any jagged pixels behind.

The giant black "26" isn't helping. It has all the visual charm of a parking garage wayfinding sign. The composition feels less like a prestigious tournament identity and more like a placeholder graphic waiting for the real logo to arrive. "Don't worry guys, we'll replace this before launch." Then nobody did.

What's particularly baffling is that previous World Cup logos, whether you liked them or not, at least attempted to express something about the host nation, culture, or tournament. This one expresses the number twenty-six. Admirably, relentlessly, aggressively twenty-six.

The logo's defenders will say it's minimalist. That's a generous interpretation. Minimalism usually involves removing unnecessary elements. Here, the design consists almost entirely of one necessary element pasted onto another necessary element.

It's not offensive. It's not ugly. In a strange way, that's almost worse. For the biggest event in world sport, they've somehow produced a logo with the emotional impact of a default folder icon.

Five stars for crop quality. One star for everything that happened afterward. Posted on 8 June 2026

The Skills AI Hasn’t Managed to Fire

Companies are trimming payrolls and feeding the money into artificial intelligence, and the math is changing faster than the weather on a bad coast. Across technology, finance, and other fields, management is flattening org charts, chasing efficiency, and putting capital into AI systems instead of more hires.

Anant Kale, co-founder and CEO of AppZen, an AI-focused finance company, sees no miracle fix in simply handing every worker a new tool. Some roles get sharper with AI. Some become far more productive. Some disappear in the shape they used to wear. In finance especially, AI agents are taking on process-heavy work once done by large operations staffs, pushing people toward work that carries more judgment and more value.

That shift leaves a bruise on the talent pipeline. Entry-level jobs long served as the back room where people learned the trade. If AI takes the repetitive chores, companies will need new ways to train beginners. The upside is that newcomers can spend less time grinding through transactions and more time learning outcomes, decisions, and strategy.

Kale argues that the durable edge is not prompt tricks but knowing how work actually moves. The workers who stay valuable will understand process design and optimization, AI governance and oversight, and business judgment. Technical AI skills will age. Systems thinking, domain expertise, leadership, problem-solving, and the ability to keep learning likely will not.

Posted on 7 June 2026

Auto Industry Swerves From Certainty to Catch-All Strategy

At the A. Alfred Taubman Center for Design Education, the BFA Degree Show opened on May 15th with a brisk, bright-eyed sense that the car world has driven into fog rather than sunrise.

Jason White, assistant professor at CCS, highlights a group of standout projects against a wider industry mood that has shifted sharply. For years, the route ahead looked almost diagrammatic: electric vehicles would sweep through new-car sales, old marques would emerge reborn in glossy, radical forms, and the public would clap politely while charging. That script has not unfolded.

Instead, the global automotive business is confronting a stubborn mismatch between aspiration and appetite. Manufacturers made large, confident plans for an EV-led future, but the market has answered with hesitation. Buyers, by and large, have not committed themselves fully to the version of tomorrow presented to them.

White frames this moment through a line from Terminator 2: Judgment Day, spoken by Sarah Connor, describing the future as a black highway at night and history being invented on the move. It fits. The certainty has drained away; the map is less map, more smudge.

For now, the industry response is practical rather than prophetic: keep the buffet open. In other words, major automakers are trying to serve every sort of customer at once.

Posted on 6 June 2026

Audi’s Nuvolari Is a Hybrid Supercar Meant to Drag the Brand Into Its Next Design Era

Audi has wheeled out the Nuvolari, the second big marker of Massimo Frascella’s influence since he arrived as chief design officer in June 2024. If the earlier Concept C was the manifesto, this is the loud, expensive punctuation mark at the end of the sentence: a monolithic supercar built to broadcast the brand’s new Radical Next design language from the top of the range downward.

Like the well-received Concept C, the Nuvolari was developed in-house and carries the same stripped-back, rationally minimalist attitude. But where the Concept C points to a purely electric future, the Nuvolari takes a more theatrical route. Its drivetrain pairs a twin-turbocharged 4.0-litre V8 with three axial-flux electric motors in a plug-in hybrid setup, suggesting Audi still sees value in mixing voltage with violence when it wants to stir enthusiasts.

The timing is telling. Order books are due to open in late 2026, with first deliveries scheduled for 2027. That puts the Nuvolari into customers’ hands before the production version of the Concept C, despite the latter’s role as the cleaner, EV-led vision.

This is more than a halo car. Audi is using it as a battering ram: to reinforce emotional attachment to the brand, reset expectations around its design direction, and make Frascella’s second statement impossible to ignore.

Posted on 5 June 2026

Ugly Buildings, Elevated Cortisol

Housing fights are often staged as spreadsheets versus sympathy, but the nervous system keeps sneaking into the room. Research from David Broockman, Chris Elmendorf, and Joshua Kalla finds that resistance to new housing is strongly shaped by whether buildings look harsh, awkward, or out of place. People are not only reacting to who might live there; they are reacting to what they must look at.

Environmental psychology and neuroscience help explain why. Streets with trees, steady building heights, and three- to five-story forms tend to feel legible and sheltered. That prospect-and-refuge balance lowers stress markers, including heart rate and cortisol. Green views even correlate with faster hospital recovery. By contrast, visual disorder, blank glass or metal facades, and jagged forms can keep people subtly on edge. Humans seem to prefer rhythm with variation: repeated windows, shopfronts, awnings, textured materials, rounded details, and the fractal richness common in nature. Slow streets matter too. Narrow lanes, pavers, outdoor seating, and active storefronts make people linger instead of flee.

A 2025 study by Adrian Pietrzak and Tali Mendelberg found that mismatched height draws more objection than height alone. A taller building that belongs can fare better than a shorter one that doesn’t. In Los Angeles, the Livable Communities Initiative says five years of outreach found near-universal support for housing when it is human-scale, well-designed, and paired with walkable streets. Support collapsed by 90% to 95% for low-quality, out-of-scale projects.

Posted on 4 June 2026

Ferrari’s Electric Debut Runs Cold

Ferrari finally brought its electric future into the open on 1 June 2026, and the result landed like a champagne coupe dropped on a marble floor. The Luce is a striking thing to look at, sure enough, but it doesn’t convince as a Ferrari. It feels less like an object of longing than a carefully resolved consumer product, and that is where the trouble starts.

For years, mainstream attention rarely found car designers unless something had gone badly sideways. Jaguar’s recent rebrand was one example, swallowing the car beneath the noise. Ferrari has managed a rarer feat: the product itself has taken the hit. The Luce has drawn a wave of disbelief and disgust that cuts across the usual fault lines of taste.

Part of that was always coming. An electric Ferrari was bound to offend traditionalists. Enzo Ferrari built the marque on the primacy of the engine, treating the car around it as secondary. He also insisted the best Ferrari was always the next one. Those two ideas now collide head-on in the Luce.

LoveFrom’s contribution to the Ferrari story is fascinating as industrial design. As brand expression, it is another matter. The car may point toward Maranello’s future, but it arrives stripped of the emotional voltage that made the badge matter in the first place.

Posted on 2 June 2026

Anthropic Wants Claude to Be Your Design Department’s Fastest Intern

Anthropic Labs has unveiled Claude Design, a research-preview tool released on Apr 17, 2026 that turns Claude into a brisk, slightly unnerving creative accomplice. Built on Claude Opus 4.7, the company’s strongest vision model, it is rolling out gradually to Claude Pro, Max, Team, and Enterprise users.

The premise is simple and rather intoxicating: ask for a visual thing and get a usable first draft, whether that’s a prototype, slide deck, one-pager, mockup, landing page, campaign asset, or something more experimental involving voice, video, shaders, 3D, and embedded AI.

For designers, it promises more room to try wildly different directions without burning days on each one. For founders, marketers, and product managers, it lowers the terrifying blank-page moment. Edits can happen through chat, direct text changes, inline comments, or Claude-made sliders for layout, spacing, and colour.

A notable trick is automatic brand alignment. During setup, Claude Design reads a team’s codebase and design files to assemble a design system, then applies the right colours, type, and components across future projects. It can also import prompts, images, DOCX, PPTX, XLSX files, website captures, and codebases.

Teams can share work across an organisation, export to Canva, PDF, PPTX, HTML, or pass a handoff bundle straight to Claude Code. More integrations are promised in the coming weeks.

Posted on 1 June 2026

Meta Pins Its Hopes on an AI Pendant

Meta is said to be preparing another wearable: an AI pendant slated for testing within the next year. The plan appears in an internal memo and suggests the company is still determined to find a gadget people will actually clip on willingly.

The pendant would likely draw on technology from Limitless, the AI-device startup Meta bought at the end of 2025. Limitless made a small wearable users could pin to a shirt or hang on a chain, with the purpose of recording conversations. Meta described that acquisition as a way to speed up its push into AI-enabled wearables.

There is, of course, a mild whiff of history repeating itself. Earlier AI wearables have largely failed to charm the public, whether because people were uneasy about privacy, put off by clumsy promotion, or simply unconvinced they needed one more thing listening near their collarbone. Even so, rivals including OpenAI continue to pursue the category.

The same memo indicates Meta wants to broaden its range of AI glasses and introduce a business offering called Wearables for Work. Taken together, the projects look like an attempt to improve the fortunes of Reality Labs, Meta’s hardware arm, which posted a $4 billion loss in the first quarter of this year.

Posted on 31 May 2026

A Labyrinth Built from Bad Carpets and Better Precision

A24’s Backrooms takes an internet curiosity—a yellow, deserted room that became a creepypasta fixture in 2019—and enlarges it into a feature about Clark, played by Chiwetel Ejiofor, an architect diminished into the management of a big-box furniture mart called Capn Clarks Ottoman Empire. He passes through a wall and into Kane Parsons’s unnerving maze of liminal interiors: beige carpet, fluorescent glare, drop ceilings, corridors without purpose, and rooms behaving badly.

Production designer Danny Vermette built much of this world physically, though Parsons first devised it in Blender, the same software he used as a 16-year-old making the original shorts. His initial file was so vast it crashed Vermette’s computer, and the team had to choose carefully what to construct. They ultimately spread 30,000 square feet of set across four sound stages, reserving practical builds for spaces demanding bodily discomfort or vertical strain: 20-foot risers, sloped floors, narrow crevasses, tunnels, doors absurdly high on walls.

The horror depends as much on design failure as on emptiness. Vermette linked liminal unease partly to the careless volume-driven spaces of the 1970s through 1990s. To match the shorts, the crew tested 30 wallpaper-carpet pairings and multiple lighting schemes. Outside the backrooms, unease persists: Async echoes drab 90s data centres, Mary Kline’s office and the Ottoman Empire revel in aggressively unaesthetic 1990s fabrics and furniture. Vancouver architects may even spot technical drawings in Clark’s office borrowed from Vermette’s architect father-in-law.

Posted on 30 May 2026

When a Session Expires, Access Can Vanish

On the modern web, session timeouts are often treated as a small engine-room matter of security. In practice, they can bar the door to essential services. For many of the 1.3 billion people worldwide with significant disabilities, being logged out mid-task can mean a lost ticket, a missed appointment, or a ruined loan application.

The burden falls unevenly. Motor impairments such as cerebral palsy can slow input; the DWP Accessibility Manual notes adaptive technology may need repeated attempts to register actions. Cognitive differences matter too: about 20% of people are neurodivergent, and conditions including ADHD, dyslexia, autism, Down syndrome, dementia, and traumatic brain injury can make rigid timers hostile, especially for those with time blindness. Vision loss adds further delay: more than 43 million people are blind and 295 million have moderate to severe visual impairment, so screen-reader navigation simply takes longer. Bogdan Cerovac’s experience showed how a one-second countdown can flood assistive tech and make a page unusable.

Common failures are silent expiry, warnings that come too late, nonextendable sessions, and erased form data. The DS-260 visa form logs users out after roughly 20 minutes idle, saves only on page completion, and offers only an approximate estimate. Better practice exists: the U.K. pension credit application warns at least two minutes before expiry and allows extension, aligning with WCAG 2.2 AA. Activity-based limits are often harsher than clear absolute ones. Autosave through cookies, localStorage, or sessionStorage can preserve work across reauthentication. Some limits remain justified, such as 10-minute ticket holds or shared computers, while untimed reading or shopping rarely needs forced expiry. With 62% of adults with disabilities owning a computer and 72% having high-speed home internet, accessible session design is not edge work but basic respect. Posted on 28 May 2026

AI Needs More Than a Human in the Loop

AI projects seldom collapse because the model is mathematically inept; they collapse because the organisation has mistaken supervision for authority. Dan Leiva, founder of CXAmplify and a former operations leader at Apple, Intuit, eBay, and Travelers, argues that the peril lies in the stretch between technical preparedness and human preparedness.

He identifies three recurring failures. First, decision boundaries drift: work that should require judgment quietly becomes automated approval. Second, frontline teams cannot explain why the system acted as it did, so questioning withers into rubber-stamping. Third, there is no safe way to halt the machine when it behaves plausibly yet wrongly.

Leiva’s remedy, developed in Amplified, is a three-class model. Class 1 decisions are low-impact, reversible, and rule-based, suitable for autonomy. Class 2 decisions affect experience, access, or treatment and require a named human decision-maker. Class 3 decisions carry financial, legal, ethical, or health consequences; here the human must own the outcome and AI merely informs it.

He also calls for a Red Button Protocol with four essentials: named authority at several levels, immediate effect, traceable logs, and protection from retaliation. In complex AI chains, one person must own the final customer-facing decision across all systems. And alongside efficiency metrics, organisations should track Key Human Indicators such as override rates, escalation quality, explainability confidence, and red-button use and resolution times. Posted on 26 May 2026

Humanoids Face a Long Slog Before the Boom

Humanoid robots are still more promise than graft. Annual shipments remain under 100,000, with demand propped up by subsidies, pilot projects, and strategic tie-ups rather than hard-nosed labour economics. The real market, though, may start to bite in 2032. Interact Analysis projects shipments topping 700,000 units by 2035, with revenue reaching about $15 billion.

That future depends on a few brutal hurdles being cleared. Humanoids need to make financial sense, embodied AI has to become reliable enough for autonomous work, and the machines must pass through regulation, certification, insurance, and efficiency tests that are still half-built or missing entirely. Marco Wang of Interact Analysis points to weak embodied AI, scarce data, uneven hardware durability, and shaky manufacturing consistency as the main choke points.

There’s also the human factor. Safety worries, uncertain reliability, fear of job loss, and the uncanny valley all keep adoption sticky. On factory floors, cheaper cobots already do useful work, so humanoids must prove they can outperform them and integrate cleanly.

Near-term deployment is centred on structured settings like manufacturing and warehousing. Public services follow, especially through Chinese state-backed programs. Home use stays distant because safety and messy environments are harder. By 2035, China is expected to take more than 65% of real-world humanoid shipments, with the US a distant second; together they should account for over 85% of demand. Wheeled platforms are emerging as the more practical near-term design.

Posted on 25 May 2026

Google Turns the Pixel Home Screen Into a Tiny Nightclub

There is, in the contemporary handset economy, no indignity so severe that it cannot be repackaged as “fun.” Google has now lacquered Android in the same mirror-tile absurdity that recently engulfed Spotify’s app icon for its 20th anniversary: disco-ball icons, gleaming with the sort of tasteless confidence usually associated with nightclub banquettes and warm prosecco.

The option arrived Friday for Pixel users through the custom icons feature introduced in March’s Pixel Drop. That update had already offered AI-generated styles such as Scribbles, a hand-drawn look; Treasure, in gold; and Easel, a painted, more chromatically agitated treatment. Previously, users were largely confined to recoloring icons so they harmonized with wallpaper and theme.

Earlier in the week, Android ecosystem head Sameer Samat had floated the notion on X by posting a discofied Chrome icon. By Friday, the whole thing had been made real, complete with a fully glittered Pixel home screen and a half-mocking prompt asking whether people truly wanted it.

They apparently do, or at least they want the spectacle of wanting it. Spotify’s version had drawn predictable disgust, with the company stressing its temporary nature. Google, by contrast, seems content to lean into the joke. The result is hideous, camp, faintly charming. A home screen becomes bottle service.

Others are joining in too: Lovable has released a tool that applies the same disco-ball treatment to logos.
Posted on 24 May 2026

A Flying Saucer Lands on the Inner Mongolian Grassland

Out on the Ulanqab Grassland, 160 kilometres west of Beijing, BÜro Ziyu Zhuang has plonked what looks like a crashed UFO beside Laoli Lake and, somehow, made it sound perfectly sensible.

Called Prairie Ark, the public gallery in Inner Mongolia is shaped as a sloping disc, buried into the ground at one end and lifting toward the sky at the other. The idea is not to trot out the usual grassland clichés of yurts, nomads and a bit of Genghis Khan wallpaper, but to use the site’s isolation to create something more stripped back and immersive.

The roof runs straight out of the terrain, so visitors can wander up from the grass and enter from above, while openings at basement and ground-floor level fold the building into the landscape. Ziyu Zhuang described the entrance as a cut in the land rather than a neat architectural threshold, likening it to the spaceship lodged in the mountains in Prometheus.

Inside, Prairie Ark is built for exhibitions, conferences, brand events and informal gatherings. A layered grid ceiling punctured by skylights drags daylight deep into the interior, while the ground floor remains open and free of partitions.

Nearby, the studio added Nomads Beacon Tower on the east shore of Laolihai Lake. Referencing rising beacon smoke and the Great Wall’s towers, it is reached by a route that can disappear underwater in summer, leaving it stranded like a moody monument. Posted on 23 May 2026

Google Returns to the Face, Older and More Cautious

Google, having once planted its face in the gravel with Google Glass, is back at the bridge of the nose. More than ten years after that 2013 misadventure — withdrawn in 2015, a mere seven months after reaching the UK, amid fuss over cost and surveillance — the company says a new pair of smart glasses will arrive this autumn.

Unveiled at Tuesday’s developer conference, the device comes in two designs, one by Warby Parker, the other by Gentle Monster. A camera sits in the frame; speakers hide in the arms. The idea, said Google executive Shahram Izadi in Mountain View, California, is to keep wearers “hands free and heads up”. The glasses will pair with Android and Apple’s iOS, and are built to deliver “all-day help with Gemini” spoken “into your ear privately” rather than through a visible display.

Google is also developing a version with an in-lens display for text and other information, though not for immediate release. Izadi said more details would emerge later this year, adding that developers are already building apps for it.

The wager is obvious: Meta’s AI Ray-Bans, with camera and speakers, have sold seven million pairs. But the old unease has not aged out. People are still being filmed without consent, in streets and living rooms alike, and only discovering it when the footage migrates online.

Christine Tsai of 500 Global called Google’s return good for consumers and start-ups alike. Smart glasses, she said, may be “that next modality” after the smartphone. Developer Anil Shah, behind tixfix.ai, said the appeal lies in letting Google Maps and Google Voice escape the phone altogether: “being able to just talk with the smart glasses without opening the app would be a very nice integration.”

Posted on 20 May 2026

Chongqing’s Latest Answer to Gravity

Chongqing has never really met a hill it didn’t want to complicate. This is the city with the world’s deepest subway station and that famous suspended rail line threading through an apartment block as if that were perfectly normal. Now it has added another entry to the list: the Wushan Goddess Escalator, officially the longest outdoor escalator system on Earth.

It sits in the Gaotang area of Wushan County and links the upper and lower parts of Shennv Avenue, where getting from one to the other used to take about an hour. The new arrangement cuts that to roughly 20 minutes, which is the sort of efficiency gain that sounds dull until you’ve had to climb it.

The numbers are, inevitably, excessive. The full system runs about 905 meters and rises 242.14 meters, roughly the height of an 80-storey building. To achieve this, Chongqing hasn’t relied on one heroic escalator but on a small civilization of machinery: 21 escalators, 8 elevators, 4 moving walkways, 2 pedestrian bridges and 2 overpasses.

A paid trial began on February 17, with each trip priced at 3 yuan, or about $0.40. Officials will decide after the trial whether that flat fare stays in place. Either way, it’s a neat illustration of Chongqing’s gift for turning awkward geography into public transport with ambition, steel and a faint sense of absurdity.

Posted on 19 May 2026

YouTube Expands AI Likeness Protection Beyond Its Inner Circle

YouTube is finally handing more adults a little digital holy water for the demon of AI impersonation. In the next few weeks, every creator age 18 and up will be able to use a likeness-detection feature that flags videos uploaded to YouTube that may have copied their face with AI.

The company said on its Community page that the point is to give users “more peace of mind” and “easy access to request the removal of unauthorized content.” Although the feature is officially framed for creators, YouTube spokesperson Jack Malon told The Verge that anybody can use it. “With this expansion, we’re making clear that whether creators have been uploading to YouTube for a decade or are just starting, they’ll have access to the same level of protection,” Malon said.

That matters because AI video fakery is getting slick enough to fool people, and the fallout isn’t just embarrassing — it can be manipulative, malicious, or commercial. Creators may also catch brands using their image to sell things without consent.

YouTube first teased the tool in 2024, then launched it in late 2025 for Partner Program creators, followed by journalists and politicians. To enroll, users must use YouTube Studio on desktop, open Likeness under Content detection, scan a QR code, submit a government ID, and record a selfie video. After that, YouTube scans uploads for face matches. Users can review flagged clips and file removal requests. The system can ask about copied voices, but it still can’t detect voice-only misuse.

Posted on 17 May 2026