GeorgesHung

Density is a number. Intensity is how a city feels to live in.

Asia and the Middle East will write the next chapter of the world’s cities. They will also write it under the hardest conditions: extreme heat, scarce water, some of the oldest populations on earth and some of the youngest. Two forces will decide how they respond. AI, which can plan and run a city with new precision while adding a vast new load to it. And sustainability, which in this region means adapting to heat and water as much as cutting carbon.

My argument is simple. The goal is not a denser city, or a smarter one. It is a more intense one: a city that offers more life, comfort and opportunity per unit of land, carbon and water. I call it Intencity.

Why now

Five forces are converging. Each is serious alone. Together they ask for a different way of planning, building and running cities.

19 of 33

of the world’s megacities are in Asia. Cities will absorb two-thirds of global population growth to 2050.

1.6 → 5.6 bn

air conditioners worldwide by 2050. Heat now decides how far people walk and how much energy a city burns.

< 500 m³

of water per person a year across MENA by 2030, below the threshold of absolute scarcity.

1 in 4

people in Asia-Pacific will be over 60 by 2050. Saudi Arabia’s median age is 29.

415 → 945 TWh

data-centre electricity, 2024 to 2030. AI is arriving as a tool and as a load.

Georges Hung at the International Design Forum, Guangzhou Design Week, 2021Guangzhou. Image: Guangzhou Design Week – IDC 2021.

What the consensus gets wrong

For a decade, urban thinking in the region has rested on three confident ideas: build up, build smart, and bring daily needs within fifteen minutes. Each has merit. Each now needs testing.

• Density is not intensity. For the same population, high-rise districts can carry 132–154% more life-cycle emissions than dense low-rise ones. Height is a choice with a carbon cost. What makes density worth living with is the public life, proximity and comfort it delivers.
• Fifteen minutes is a principle, not a template. A fifteen-minute walk means something else at 45°C, on a humid evening, or at the age of eighty. Riyadh’s new metro—six lines, 176 km, 85 stations—will succeed or fail in the last few hundred shaded metres between station and front door.
• A smart city is not automatically a good one. Hangzhou’s City Brain cut congestion sharply, and raised hard questions of privacy and control. Technology-first megaprojects have proved fragile; The Line at NEOM has been scaled back. Cities grow best through evidence, iteration and trust.

AI can make a city more efficient. Only design can make it more worth living in.

AI and sustainability, properly understood

Used well, AI makes design more evidence-based. Digital twins such as Virtual Singapore test wind, sunlight and heat before anything is built; generative tools test hundreds of massing options against daylight and carbon in hours. Across buildings alone, AI could save around 300 TWh of electricity. Yet AI also runs on power, water and land, and in the Gulf it is being planned at gigawatt scale. Data centres are urban infrastructure. They belong in the same plan as housing, transit and water.

The deeper change is social. As information, transactions and routine work move to screens, the value of the city shifts to what only presence can give: learning from others, shared public life, trust and belonging. The public realm matters more, not less. And in hot climates, sustainability starts with comfort—shade, water, planting and transit are what let people live well outdoors and use less energy indoors.

Defining Intencity

The intense city is one where daily needs are close and comfortable to reach, public life is open to every age, resources are used with care, and technology serves the public with its trust. Five tests set it apart from a city that is merely dense or merely smart.

1. ProximityCan residents reach daily needs, work and transit within a comfortable walk or short ride, in every season?
2. ComfortAre streets and public spaces shaded, cooled and usable for most of the year and most hours of the day?
3. ResourceAre carbon, energy and water per resident falling, including the demands of digital infrastructure?
4. BelongingIs there free, safe public space and culture for every age, income and background, rooted in local story?
5. IntelligenceIs data used transparently, with consent and accountability, to improve public life rather than only to control it?

Six shifts

1. From density to intensity. More life per hectare, not more floors. Towers belong on transit nodes, and must earn their height through the public life they carry—as Singapore’s Kampung Admiralty does, stacking a hawker centre, clinic and gardens beneath homes for older residents.
2. From car city to shaded transit city. Redefine the fifteen-minute city as fifteen minutes in comfort. Continuous shade, covered links and places to pause, measured by walking time in the hottest month.
3. From green image to climate infrastructure. Trees and parks are cooling systems, flood defences and public health. Native planting, reused water, reflective materials and district cooling, with shade treated as a utility as essential as lighting.
4. From smart city to civic intelligence. AI designed around trust: digital twins shared with the public, open data, clear rules on use, and a human alternative for every digital service.
5. From one generation to a city of all ages. Schools, clinics, markets and parks placed together, so grandparents and grandchildren share the same daily routes. Design every street for the slowest walker.
6. From megaproject to living fabric. Renewal before replacement. Existing buildings hold carbon, memory and community. Record the stories of a place before changing it, and let local character, climate and culture guide new form.

The designer’s role

“We as architects and urbanists are conduits of experientially driven urban spaces—as connectors, curators, and life creators.”

• ConnectorLinking transit, energy, water and data networks with the public life they should serve, and bringing climate engineers and data scientists around one idea of place.
• CuratorChoosing among uses, evidence and tools—AI included—with judgment. As machines produce more options, the value lies in choosing well and explaining why.
• Life-creatorMaking the occasions for presence no algorithm can supply: the shaded square, the shared table, the street that invites people to stay.

What this means

The city worth being in

The future of these cities will not be decided by how tall they grow or how many sensors they install. It will be decided by whether people can walk in the shade to meet a friend, whether a grandmother and a child share the same park, and whether the city still feels like itself.

The future city will not be measured by how high it rises or how much it knows, but by how fully people can live in it.

Related service · Vision & Brief →

Cities are adding people faster than they can add ground. The answer has been to build up. The question is whether public life follows.

In Hong Kong, Singapore, Tokyo and the cities of the Gulf, density has gone vertical. Too often, public life has not. It stops at the podium roof, behind a turnstile, or at the door of an air-conditioned mall.

My argument: the next generation of dense development will be judged less by its height than by the public life it carries in section. I call this the vertical commons—publicly accessible, climatically comfortable space, connected to street and transit, distributed through the height of a building or district, and cared for over the long term.

Why now

Four pressures are converging on the dense city. Each alone would change how we design the public realm. Together, they make the case for a new model.

8 → 33

megacities, 1975 to today. Nineteen are in Asia. For land-constrained cities, the only reserve left is the section.

37%

of global CO₂ comes from buildings and construction. As grids decarbonise, every storey carries a material argument.

August

decides occupation. In hot cities an unshaded plaza is an empty plaza—a public realm that fails in summer fails for a quarter of the year.

Private

developers now deliver much of the new public space in dense cities. That changes who decides who is welcome, when, and for what.

Workplace amenities wrap the inverted HQ tower, opening onto green terraces and natural ventilationCoolpad HQ Tower, Shenzhen. Design: Architecture Georges Hung. Image: Georges Hung.

What the density argument gets wrong

Dense, mixed-use development protects open land, shortens trips and shares infrastructure. Nothing here argues against density. It argues against three assumptions that often travel with it.

• Dense does not have to mean tall. For a fixed population, moving from high-density low-rise to high-density high-rise raises life-cycle emissions by 132–154%. A fabric closer to Paris than Manhattan can house around 47,000 people per km² at a fraction of the carbon. Towers are not wrong—but height is a choice with a carbon price.
• Height changes social life. Four decades of research find high-rise living less satisfying for most residents, harder for families with young children, and more impersonal. Not destiny—the shared corridor, the sky terrace and the lift lobby as a place to pause are all variables designers control.
• Elevated does not mean public. Singapore’s Pinnacle@Duxton links seven 50-storey towers with sky gardens—yet one is residents-only and the other ticketed, at 150 visitors a day. A public realm can be drawn generously and still operate narrowly.

Height is a choice with a carbon price. The public life it carries is how the price is repaid.

Defining the vertical commons

Publicly accessible space, distributed through the section of a building or district, continuously connected to street and transit, holding programme people need, comfortable across the year, and secured for long-term care. A roof garden visible only from a penthouse is not part of it. Neither is a sky lobby that needs a building pass. Five tests separate a commons from an amenity.

1. AccessCan anyone enter, without a ticket, a purchase or a pass, for most of the day and evening?
2. ContinuityDoes it connect to street and transit by a legible, step-free public route—not only to a private lift lobby?
3. PurposeDoes it hold programme people need every week—food, health, care, learning, play—not only views?
4. ComfortIs it shaded, ventilated and usable in the hottest and wettest months, and after dark?
5. StewardshipWho programmes, maintains and pays for it in year ten, and is that commitment secured in title?

Six shifts

1. From podium to section. Repeat the ground plane at height: sky lobbies that behave as streets, elevated decks, terraces linked by public vertical circulation—as Hong Kong’s walkways and Mid-Levels Escalator, and Shibuya’s “urban cores”, already do. Treat public lifts and stairs as streets, not core.
2. From amenity to infrastructure. Views attract occasional visits; civic programme brings people back every week. Kampung Admiralty stacks a plaza, a 900-seat hawker centre, a medical centre and gardens beneath homes for older residents—a village in section. The test: why would someone come back on Tuesday?
3. From green image to climate instrument. Measure planting by shade, cooling and hours of outdoor comfort, not square metres. Singapore’s LUSH 3.0 requires lost greenery to be replaced at height. In the Gulf, design the terrace for 9 pm in August, not noon in March.
4. From tallness to the right height. Ask what height buys. Test a mid-rise alternative at massing stage, retain existing structure before demolition, use timber and hybrid structure where codes allow, and report embodied carbon per resident.
5. From privately owned to publicly held. Codify publicness—opening hours, street signage, free seating, permitted activities—in title and planning agreements, and publish the data so it can be audited, not assumed. Every public space on a drawing needs a matching obligation in the lease.
6. From handover to stewardship. Thriving spaces have an operator, a programming budget and a feedback loop. Budget ten years of operations at feasibility. And give height meaning: local names, craft, memory and recorded stories turn an elevated room into a place residents recognise as theirs.

The framework

The five tests—plus carbon—become indicators that can be set at brief, checked at design review and measured after occupation.

• AccessShare of public realm open 18+ hours a day without ticket, purchase or pass.
• ContinuityStep-free public routes from transit to the highest public level.
• PurposeShare of public floor area with civic or non-commercial programme; repeat weekly visitors.
• ComfortAnnual hours within the UTCI comfort band; share shaded at noon in the hottest month.
• StewardshipSecured operating budget per m² of public realm; programmed days per year.
• CarbonEmbodied carbon per resident and per m² of public realm, benchmarked against a mid-rise option.

What this means

Height as a public promise

Every tall building makes a claim on the city: its sky, its light, its infrastructure and its carbon budget. The vertical commons is how that claim is repaid. The next decade will not be defined by who builds highest, but by who carries the street upward with its openness intact—the bench anyone can sit on, the shade that holds in August, the clinic above the market, the terrace open at night.

The measure of a vertical community is not how high it rises, but how far its public life reaches.

Related service · Independent Review →

Knowledge is now instant. Attention is not.

For two centuries the school building had one clear job: to bring students, teachers and information into the same room. That job is over. Information reaches every student through a phone, and an AI model can explain, summarise and draft faster than any lecture.

My argument: the value of a place of learning has moved. It is no longer where information is delivered. It is where attention is protected, judgment is formed and belonging is built. I call this the attentive school.

Why now

Four forces are converging on education at once. Each has been discussed on its own. Few school briefs yet respond to all four.

95%

of UK undergraduates now use generative AI. When a machine can produce an answer, education must teach the judgment to question it.

65%

of 15-year-olds across the OECD are distracted by devices in maths lessons. Phone bans have spread from 24% to 58% of education systems since 2023.

−14%

Hong Kong primary enrolment in five years—while young populations elsewhere need schools faster than they can be built. One sector, two opposite briefs.

44 million

additional teachers needed worldwide by 2030, as primary attrition nearly doubled. Buildings rarely count in teacher retention. They should.

What the flexible-school consensus gets wrong

Two decades produced a clear consensus: flexible, technology-rich, student-centred, collaborative, sustainable. Each is sound. Together they became a checklist—and a checklist is not a thesis.

• Reconfigurable is not the same as effective. The HEAD study of 3,766 pupils in 153 classrooms found design explained up to 16% of a year’s learning progress—driven by daylight, air, temperature, ownership and the right level of stimulation. Secure the basics first.
• More technology is not more learning. Students given an unrestricted chatbot did better in practice, then scored 17% worse on the exam without it. A version designed as a tutor avoided the loss; a well-designed AI tutor at Harvard more than doubled learning. Design, not access, decides the outcome.
• Open is not the same as collaborative. Without acoustic separation, open learning areas invite noise, not collaboration. Collaboration needs enclosure as much as openness: a table for four with a door that closes, beside a larger room, beside a quiet corner for one.

A movable wall does not compensate for a room that is stuffy, glaring or loud.

Defining the attentive school

A place of learning designed to protect attention, form judgment and build belonging, connected to the life of its city, and able to adapt as its population and pedagogy change. It applies equally to a primary school, a university building or a corporate learning centre.

1. AttentionDoes every learner have a calm, daylit, well-ventilated space for focused work, with devices managed by design rather than by rule alone?
2. JudgmentAre there spaces to make, test, present, critique and defend work in person, where thinking is visible and cannot be outsourced?
3. BelongingDo the in-between spaces—stairs, thresholds, dining, courtyards—give every student somewhere to be known?
4. ReachIs the building connected to work, community and city, and open to learners of every age beyond school hours?
5. AdaptabilityCan it serve a smaller, larger or different population in twenty years without demolition?

Six shifts

1. From device-rich to attention-rich. Daylight without glare, fresh air, stable temperature and acoustic control—the core of both learning and low carbon. Lockers at the entrance and device-free zones marked by material and light make the calmer choice the easier one. Draw device zones on the plan, not only in the handbook.
2. From content delivery to judgment formation. When AI can write the essay, assessment moves to what cannot be outsourced: live work, oral defence, making and critique. Fewer rooms sized for listening; more studios, workshops and presentation rooms. The architecture studio, with its desk crits and public reviews, is a model for every discipline.
3. From screen networks to real community. School may be the last place young people spend long hours together in person. Belonging is built between classrooms—a generous stair, a dining hall that works as a living room, a shaded courtyard. These are the first spaces cut when budgets tighten. They should be the last.
4. From campus to city. Share labs with industry, libraries with neighbours, sports with the community. A building used six hours a day, 190 days a year, is an underused civic asset; opened for adult learning and local enterprise, it becomes neighbourhood infrastructure—and the building itself can teach sustainability.
5. From fixed capacity to adaptive estate. A school sized for today may be half empty in fifteen years, or twice as full. Structural grid, floor heights, cores and independent wings decide whether it can become elder care, community hub or housing. Test every new school against two alternative futures at concept stage.
6. From building for pupils to building for teachers. The teacher’s workplace is part of the learning environment. Distribute staff space close to teaching, with places to plan together, concentrate and recover. AI can remove hours of administration each week; the building should give that time a place to go.

The framework

The five tests, plus a sixth for the teacher’s workplace, become indicators set in the brief, checked at design review and measured after occupation.

• AttentionShare of teaching spaces meeting daylight, CO₂, temperature and reverberation targets; device-free zones on plan.
• JudgmentFloor area for studio, workshop, lab and presentation, compared with lecture and classroom.
• BelongingIn-between space per student; share of students who say they have a place where they are known.
• ReachHours a week open to the community; active external partnerships.
• AdaptabilityAlternative future uses tested; structural grid and floor-to-floor height.
• TeachersStaff workspace per teacher and its distance from teaching; staff satisfaction with the building.

What this means

The last shared room

Students no longer come to school because that is where the knowledge is. What remains is what a screen cannot give: the attention of a teacher, the friction of a good argument, the satisfaction of making something by hand, and the experience of being known by name. For many young people, school is now the last shared room—the one place they spend whole days together, in person, away from the feed. The task is no longer to deliver information. It is to build places worth being present in.

The school is no longer where knowledge is kept. It is where attention is given.

Related service · Vision & Brief →

When an algorithm + app can do the shopping, a retail centre has to earn its reason.

The retail centre was built around a simple exchange: people travelled to find, compare and buy. AI is taking over that errand. Yet physical retail is not disappearing—close to four in five purchases still happen in a physical place. When in-chat checkout launched in 2025, it was withdrawn within six months. People were happy to let AI research their purchases. They were far less willing to let it decide.

My argument: the future retail centre will not be a place for transactions. It will be a place people choose to go—for what a screen cannot give, for the needs of every week, and for the company of others. I call this the reason to go.

Why now

Four forces are converging on physical retail. None is new on its own. Together, they change what a retail centre is for.

+693%

AI-referred traffic to US retail sites over the 2025 holidays. Discovery has moved upstream—what AI cannot read, it will not recommend.

> 50%

of US consumer spending comes from households over 50, while 63% of Saudi nationals are under 30. The same format cannot serve both.

1 in 6

people worldwide are affected by loneliness. Few places still offer free, safe, all-weather space to simply be together. The retail centre is one of them.

93 million

cross-border trips by Hong Kong residents in 2024, up 48%. Shoppers travel for value and experience; centres now compete with the whole region.

What the mixed-use retail story gets wrong

The post-pandemic consensus held that mixed-use centres—shaped by Gen Z, digitally integrated, rich in experience—would replace the traditional mall. Much of that is right. Three assumptions inside it need testing.

• One generation is not a strategy. Gen Z is digitally fluent and enthusiastic about physical stores, but spending power sits with people over 50—healthier, more active and more digital than any older generation before. The real question is how to serve a teenager, a young family and an 80-year-old in the same visit, with the same dignity.
• Digital add-ons are not a digital strategy. Virtual fitting rooms treated technology as a feature inside the building. Today the decision to visit is made before leaving home, often by an AI summarising what is open, in stock and worth the trip. Strategy is less about screens on walls, more about accurate, machine-readable data.
• A mix of uses is not a place. Gyms, co-working and galleries under one roof can still feel like a mall. A tenant mix describes what is leased; a place is defined by what is shared—free public space, programming that brings people back, and management that treats visitors as citizens, not traffic.

The question is not which generation to design for. It is how to serve all of them in the same visit.

Experience, properly understood

Almost every new centre claims “experience”. In practice the word means a tenant category, a spectacle or a mood-board adjective—all designed for the first visit. Spectacle depreciates; once seen, often first on a phone, there is little reason to return. Pine and Gilmore’s original idea was more precise: goods are the props, services the stage, and what people value is the memory. Understood that way, experience can be designed in seven spatial qualities.

• ArrivalThe passage from car park, metro or street: threshold, compression and release, the first view.
• TimeThe rhythm of a day and a year—morning coffee, the school run, Ramadan nights. Design for the fiftieth visit, not the first.
• ParticipationTasting, trying, making, learning and repairing. Active verbs create memory.
• SensesDaylight, acoustics, temperature, texture and scent. Honest materials that age well are read as care.
• Other peopleEdges to sit on, places to linger without spending. What attracts people most is other people.
• EncounterThe tailor, the barista who knows your order, the specialist who gives advice—the part an algorithm cannot copy.
• IdentityLocal craft, food, story and climate. An experience that could exist anywhere gives no reason to come here.

Two questions replace the tagword: Would someone come back next Tuesday? Would they come without buying anything? Measure by repeat visits and dwell time, not by how often a place is photographed.

Retail itself shifts from purpose to ingredient—but the ingredient that binds the others. It moves from errand to choice, from inventory to theatre, and above all gives life to the ground floor: shops light the street after dark and put “eyes on the street”. The operator becomes an editor as well as a landlord. The centre survives; “retail” becomes its adjective.

Experience is not a feature added to a space. It is a quality of the whole sequence.

Defining the reason to go

A place of exchange that offers what cannot be delivered, meets the needs of every week, welcomes every generation without requiring a purchase, is legible to people and machines alike, and can adapt as demand changes. Five tests separate a destination from a distribution point.

1. PresenceDoes the centre offer what only happens in person—trying, tasting, repair, advice, performance, meeting?
2. EverydayAre there anchors—food, health, services, learning—that bring people back every week?
3. BelongingIs there comfortable, free-to-use space where every age, from toddler to great-grandparent, is welcome to stay?
4. LegibilityCan people find their way easily, and can AI tools read accurate, live information on what the centre offers?
5. AdaptabilityCan space change use, size and tenure without major structural work as demand shifts?

Six shifts

1. From transaction to reason. People use AI to narrow choices, then want to decide themselves. Stores become smaller showrooms with more room for demonstration, fitting, personalisation, repair and workshops; stock moves behind the wall. Frontages show activity, not inventory.
2. From search engine to answer engine. Legibility now has two audiences: people, who need clear wayfinding, and machines, which need a live data layer of tenants, stock, services, events and access. Treat that data as infrastructure, and give collection, returns and last-mile logistics a place in the plan.
3. From one generation to every generation. Design for the least able visitor and everyone benefits. AEON MALL runs walking programmes and dementia-awareness training across its malls. In young markets, families, teenagers and evening life shape the brief. Test the plan with the oldest and youngest visitors.
4. From anchor store to everyday anchors. Grocery, clinics, fitness, childcare, libraries and government services create a weekly habit and fill the quiet hours. A clinic above a market and a learning centre beside a food hall make the centre part of the neighbourhood.
5. From mall to public realm. With one in six people lonely, free-to-use space is a public-health asset. Open-air streets, squares and rooftop gardens that stay open after the shops close—and, in hot climates, designed for the evening, when life in the Gulf moves outdoors.
6. From fixed GLA to adaptive asset. Demand changes faster than buildings. Long spans, generous floor heights and separable servicing let floors become homes, clinics, classrooms or logistics. Pop-ups and incubator units keep the offer fresh and reduce vacancy risk.

What this means

Worth the journey

For fifty years the retail centre was a machine for transactions. AI now finds the goods faster. What the algorithm cannot do is sit beside a friend, taste the food, ask the tailor, walk with a grandparent in the cool of the evening, or feel part of a crowd. Those are the reasons people will still go. The centres that thrive will be designed around them.

When an algorithm can run the errand, the centre has to give people a reason to go.

Related service · Independent Review →

When AI can do much of the work at a desk, the office has to offer what a desk cannot: each other.

The contest between home and headquarters is largely settled. Hybrid is the norm, and presence is rising where the office gives people a reason to come. The larger change is AI. It drafts, summarises, codes and analyses—and moves people’s work toward direction, judgment, review and invention. Its help is uneven: novices in one study became 34% more productive, while experienced developers in another took 19% longer, believing all the while that AI had made them faster.

My argument: the office’s value has moved. It is no longer where individual work gets done. It is where people learn from one another, decide together, and invent what neither they nor their tools could produce alone. I call this the studio workplace.

Why now

Four forces are converging on the workplace. Each is well documented. Few workplace strategies yet respond to all four together.

15×

rise in active AI agents in a single year—yet only 13% of users are rewarded for redesigning how work is done. The technology is ready faster than the workplace around it.

66%

US office occupancy on Tuesdays in late 2025, the highest since 2020. Hybrid works; people come in when there is a reason, and on the same days as colleagues.

20%

global employee engagement in 2025, the lowest since 2020. Managers fell from 27% to 22%—they carry the change and are the least supported.

97%

occupancy of Riyadh’s Grade A offices, while Hong Kong buildings over 30 years lose tenants. The question is no longer whether offices are needed, but which ones.

What the flexible-office consensus gets wrong

The post-pandemic consensus promised flexibility, collaboration, wellbeing and seamless technology. Together they described an office defined mainly by what it no longer required. Three assumptions need testing.

• Open is not the same as collaborative. When two Fortune 500 companies moved to open plan, face-to-face interaction fell by around 70%. Shared space supports collaboration only alongside enclosure: rooms where a team can argue, quiet places to think, settings where a conversation will not disturb twenty others.
• Home versus office is the wrong question. Two days at home cut resignations by a third with no loss of performance. What proximity adds is learning: juniors seated near teammates received 18% more feedback. Presence trades some senior output today for junior growth tomorrow—a trade design should make deliberate.
• AI is not another collaboration tool. It changes the work itself, unevenly. 41% of workers have received AI “workslop”—content that looks finished but isn’t, taking close to two hours to fix. AI moves effort from producing to reviewing, and review is human, social work that needs a place.

When everyone can see and hear everyone, people withdraw into headphones and screens.

Innovation as the differentiator

Like “experience” in retail, “innovation” has become a tagword—sensors, booking apps, a branded ideation room. A workplace cannot supply ideas. It can change the conditions in which they form, spread and get tested. A study of 61,000 Microsoft employees found remote work made networks more siloed, with fewer ties across groups—the weak ties where new combinations of knowledge begin. Six conditions matter.

• CollisionShared stairs, cafés and routes that cross team boundaries, so people meet those they would never schedule a meeting with.
• VisibilityWork on walls and screens where others can see, question and improve it.
• PrototypingWorkshops, labs, mock-up rooms and sandboxes close to where ideas are discussed.
• CritiqueRooms for honest review of work, human- or AI-made, with the right size, acoustics and display.
• ApprenticeshipJuniors near experienced colleagues, so tacit knowledge passes on by watching, asking and doing.
• EcosystemGround floors, labs and events shared with partners, universities, start-ups and the city.

Two tests: Did anyone meet someone this week, outside their team, who changed how they approached their work? Is the work visible enough that others could improve it? And the best organisations treat their own workplace as a prototype—pilot a floor, measure, ask, revise. A workplace built as a finished product ages from the day it opens.

A workplace cannot supply ideas. It can change the conditions in which they form, spread and get tested.

Defining the studio workplace

A place where people come together to learn, decide and invent; where work is visible and critique is normal; where focus and wellbeing are protected; and where the building can adapt as work and technology change. The architect’s studio—making, critique and apprenticeship, in company—is the right model.

1. ConveningDoes every office day have a reason to be there, shared with the people who matter to the work?
2. ApprenticeshipAre less experienced people close enough to experienced colleagues to learn by watching, asking and doing?
3. FocusCan anyone find a quiet, daylit, well-ventilated place for concentration or a private conversation—including with AI?
4. InventionDo space and routine create encounters across teams, make work visible, and allow ideas to be prototyped?
5. AdaptabilityCan floors, leases and uses change as headcount, technology and work patterns shift?

Six shifts

1. From desk to studio. The unit of design becomes the project, not the desk: team neighbourhoods with project walls, shared screens and tables for the whole team. When AI drafts in minutes, the scarce activity is deciding what is good—so put the work where people can see it.
2. From individual output to human–agent teams. Machines take on more execution; people set direction, review and decide. More small rooms for two to six, more enclosed space for thinking, and far more acoustic privacy. A floor where everyone talks to an assistant cannot be open plan.
3. From attendance to apprenticeship. AI is absorbing the routine tasks through which juniors once learned. The office is where that ladder is rebuilt—like a teaching hospital. Seat juniors near experienced colleagues on anchor days, with focus rooms close by to protect senior time.
4. From open plan to a range of settings. Libraries for quiet, team rooms for noise, workshops for making, terraces and cafés for encounter—each acoustically separated. Wellbeing starts with daylight, fresh air, thermal comfort and sound, and benefits everyone, including neurodivergent colleagues.
5. From tower to district. The office competes with home, and home is quiet and free. What the office can offer is the city: food, culture, transit and chance encounters. Share ground floors and podiums with the public; in Riyadh, the district around the office will shape where talent chooses to work.
6. From fixed lease to adaptive asset. Occupancy peaks midweek; headcount changes faster than leases. Retrofit before replacement—light retrofits can cut energy around 17% with payback under a year. Design new offices with generous floor heights and separable servicing for future conversion.

The framework

The five tests become indicators set in the brief, checked at design review and tracked in use.

• ConveningShare of in-office time spent with others; peak-day utilisation versus average.
• ApprenticeshipJuniors seated near senior colleagues on anchor days; frequency of feedback reported.
• FocusEnclosed, acoustically private settings per 100 people; daylight, air and reverberation targets met.
• InventionCross-team interactions; floor area for making, prototyping and review; pilots run per year.
• AdaptabilityFloor area convertible without structural change; lease flexibility; retrofit plan in place.

What this means

Worth the commute

For a century, the office was where work lived: the files, the machines, the people who knew how things were done. Much of that now lives in the cloud and, increasingly, in AI. What remains is what cannot be sent through a network: learning a craft by watching someone better, reaching a hard decision in the same room, the idea that comes from a stranger in the lift, and the sense of belonging to something larger than one’s own task. Those are the reasons people will make the journey.

When AI agents can do most of the work online, the workplace must do more to connect us.

Related service · Vision & Brief →

The home is being asked to do more, with less, for households that change faster than buildings.

At different hours, the home is now an office, a classroom, a gym, a clinic and a place of care, as well as a place to sleep, eat and gather. At the same time it is getting smaller, more expensive and home to fewer people.

My argument: the answer is not a cleverer flat, full of moving walls and robots. It is a better relationship between the private home and the shared building around it—adaptable cells within a generous, shared hive. I call this the living hive.

Why now

Five forces are changing what a home must do in the cities of Asia and the Middle East.

36%

of South Korean households are a single person—now the largest type. Nearly half of those living alone often or sometimes feel lonely.

1.3 days

a week of working from home, settled as a permanent part-time use most homes were never designed for.

8 m²

Hong Kong’s new legal minimum for subdivided flats. The question is no longer only how many homes, but how well they serve people for decades.

1.6 → 5.6 bn

air conditioners by 2050. In hot, humid cities, comfort is the largest use of household energy—and a growing cost.

AI

and robotics are entering the home. The technology is ready. What it should do—and never do—is still undecided.

The Living Hive — an isometric flat at 8 pmFLEXhome+, Hong Kong. Design: Architecture Georges Hung. Image: Georges Hung.

What the flexible-home consensus gets wrong

The post-pandemic home was imagined as transformable, modular, smart and sustainable. Each idea has merit. Each carries an assumption the evidence does not support.

• Flexibility is not the same as moving parts. Schneider and Till’s century-long study of flexible housing finds “soft” flexibility—generous, loosely defined space—serves people better over time than “hard” mechanisms, which are costly, need maintenance and are often left in one position. Adaptability needs slack.
• Modular is not the same as adaptable. Kurokawa’s Nakagin Capsule Tower was designed so its 140 capsules could be replaced. In fifty years, not one was; it was demolished in 2022. What matters is what can change after completion: services, partitions, connections and ownership rules.
• Smart is not the same as wise. A home that analyses its residents and rearranges itself raises hard questions: who owns the data about when you sleep, and what happens when the motors fail? People want homes that respond to them, not homes that decide for them. And in humid climates, comfort is decided by orientation, shade and ventilation before any product.

A home that fits perfectly today has no room to change tomorrow.

Flexibility, properly understood

“Flexibility” has become as loose a word in housing as “experience” in retail or “innovation” in the workplace. To design for it, ask: flexible over what period, and for whom? Change happens on four timescales.

• DailyWork, study, rest, cooking, play, guests—rooms that serve more than one use, light furniture, storage that clears surfaces, a door that closes.
• SeasonalClimate, holidays, Ramadan, visiting family—shaded outdoor rooms, spaces that open and close, shared guest rooms in the building.
• Life-stageChildren, ageing parents, illness—rooms that divide or join, step-free routes, a bathroom that can adapt.
• GenerationalHousehold size, ownership, use of the building—units that merge or split, accessible services, structure that allows conversion.

Soft flexibility first; hard flexibility only where space is extremely scarce and the mechanism is simple, robust and works by hand. Two tests: Can the home change without a contractor? Will it still work if the moving parts stop?

Defining the living hive

A home made of adaptable cells within a generous shared building: one that changes with its household by the day and by the decade, supports health and care, connects residents to one another, and uses carbon, energy, water and data with restraint.

1. AdaptabilityCan the home change for daily, seasonal, life-stage and generational needs, mostly without building work?
2. WellbeingDoes every habitable room have daylight, fresh air, acoustic privacy and thermal comfort, by design first?
3. SharingDoes the building provide shared rooms that make each home more generous and give residents reasons to meet?
4. CareCan the home support children, ageing, illness and caregivers, including ageing in place?
5. StewardshipAre carbon, water, maintenance and residents’ data managed for the long term, with residents in control?

Six shifts

1. From fixed rooms to adaptable cells. Names fix uses. Design well-proportioned rooms with no fixed purpose around a fixed wet core, with partitions that move without touching structure or services. A cell that is a nursery today can be a study in ten years and a carer’s room in thirty.
2. From private flat to shared hive. Guest rooms, workrooms, large kitchens, laundries and gardens are used occasionally. When the building provides them, private homes can be smaller without being poorer—and a shared kitchen or roof garden gives people living alone an easy reason to meet.
3. From home office to home of many uses. Work, study, care and recovery each need a quiet, daylit place with a door, not a corner of the dining table. In compact housing, a bookable workroom on each floor often serves better than a study squeezed into every flat.
4. From automated home to assisted home. The most useful technology works in the background: balancing energy and cooling, monitoring air, detecting leaks, and alerting family or carers to a fall. Robotic furniture must fail safely and work by hand; data stays at home, shared only with consent. The resident decides.
5. From cooled box to healthy, low-carbon home. Comfort begins with orientation, shade, cross-ventilation and the ratio of glass to wall—then systems, with radiant cooling paired with dehumidification in humid cities. Fit-outs are replaced every ten to fifteen years: choose durable, low-emission materials, designed to come apart and be reused.
6. From housing unit to home for a lifetime. Children arrive and leave, parents move in, people live alone for longer and grow old at home. Neighbouring units that can merge or split, a room that becomes a self-contained suite, step-free routes and adaptable bathrooms let people stay where they belong.

The framework

The five tests become indicators set at brief, checked at design review and measured after residents move in. Targets vary with climate, culture and tenure; the commitment to measure does not.

• AdaptabilityRooms suitable for two or more uses; partitions movable without structural work; units able to merge or split.
• WellbeingDaylight, fresh air, acoustic and thermal targets met in every habitable room.
• SharingShared floor area per resident; bookable shared rooms; share of residents who know a neighbour by name.
• CareStep-free routes; adaptable bathroom and bedroom; provision for carers and ageing in place.
• StewardshipCooling energy per m²; fit-out embodied carbon; water reuse; local data processing and manual override.

What this means

A better hive

For most of the last century, the home was designed as a fixed set of rooms for a fixed kind of family. That family is no longer typical. The future home will not meet today’s demands by becoming a smarter box. It will meet them by becoming part of a better hive: adaptable cells that change with the household, a shared building that makes small homes generous and neighbours visible, and technology that serves quietly, with people in control.

Beyond smart — A home that lives, adapts, and grows with its inhabitants.

Related service · Vision & Brief →