I went to New York following creativity, design and emerging technology wherever they intersected, with a fellowship folded quietly into the trip rather than the trip built around it. The goal, as much as I had one, was simple: attend as much as I could. Find out as much as I could. Meet as many people as I could. Write. Think. Walk. Process. Come home, and tell whoever wants to listen.
What I came home with wasn't a verdict on AI. It was the sense that this isn't a conversation any one country is having ahead of the rest of us. It's global, and everyone I met — in New York, and later in Melbourne — was standing at roughly the same point in it. All of us, in a sense, paddling toward the same unfamiliar shore, working out together how to live alongside a new kind of intelligence. Which is the part I still find strange: the pace of the technology hasn't pulled people apart. If anything, it's pulling us closer together.
For fifteen nights I moved through a compressed, overlapping version of the city: a lecture theatre at NYUNew York UniversityA private research university based in Manhattan., a stage / co-working floor at ZeroSpaceZeroSpaceA production and creative studio in Brooklyn specialising in digital, virtual and live content — and a regular host for New York's creative AI community., a café crowded with laptops for an AI TinkerersAI TinkerersA global community of builders meeting locally to demo and discuss applied AI. build sprint, rows of seats and a lit stage at BrainStationBrainStationA global tech and design education company running courses across creative and digital skills., a critique studio at Parsons School of DesignParsons School of DesignA design and art school in New York City, part of The New School., a gallery on the Bowery at the New MuseumNew MuseumA contemporary art museum on the Bowery in Manhattan, dedicated to new art and ideas..
I'd gone in expecting to come home with a list of tools. What I came home with was better, and stranger: the same question, already waiting for me in Melbourne before I'd even landed.
Different city.
Different people.
Different ecosystem.
The same conversation.
Not, "Can AI do this?" I'd stopped hearing that question well before I left New York, and Melbourne, it turned out, had moved past it too. What I heard instead, in both cities, in a dozen different accents, was something harder: not what a tool could do, but how to build with it well — and what it should actually change about the way we work.
I stopped being surprised by how different the two cities were. I started being surprised by how alike they sounded.
Before the first speaker had even taken the stage at BrainStationBrainStationA global tech and design education company running courses across creative and digital skills., the room was already buzzing. People weren't sitting quietly, waiting for it to start. They were introducing themselves to complete strangers, swapping ideas, comparing what they were building, asking what had brought each other there. No rows of people hidden behind laptops. It felt less like an audience assembling and more like a community that had simply found the same room again.
I found myself talking to someone from Eastern Europe who'd spent the previous two weeks immersing himself in New York's AI and startup scene. Minutes later I was talking to a Yale computer science student who led the university's AI student group. Within minutes, the three of us were somewhere deep in creativity, education, emerging technology, and where all of it might be heading.
It wasn't networking, not in the way I understood the word. Nobody was selling anything. People were just curious — asking real questions, actually listening to the answers. Looking back, that conversation in the foyer set the tone for everything that followed. At BrainStation, at NYU, at Parsons, at the AI Tinkerers sprint, at ZeroSpace, at Columbia Startup Lab talking to Alon Grinshpoon about echo3D, at the New Museum, in one conversation after another with founders, designers, educators and artists — the technology kept changing shape, but the questions people were asking stayed remarkably similar.
Once the session properly began, nobody reached for a laptop. There was nothing to open, nothing to type into — just sharp brains and open ears, a room full of people listening the way you listen when you're trying to catch something before it disappears. I'd spent the last couple of years, like most people, treating AI as something you talked to — better prompts, sharper instructions, one tool at a time. This felt like the conversation had moved on without telling me.
Nabeel Barqawi, a product leader at Disney, and Xavier Rumph, a design strategist, put words to it before I could — and challenging us to think beyond individual tools felt less like a new idea than a continuation of the conversation that had already started in the foyer, before the event had even officially begun. You don't need a computer science degree to build with this, they said. You need to think in systems. Not what can this tool do, but what problem are we actually solving, and what's the simplest system that solves it?
I sketched it on the back of my notebook as they talked: VS CodeVisual Studio CodeA free source-code editor made by Microsoft, widely used for software and web development. as the workspace, Claude Code as the collaborator, GitHubGitHubA platform for hosting and collaborating on code, owned by Microsoft. managing the project, VercelVercelA cloud platform for deploying and hosting web applications and front-end projects. shipping it — nothing added unless it earned its place. Build something small. Solve one real problem. Iterate.
A few days earlier, in a very different kind of room — a panel table, NYU's Creativity Leap — I'd heard what turned out to be the other half of the same idea, though it took me a while to notice they were the same idea. Not because I can code, either; I can't, not really. But as the technical barrier drops, the question at BrainStation about what to build starts to matter less than this one: what's still worth building, once the building gets easy?
Dr Natalie Nixon, Daisy Auger-Domínguez, Dag Kittlaus and Devon Turnbull kept circling back to a single concept: the future belongs to the imaginative, the restored and the creative. Not making more of anything — noticing better. Knowing which idea earns the extra hour, and which one doesn't.
Someone on the panel called it flourishing, which sounded strange in a room usually reserved for the word productivity. Prototyping, someone else said, is the antidote to polish — perfectionism is what keeps most ideas from ever leaving the page. If AI does anything worth celebrating, it's making the ugly first draft cheap enough to actually attempt.
Not to become more productive. But to become more creative, more thoughtful and, ultimately, more human.
If one idea trailed me around New York like a shadow, it was this: the more capable the technology gets, the more your judgement is worth. Not taste in the design-school sense — colour, type, style — but taste as discernment. Knowing when an idea is finished. Knowing when it needs another pass. Knowing when it shouldn't exist at all.
Someone at NYU offered an example I still haven't shaken. Ask AI to design a house and within twenty minutes you'll have floor plans, elevations, a construction set that looks entirely convincing. But if the person asking never learned structural principles, materials, the physics that keeps a roof standing, that house may never actually stand. The technology can produce an answer. It has no idea whether the answer is any good.
I thought about that house again the next day, walking through Parsons School of DesignParsons School of DesignA design and art school in New York City, part of The New School. — half-finished garments on the racks, "Knicks in 5" written diagonally on the whiteboard. Design was never really about the outcome. It's the mess before it: the wrong problem, defined three times before the right one surfaces, the failed prototype, the critique that stings. AI can speed up parts of that process. It can't sit through the critique for you.
As an educator, that's my job now, made explicit rather than implied. If AI can produce a polished outcome in minutes, mine shifts too — from assessing what a student made to understanding how they got there. What did they reject, and why? What did the feedback actually change? Critical thinking stops being one unit in the curriculum and becomes the whole of it.
Some of what surprised me most in both cities had almost nothing to do with the technology. It was the room itself. One morning I joined an AI TinkerersAI TinkerersA global community of builders meeting locally to demo and discuss applied AI. build sprint — a café, a few dozen laptops, one simple rule: say what you're building, spend a few hours building it, then show the room what you learned, finished or not. Nobody performed. Nobody pitched. It felt closer to a study group than a startup event.
A little over a week later, in Melbourne, I found the same room wearing a different accent. People compared models instead of showing off finished products. They troubleshot each other's problems out loud, argued gently about which workflow was actually faster, admitted when something hadn't worked. Nobody there had met the New Yorkers I'd just spent two weeks with. They were having the identical conversation anyway.
I'd braced myself for the version of AI that isolates people behind screens. What I found instead, in both rooms, was the opposite: designers learning from developers, educators from founders, engineers from artists, everyone admitting out loud what they didn't know yet.
Another evening, another room — ZeroSpaceZeroSpaceA production and creative studio in Brooklyn specialising in digital, virtual and live content — and a regular host for New York's creative AI community., a creative studio in Brooklyn hosting NYC Creative AI. What struck me wasn't a tool or a launch. It was the vocabulary. Nobody was talking about prompting a single model anymore. They were talking about ecosystems — different tools, different models, different workflows, wired together and talking to each other.
Terms like MCP and CLI came up all night, not because the room had suddenly filled with engineers, but because creative practitioners are starting to think in terms of interoperability. Matt Miller walked through MCP integrations inside ComfyUI, and the idea underneath the demo landed harder than the demo itself: stop treating AI like a chatbot you visit. Build a workflow where specialised tools hand context to each other, automate the repetitive parts, and free you up to spend your attention on the decisions that actually need a human.
Earlier that same evening, a different speaker had walked the room through something smaller, and just as sticky: a repeatable method for reverse-engineering why a paid social ad works. Pull the winning ads from the Meta Ad Library and TikTok Creative Center, ignore the likes, sort by how long they've been running — old and still live means it's converting. Hand the winner to Gemini with a decode prompt: name the hook in the first three seconds, map the structure as an arrow chain, describe the visual grammar, name the emotional driver, write the one-line thesis for why it works. It was the same instinct I'd been tripping over all week, aimed at advertising instead of art or code: don't copy the surface. Understand the structure underneath it. Then build your own version.
Coming from design and education, I found this genuinely exciting rather than alarming. It suggests a future where creative people don't need to read every line of code, but do need to understand how the pieces fit together. The advantage isn't knowing one tool well anymore. It's knowing how to orchestrate a handful of them towards something that actually matters.
One afternoon took me somewhere quieter: Columbia Startup LabColumbia Startup LabColumbia University's co-working space in SoHo for recent-alumni founders., just off the university in SoHo, home to a rotating cast of early-stage teams. Among them, echo3DAlon GrinshpoonCEO & co-founder of echo3D, a 3D/AR digital asset management platform, and a Techstars mentor., a 3D and AR asset platform led by CEO Alon Grinshpoon.
Alon started the company seven years ago after building 3D interfaces for operating theatres — watching a patient's heart, reconstructed from CT and MRI scans, float above the surgeon mid-procedure. The bottleneck turned out to be embarrassingly mundane: someone had to physically be in the room at five in the morning to swap files between cases. There wasn't a product for that, so he built one.
The surgical work itself had come out of Columbia, where he did his master's in computer science — patented research, built with a small team of engineers and clinicians chasing a genuinely unglamorous problem: getting a scan into a surgeon's field of view without anyone having to look away from the patient. Before that, an electrical engineering and computer science degree at Tel Aviv University, then a stint as a software engineer at NVIDIA. He's a Techstars mentor now, one of those people who ended up teaching the thing he'd only just finished learning himself. Seven years on, the client list has grown a long way past hospitals — Chevron, Energizer, Chanel, Lowe's — sitting alongside the design students on his free education tier, using the same platform for entirely different reasons.
What stayed with me was something small he did mid-sentence, without making a thing of it. He pulled up a 3D model, exported its texture, dropped it into an AI image tool, asked it to turn a sand-coloured surface into brick, uploaded the result straight back into the platform as a new version. No trip to Adobe. No hours lost. The tool did the labour. The only thing that actually required Alon was the judgement — what to try, and whether it had worked.
"We're literally limited only by imagination."
The afternoon that slowed everything down
By then I'd spent the better part of two weeks in rooms full of screens. Architectures, workflows, agents, systems. Smart, generous, exhausting conversations. And then, on an ordinary afternoon with nothing scheduled around it, I met a painter, and everything else went quiet for a while.
His name is Bobby Hill. On the surface, our conversation had nothing to do with AI. We didn't mention it once, as far as I remember. But looking back, everything I'd been hearing all week — about systems, judgement, taste, attention — was sitting quietly inside what he told me that afternoon, and I hadn't noticed until much later.
He told me he asks himself questions before he falls asleep, and that the answers tend to arrive as dreams — the trick, he said, is writing them down before they slip away. His inspiration comes from being alive, paying attention, and he said even his own process keeps changing. He's learned, over the years, to trust his instincts sooner than he used to.
He's been in that studio since 2008, but the city in him goes back much further. Bobby grew up in Harlem, went to school in the Village, and has watched New York reshape itself around him for decades. He remembers 1982, '83 — Madonna, the punk look, crucifixes everywhere downtown. A different, wilder New York, he said, than the one either of us was standing in that afternoon.
His process is deliberately physical, which felt pointed after everything else I'd seen that week. Canvas. An old-school projector to block in the composition. Primer, before a single line goes down. Then house paint, layer over layer, built up and scrubbed back with a brush until the surface finally says what he needs it to say.
Trusting the process came up again and again. His paintings rarely finish where they started. One series that began in anger, after the murder of George Floyd, slowly became a tender image of a father and a child. Once, he left a set of unfinished pieces on a roof and forgot about them — literally forgot — for years, before rediscovering them and finishing them with fresh eyes. It changed everything about the pieces. He doesn't force an outcome. He lets the work tell him where it's going.
He's intensely focused when he's working — almost obsessive, he admitted — but insisted the focus has to be earned first. Before he can create, he has to fill his own cup — stay open, keep moving, get outside. The answers, he said, usually arrive through other people: an overheard sentence, a sign you'd miss if you weren't paying attention. While I was flipping through a stack of his prints, he happened to be talking about Thailand. Without thinking, I reached for the one Thai-inspired piece in the pile. He just smiled. Synchronicities, he said, are everywhere, if you let them be.
Our own meeting was one of those. It was my second-last day in New York, back in January, and I'd gone looking for something real from an actual New York artist. I crossed the Brooklyn Bridge into Dumbo and found Bobby at a street stall — his first in ages, he told me. We talked for a long time and stayed in touch. When we met again this July, he still hadn't set up another stall since that day.
Slow down. Pay attention. Talk to people. Listen. Follow your curiosity. Trust the process. Remain open to synchronicity.
Being a real creative, he told me, can be lonely — but that solitude is necessary. It's what lets you fully disappear into the work, reach the state he calls flow. He sees it as healing, too. He believes everything happens for a reason, and that his own hardships fed directly into the most focused, most productive period of his life. The hardships, in his telling, were a gift. Without them, he doesn't think he'd have made nearly as much. And if you're stuck? His advice was almost embarrassingly simple. Just start. Make mistakes. Start again.
I keep coming back to that afternoon. In a world where AI can generate an idea in seconds, maybe the responsibility isn't producing more of them. It's building the patience, the attention, the discernment to know which ideas are actually worth chasing. Bobby never once mentioned artificial intelligence. He didn't need to. He'd already answered the question everyone else in New York was still circling.
One of the most striking hours of the whole trip had nothing to do with a conference at all. It was an afternoon at the New MuseumNew MuseumA contemporary art museum on the Bowery in Manhattan, dedicated to new art and ideas., wandering New Humans: Memories of the Future. Walking through it, I kept recognising the same question in different clothes. Nobody in that room was asking what AI could do. Every wall text, every looping video, was really asking what kind of future we're actually building.
The work sat in the blurred space between human and machine — identity, embodiment, synthetic creativity, what stays uniquely ours in an age of intelligent systems. Some pieces imagined futures full of possibility. Others were more unsettled, asking harder questions about the social and ethical weight of what we're building. What happens to memory when a machine can generate endless new content? What happens to identity when creativity itself becomes computational? What does authorship even mean, once ideas are made together with something that isn't a person?
None of those questions had tidy answers on the wall text. Nobody in the room seemed to expect them to. I stood in front of one canvas for a long time, a lecture theatre and a workshop room and a critique studio still sitting somewhere in the back of my head, and thought: this is the same question, wearing paint instead of a slide deck.
Home, with the same question
A little over a week after landing back in Melbourne, I found myself in an industrial co-working space — exposed pipes, ring pendant lights, the word Inspire painted across the brick behind the stage — a full house of people who use Claude, there to compare notes rather than watch a keynote. It was the room I'd opened with, though I didn't know that yet when I walked in.
The format echoed New York almost exactly: practical, hands-on, no polish for the sake of polish. One speaker walked the room through a four-step setup for prototyping in code — Claude for the vibe-coding, VS Code for editing, GitHub for managing, Vercel for deploying. Another dug into agents and system prompts, and a plugin called FAT Agent — Fix, Audit, Test — that acts like a post-launch QA engineer: auditing a live site for SEO, performance, security, accessibility and content, then walking you through fixing each one.
What struck me hardest was hearing the same questions come back, almost word for word — how do we stay creative, how do we build with this well, what's actually worth automating and what isn't — asked by people who'd never been in a room with any of the New Yorkers I'd just spent two weeks with. Different accent. Same conversation. It was the clearest proof yet that this isn't one city figuring something out ahead of another. It's one long, unfinished, genuinely global conversation, and I'd just watched two of its rooms in real time.
Sitting there, I realised it wasn't a second event. It was another chapter of the same one. What's stayed with me isn't the technology. It's how consistent the ideas were, wherever I happened to be standing. Not: can AI replace us? But: how do we build responsibly? How do we stay creative? How do we design something that actually makes someone's life better?
It wasn't the only echo waiting for me. A few days later, back at school for the start of Term 3, I sat in on a session with Dr Tim Kitchen — an education consultant who's spent years thinking about creativity and generative AI in the classroom. Listening to him through the lens of everything I'd just seen in New York, it felt less like a professional learning workshop and more like the next chapter of the same conversation: education has to keep human thinking, creativity and judgement at its centre, whatever the tools get better at doing for us.
I think about Bobby's three words more than I think about any panel, any workshop, any system diagram. He said them about paint. It turned out to be true of everything else I saw in New York and Melbourne too — he'd just worked it out a few months before anyone else got around to it.
Epilogue: somewhere over the Pacific
Of all the people I could have shared a seventeen-hour flight home with, I happened to sit beside a retired lecturer who had spent her career teaching art, design and creativity. She also had two sons living in New York. It felt like one final conversation the city had arranged before I came home.
Somewhere over the Pacific, I found myself telling her everything — the AI events, the designers, the educators, the artists, the founders, two weeks' worth of conversations that had filled my notebook. She listened quietly, the way people do when they're actually listening.
When I finished, she smiled and offered a piece of advice that stopped me in my tracks.
"Slow down."
Then she added something I don't think I'll ever forget.
"Don't lose yourself. Don't lose what it is to be human."
It was such a simple observation. Yet somewhere over the Pacific Ocean, it felt like the final piece of the puzzle — the thing every room in New York and Melbourne had been circling without quite landing on it.
At BrainStation, people spoke about systems thinking. At NYU, they spoke about imagination and human flourishing. At Parsons, they spoke about process and critical thinking. At ZeroSpace, they spoke about creative ecosystems. Tim reminded us that education has to keep human thinking, creativity and judgement at its centre. Bobby spoke about curiosity, conversation and the art of paying attention. And now, a stranger I'd met only hours earlier had distilled all of it into two quiet sentences.
Slow down.
Don't lose what it is to be human.
Looking back, that's what the trip was actually about. It was never really a study tour of artificial intelligence. It was an exploration of creativity. Of education. Of design. Of community. Of paying attention. And perhaps that's the most encouraging thing I brought home with me: across two cities separated by more than sixteen thousand kilometres, I kept encountering the same conversation. Not about machines. About people.
