One caveat before we jump in: I’m working with an AI on these pieces now (shock, horror). You probably know this already. I’ve never claimed to be much of a writer. I am, however, a deep thinker, and I recognise patterns (I can’t turn it off; thanks, ADHD). So I work with Claude (and others) to help me research, and to structure and challenge my thinking. I see it as a writing partner for a skill I feel weaker in, and I’ve learnt a lot from doing it. The most important thing for me is this: the ideas start with me, and the point of view is always my own.

Let’s jump in…

Think about the last big fragrance launch you saw. A famous face, golden light, a slow-motion turn of the head, a billboard on Sheikh Zayed Road. It’s about as pure a piece of promise-making as marketing gets: a feeling, a version of yourself you’re meant to want. I’m not above it. That kind of thing works on me, and it works on most of us.

Now picture a different shopper. She opens ChatGPT and types: which perfume has the best sillage in Dubai heat? I want people to notice it when I walk past.

That’s the shift I keep coming back to )please see below for proof of real human thinking about this). For most of history you sold to one kind of buyer, a person, and the whole machinery of advertising was built to make that person feel something. Now there’s a second buyer, one that reads everything, feels nothing, and more and more often decides what the human sees first.

So, who are you selling to?

The honest answer for almost everyone is: both. And they want completely different things.

^{My crazy doodling trying to figure out who is selling to who, and how}

The promise and the proof

Let me put it as plainly as I can, because everything else in this piece hangs off it.

People buy the promise. Who will I be if I own this? How will it make me feel? Will people like me, or people I’d like to be, see it on me? Status, desire, belonging, identity, and mostly at a subconscious level. Brands spend fortunes on this because it works.

Machines buy the proof. Not proof in a legal sense, but whatever they read as evidence: ratings, review volume, consistent facts, independent mentions, clear specs, a price that’s actually the price. An AI assistant has no ego to flatter and nothing to signal to anyone. It wants to give a good answer and not get it wrong.

We now have some research on what moves these agents, and it’s quite revealing. A team of researchers built a test marketplace and watched how AI shopping agents from the big labs chose between products (the paper is called What Is Your AI Agent Buying?). A few findings stood out to me:

• Labelling something “sponsored” made agents less likely to pick it. Every model did this. The paid badge that works on a scrolling human works slightly against you with a machine.

• A platform’s “Overall Pick” badge did the opposite, and by a lot. A product picked 10% of the time jumped to somewhere between 20% and 43%, depending on the model. Third-party endorsement is gold.

• A tenth of a star mattered. Raising a product’s rating by just 0.1 took it from 10% to as much as 20% of picks. Review count mattered too.

• Rewriting the product description helped, in most models. When the researchers let a seller-side AI tweak descriptions to better answer the shopper’s question, the product’s share went up by as much as 15 points.

• The machines don’t agree with each other, and they change their minds. In one category, a fitness tracker was chosen 77% of the time by one model and 6% by another. A model update can reshuffle the whole picture.

So what does this tell us? The machine isn’t a neutral judge. It has habits of its own, and they can change every time the model is updated. But look at what those habits reward: someone else vouching for you, a strong rating, plenty of reviews, a description that actually answers the question. Then look at what they punish, which is the thing you paid to promote. With people, money can buy you attention. With a machine, it mostly can’t. What gets you picked is what other people say about you, and how clearly you’ve explained what you do.

Why the promise worked, and why it’s wobbling

There’s a reason the glossy campaign worked for so long, and Oliver Feldwick explains it brilliantly in “Cheap Talk and Costly Signalling“, from his Substack Scraps, Straws and Silicon. His argument, in short: a lot of marketing has worked as a costly signal. We trust a message more when it would have been expensive to fake.

You can see this all over Dubai. In 2023, the number plate P7 sold for AED 55 million at a charity auction. Nobody needs a two-character plate to get to the mall. It means something because almost nobody could afford it. If anyone could have one, it would say nothing at all.

Advertising worked the same way. A huge campaign didn’t just tell you about the perfume. It told you the brand was big enough, and confident enough, to spend that kind of money. The celebrity face, the Paris show, fireworks over the Burj. Money you could see, standing in for quality you couldn’t.

AI is squeezing this from both ends.

It’s now cheap to look expensive. A campaign that looks like it cost millions can be made in an afternoon, so looking expensive doesn’t tell you much any more. Creators are feeling this too. When anyone can make a beautiful carousel in five minutes, a beautiful carousel no longer proves you put the work in.

And the machine doesn’t care about that type of signalling. People used the size of a campaign as a shortcut for “this must be good”. An AI agent doesn’t need shortcuts. It reads the reviews, the return complaints and the price, so the money spent on looking impressive barely registers.

So the promise is being squeezed by the people making it and by the machines reading it. That doesn’t make the fancy signalling extinct, It means the kind of effort that convinces people is changing, and I’ll come back to what that looks like at the end.

Five ways selling works now

When I first mapped this, I came up with five different ways buying and selling now happen. Each one has a different buyer, and each buyer responds to something different.

1. Human to human: still buys the promise

The market stall never went away. People buy from people, and they want connection, emotion and a sense of belonging.

You might expect AI to make this matter less. It’s doing the opposite. In a Gartner survey published in September, 43% of US consumers said they now rely more on real people for shopping advice because there’s so much AI content around. 57% said AI-made brand content had made them trust brand messaging less. And 35% said they rely on influencers less.

I don’t read that last number as people giving up on creators. I read it as people giving up on content that feels like it was made by nobody. When anything can be made in seconds, the thing you can’t fake is a person who keeps showing up.

Communities show this more clearly than anything. When a fan spends weeks making something for a brand or a person they love, that effort is proof no budget can buy. It’s also why audiences turn on AI-made work so fast when they feel misled. They weren’t only buying the thing. They were buying the care that went into it.

2. Humans to machines: getting found and getting cited

If people ask an AI what to buy, you have to be visible to the AI first. The industry calls this GEO or AEO, generative or answer engine optimisation. Think of it as SEO where the reader is a machine that summarises, rather than a search engine that lists.

Bain and the Comité Colbert found in June that around 70% of luxury shopping prompts don’t name a brand at the start. People ask “what’s a good first fine-jewellery piece?”, not “Cartier”. And 90% of the links AI tools cited came from outside the brands’ own websites.

Read that again. Nine in ten. What the machine believes about you is mostly what other people have said about you.

As someone who came up through PR and publishing, this is the bit that made me sit up. A feature in a respected title, a detailed review, an expert quote, a long Reddit thread. These used to be about the humans who read them. Now they’re also what the machines read and repeat to millions of people asking for advice. Press has become reputation a machine can read.

3. Machine to human: the AI as salesperson

This one’s already big. Amazon says its shopping assistant Rufus was used by 300 million customers in 2025 and added close to $12 billion in sales that wouldn’t otherwise have happened.

And the people AI sends your way are good customers. Adobe’s July figures show visits from AI tools now convert 60% better than other traffic. Shopify said in its Q2 results that orders coming from AI search had tripled on the year before, and that AI channels were bringing in new buyers at about twice the rate of Google, social or direct.

A year ago, AI traffic converted worse than everything else. That flipped in about twelve months.

4. Machine to machine: buys certainty

What happens when one AI agent buys from another, with no human watching in the moment? The pipes for this are being laid fast: payment standards from OpenAI and Stripe, from Google and Shopify, and a “know your agent” framework from Visa, Mastercard and Ant International that’s basically ID checks for bots.

It’s early. OpenAI quietly retired its in-chat checkout in March after very few merchants went live. The rails are ahead of the trains, which is normal.

But it tells you what machines want from a seller, and none of it is aesthetic:

• Certainty. Accurate prices, live stock, clear delivery times.

• Readability. Product information laid out so a machine can use it.

• Verification. Proof you are who you say you are.

• Accountability. A clear answer to who fixes it if it goes wrong.

My bet is that reputation becomes a kind of currency between agents. Something like a credit score for whether a business does what it says.

5. Humans to humans, through agents: the doorman

This is the one I think most brands are missing, and it matters most for luxury, wellness and personal brands.

It isn’t really robots selling. It’s still a person buying from a person or a brand they love. But there’s now an agent standing in the doorway.

The agent picks the shortlist. The human picks the winner.

You can see it already. Adyen found the share of US shoppers using AI assistants nearly tripled in a year, from 12% to 35%. Bain found 82% of top-spending luxury customers used an AI tool before their latest purchase, and 47% of people who bought in store had used AI before they walked in.

The Gulf is further along than most. In Checkout.com’s survey this spring, 79% of UAE shoppers said they’re comfortable letting an AI agent complete a purchase for them, and 71% would let it switch brands if it found better value. That second number should worry any brand that’s relying on loyalty alone.

So you have to win twice. You win the agent with proof, and you win the human with the promise. Fail the first and you never make the shortlist. Fail the second and you make the shortlist and still lose.

There’s one way past the doorman, and Oliver makes this point well: be the name. If someone tells their agent “get me a scented candle”, the agent goes off and compares forty of them. If they say “get me a Diptyque candle”, it doesn’t compare anything. It just goes and gets it. The brands and people who get named, rather than described by their category, skip the shortlist altogether.

Who does the agent work for?

Here’s where it gets uncomfortable, and where I think the next few years get decided.

The optimistic version of all this is lovely. Everyone gets a personal shopper who has read every review ever written and has no commission to earn. Something that quickly finds the thing that’s actually good, rather than the thing that shouted loudest.

But that only works if the agent is really on your side. And the company that builds the agent gets a say in what goes on the shelf.

You can already see the fight. Amazon started letting brands pay for sponsored prompts inside Rufus last November, so the friendly assistant is now partly an ad with a nice voice. In September, Amazon blocked Meta’s new shopping agent, Muse, from its site, saying it hadn’t identified itself as a bot. It has been in a legal fight with Perplexity over the same thing. Nobody wants to be the shelf another company’s robot shops from.

People can sense this. In Adyen’s research, 41% of shoppers said they want to know why an AI picked one product over another. I think “whose side is your agent on?” is going to become one of the defining consumer questions of the next decade. I’d like it to be answered properly, in the open, rather than buried in terms and conditions.

Good news for small brands?

Maybe. I’m cautiously optimistic.

If machines lean on evidence rather than budget, a small brand with a better product and real reviews has a better chance than it’s ever had. Analysis from Peec AI found that one indie skincare brand, Kaja Beauty, picked up 4.2% of all skincare citations across the big AI platforms. That’s almost level with Wikipedia. A small brand can win this.

Better products beating bigger marketing budgets. I’d love that world.

Two catches, though.

Proof can be faked. If machines reward reviews, ratings and “independent” mentions, people will manufacture all three. Fake reviews, paid placements dressed up as editorial, invented test results. The machines will have to get better at telling real proof from staged proof, and some of the early ones won’t be.

There’s no single top spot to chase. Because different AIs prefer different things, and change their minds with every update, there’s no trick that holds. The only strategy that survives is real proof, built over time, in lots of places.

Delegated or desired?

Underneath all of this, I think buying is splitting in two.

Delegated purchases are things you don’t want to think about. Washing-up liquid, printer ink, the vitamins you reorder every month. Agents will take these over, and fast. The brand that wins is the one the agent trusts and finds easy to buy from.

Desired purchases are about who you are, or who you want to be. The bag, the ring, the hotel, the holiday. Nobody wants a bot to buy their first Birkin. The wanting is the point.

Luxury and personal brands live almost entirely in the second group. But even there, the agent decides whether you make it onto the list.

Oliver has a really useful way to sort this, with two questions. Can a machine check whether it’s any good? And does anyone see you choose it?

• Checkable, and nobody sees (a phone charger, a water filter): the agent decides. Your job is to make the shortlist.

• Checkable, but people see your choice (sunglasses, a gym bag): the agent filters, the human picks. Being recognisable still counts.

• Hard to check, and private (a coach, a consultant): what you’re willing to put on the line does the work.

• Seen, and decided by taste (a handbag, a fragrance): this is where the big, visible signal still holds up best.

From showing effort to showing stake

Oliver’s answer to all this, and I think he’s right, is that trust is shifting from proof of effort to proof of stake. The question is no longer how hard you worked. It’s what you stand to lose if you’re wrong.

Guarantees are the clearest example. Patagonia has long promised to repair, replace or refund anything that doesn’t perform. That would be ruinous if the products were poor, which is exactly why people believe it. A coach who says “if you don’t get results, you don’t pay” tells you more than any glossy sales page.

Making is cheap now. These are the things that still cost something, so they’re the things that still convince people:

1. Put something on the line. Make claims specific enough to be wrong. Offer guarantees. Put your name on the work. A promise only counts if breaking it would hurt.

2. Be in the room. Events, dinners, live sessions, showing the thing working in front of people. Time and presence can’t be generated. (It’s a big part of why I’m launching in person events again soon.)

3. Show your workings. How you made it, the decisions you took and why, and where AI helped. Being upfront about how you use AI lands far better than being found out.

4. Edit. Anyone can produce ten thousand words now. Choosing the three hundred that matter is the skill.

5. Build a track record. Years of turning up and doing what you said you’d do. It’s the one thing you can’t shortcut.

And for the machines specifically:

6. Ask them about yourself. Once a month, ask ChatGPT, Claude, Gemini and Perplexity the questions your customers would ask. Not “tell me about [your name]”, but “who’s the best [what you do] in [your city]?” Notice who comes up and what sources they cite. That’s your new shelf.

7. Make your proof easy to read. Results, reviews, credentials, prices, what you actually do, said clearly and the same way everywhere. If a machine has to guess about you, it’ll pick someone it doesn’t have to guess about.

8. Get talked about where machines listen. Press, podcasts, YouTube, Reddit, LinkedIn, guest essays, reviews. Other people’s words about you count for far more than your own.

9. Own the relationship. Your email list, your community, your clients. Adyen found 27% of retailers already worry about losing the direct relationship to AI. The doorman can’t stand between you and people who already know your name.

None of this means desire is going to disappear. The proof gets you on the AI agents results, and a the promise, backed by something real, is still what makes someone choose you once you’re there.

So, who are you selling to?

Sources

• Allouah et al., What Is Your AI Agent Buying? Evaluation, Biases, Model Dependence, & Emerging Implications for Agentic E-Commerce — arXiv 2508.02630

• Gartner consumer survey on AI and influencers (1,006 US consumers), 22 Sept 2026 — Gartner

• Bain & Company and Comité Colbert, Luxury and Technology report, 30 June 2026 — Bain

• Peec AI / 5WPR, State of AI Citations 2026 — 5WPR

• Amazon Rufus figures — Modern Retail, 30 Apr 2026

• Adobe AI traffic data, July 2026 — Digital Commerce 360

• Shopify Q2 2026 results, AI orders and new buyers — SaaStr

• Agentic commerce protocols, Sept 2026 update — toomanysecrets.com

• Adyen Retail Report 2026, US — Adyen

• Checkout.com, Agentic Commerce 2026 (UAE figures) — Zawya

• Amazon blocks Meta’s Muse agent, 22 Sept 2026 — VKTR

• Oliver Feldwick, “Cheap Talk and Costly Signalling”, Scraps, Straws and Silicon, 28 Oct 2026 — Substack

• P7 number plate, Most Noble Numbers charity auction, April 2023 — Gulf News

Before the trends, a confession. Today I shipped my first product.

“Shipped” is developer speak. (Well, “to ship” is also Gen Z speak for wanting two people to get together, but for this post we’re using the developer version.) It means you stopped tinkering and put the thing out into the world where people can actually use it. I’ve worked in marketing for fourteen years, and I’ve launched campaigns, magazines, brands and other people’s products, but I have never once built something myself, from nothing, and pressed the button. Today I did, and I’ve been grinning at my laptop like a child on Christmas morning ever since.

It’s called Signals by Social Mode, and it lives at signals.red. It’s a trend radar for the UK and the GCC: the big, slow shifts (and the fast, disposable ones) in how people live and spend, what the major global reports agree on, how differently each trend lands in London and in Dubai, and what a brand should actually do about it. The radar is free. The deeper thinking behind it will live here, on Substack.

I built it in just under a week, with AI as my thinking partner, no developer and almost no budget. A few years ago that sentence would have been impossible to even comprehend. Now we can sit at our dining tables working with AI, shipping things from our imagination out into the big wide web. Which is a big part of why there’s a new trend at the bottom of this list.

A few months ago I made a video about the seven things I thought were about to take over the conversation. It was my first trend-led video, and it’s still the post that’s done the most to push Social Mode forward, so here they are again, properly this time, with the reasoning behind each one. Plus two I’ve added since.

1. The trust economy

People are buying who they believe, not just what works. When every product claims the same results and every ad looks the same, the tiebreaker becomes the person you trust to tell you the truth about it. It’s why creator-led brands keep doing well, why so many businesses now put a founder’s face on the front, and why one honest review can move more product than a whole campaign.

Trust takes years to build. One bad partnership can destroy it. Expect more people to start treating it like the asset it is.

2. The intellectual influencer

Nuance is having a moment. For years the feed rewarded certainty: hot takes, hard rules, one-line answers. Now the accounts I see growing fastest are the ones willing to think out loud, cite something, change their mind, and say “it depends” before explaining what it depends on. Long captions are back, and so are essays. Not everything is ‘dead’ any more, thank God.

3. Cosplaying your job

More people are performing their jobs online, for the algorithm rather than the paycheque. Lawyers, nurses, pilots and bankers are finding that the texture of their work has an audience. Expertise has become content, and the perspective a job gives you is starting to be worth more than the job itself. Don’t tell me you don’t follow your dentist on Instagram. I know you do.

4. Wellness getting grounded

Less biohacking, more basics. After a few years of cold plunges, twelve-step morning routines and a supplement for every hour of the day, the mood is moving back towards sleep, walking, protein, daylight and friends. The expensive, complicated version of wellness is starting to look a little try-hard, and the simple version is starting to look confident.

5. Touch grass. Literally.

Running clubs, supper clubs, walking groups and book clubs keep growing. We’re spending more time online than ever and we’re lonely, and the fix people are choosing is physical and in person. Don’t get me started, though, on brands swapping the word ‘customers’ for ‘community’. Please, it needs to stop. For brands and creators, what matters is whether you can get people to come and meet you IRL, and whether they still want to hang out with each other once you’ve left. That’s a community, not just customers.

6. Muscles over skinny

The body ideal is shifting from small to strong. Strength training has gone mainstream for women in a way that would have felt niche ten years ago, and the language has moved with it: from shrinking to building, from “toned” to strong. It will keep reshaping fashion, food, fitness and the whole wellness category.

7. Mitochondria

It’s about to be a household word. Energy, longevity and cellular health are moving out of scientific papers and into mainstream wellness marketing. I’d bet on seeing “mitochondrial health” on packaging you’d never expect within the next year or two.

8. The AI ick

AI content is everywhere, and so is the ick it gives people. It isn’t really the technology they object to. It’s the pretending: an outfit that doesn’t exist, a “curated” feed nobody curated, a luxury campaign made by a prompt. I’m starting to see brands put “no AI” clauses into creator contracts, and I think “made by humans” becomes something people actively look for and pay more for. Luxury has always meant somebody made it.

(Yes, I see the irony of putting this next to a product I built with AI. I don’t think the two cancel out. The ick, I think, is about value, effort and how you describe what you’ve made, and of course whether it’s good for us and the planet.)

9. The kitchen-table builder

This is the new one, and I added it after building something at my new dining table. Very meta.

The cost of making a real product has collapsed. Not a pitch deck or an idea in a notes app, but a working website, a tool, an app, a small piece of software that does a job. Things that needed a developer, an agency and a five-figure budget two years ago, one person can now build at their kitchen table over a few evenings.

If that keeps going, and I think it will, then by the end of 2027 a lot of us will be building all kinds of things. Not unicorns. Small, useful, specific things, made for a particular community, a particular city, a particular problem the big platforms will never bother to solve.

And here’s the bit I keep turning over. For twenty years the internet pushed everything towards global scale. Go big or don’t bother. Maybe the next wave pushes the other way. Maybe it brings back the mum-and-pop shop, the kitchen-table business, the person who makes a living serving a few thousand people really well instead of chasing a few million. Local again, but with better tools.

I could be wrong. Cheap tools also mean more competition, and building something was never the hard part. Getting people to care about it still is. But for the first time in a long time, the barrier to starting is close to zero, and that changes who gets to have a go.

What links them

Read the list back and most of these point the same way. People are tired of the frictionless, the fake and the over-optimised. They want real people, real rooms and real thinking, and more and more of them want to make something of their own.

That’s what Signals is for: tracking shifts like these, properly, as they move. Have a look at signals.red, click around the radar, and tell me which trend you’re already feeling. And if you’re building something at your own kitchen table, I’d genuinely love to hear what it is.

If you’re here because you commented “Truman Show” on the Reel, welcome. This is the longer version.

At the end of The Truman Show, Truman walks up a staircase painted onto the sky, finds a door and steps out into the dark. The audience in the film cheers, and so do we. After a lifetime on air, Truman has found the one place the cameras don't reach.

I’ve started to think most creators have bricked that door up. And I think that’s where the ick comes from. Let me explain.

Front stage, back stage

In The Presentation of Self in Everyday Life (1959), the sociologist Erving Goffman argued that we all run two versions of ourselves. There’s the front stage self we perform for other people, at work, at dinner, on a first date. Then there’s the back stage, where the performance drops: the car after the meeting, the kitchen after the guests have gone home, the group chat about the meeting.

The back stage is what makes the front stage possible. You can perform all day because there’s somewhere you don’t have to.

Being a creator doesn’t have to get rid of the back stage, but it puts all of it up for grabs. Your morning coffee could become a ‘get ready with me’. A quiet Sunday could be a ‘reset routine’. A breakup could become a ‘storytime’ with three lessons at the end. Whether you actually post it almost doesn’t matter, because your audience knows you could. And once anything might be content, it gets hard to tell which parts of you are real. Even the “real me, no filter” post has been filmed, edited and captioned before anyone sees it. It can be just as much of a mind screw for the creator as it is the viewer.

When it stops being a metaphor

There’s a documented psychiatric phenomenon called the Truman Show delusion. The psychiatrist Joel Gold and his brother Ian, a philosopher, named it in the late 2000s after seeing patients who believed their lives were being secretly filmed and broadcast.

In February 2026, a case report in the journal Future Health described a 30-year-old social media influencer who developed it after a sharp drop in her engagement and a public falling-out with a close collaborator. She became convinced her home was full of hidden cameras. With treatment, including a supervised step back from social media, she recovered within weeks.

She’s one case, and psychosis is complicated, with causes that go well beyond social media. But the shape of it stayed with me.

Most creators are nowhere near that. They’re just doing a milder version of it, on purpose, every day.

So what is ‘the ick’?

People assume it’s vanity: the ring light, the ‘self-obsession’. I don’t buy it. Plenty of vain people are very charming, and we are all a little bit vain.

I think the ick is a trust problem. When someone shares their life for a living, you can never be sure whether what they’re showing you is real or made for the camera. Even when they open up, part of you wonders if it’s content. That doubt is what puts people off. It isn’t really about vanity. It’s not being able to tell where the person ends and the content begins.

It helps to remember how recently filming in public became normal. Before 2020, if you set up a phone and started talking to it on the street, people stared. It looked odd, and a bit embarrassing.

Goffman had a name for the rule it broke: civil inattention. When we pass strangers in public, we usually glance at them briefly to show we’ve noticed them, then look away to show we’re not going to intrude. It’s a small, unspoken agreement. I see you, and I’ll leave you alone. It’s a large part of why public spaces feel comfortable.

Filming yourself for social media breaks that agreement in two ways. You’re performing in a space where nobody asked to watch, so passers-by become an audience they never signed up to be. And you’re pointing a camera at a place full of people who might end up in your video without knowing. The strange looks were people noticing that the usual rules had been broken.

Then the pandemic happened. Lockdowns pushed people onto TikTok, Instagram launched Reels in August 2020, and short video became one of the main ways people found out about the world. Somewhere in that period the public mood flipped. Filming “silly little videos” stopped looking strange and started looking like work, and sometimes very well-paid work.

The bigger shift was in direction. We used to use the internet to document real life. Now real life is often arranged for the internet: the restaurant is picked for how it films and the trip is planned around the content. The offline world became raw material for the online one, which is exactly why it’s so hard to tell what’s real.

And once filming your life became work, it came with a pay cheque. That’s the other layer of the ick. A lot of people wouldn’t feel comfortable putting their life online to earn from it: their relationship, their home, their kids, their bad days. To them, it feels like something that shouldn’t be for sale. So when they watch someone else do it, and get paid well for it, the discomfort is partly a moral one.

And the money can be huge. The biggest creators can earn more from a single brand deal than most people earn in a year. That lands differently right now, when so many careers feel unstable. People are being made redundant, graduates are struggling to find entry-level jobs, and AI has a lot of people wondering whether their work will still exist in five years, hell, maybe even one year. If you’ve done everything you were told to do and still feel insecure at work, watching someone get paid to film their morning routine can feel ‘unfair’. I think some of the ick is that. It’s not quite jealousy. It’s more a sense that the rewards don’t match the effort, or the values, many people were brought up with.

Keep a door

If you’re building a personal brand, this is the most useful advice I have, and it’s the least strategic thing I’ll ever tell you. Decide what is never content. It can be different for each person but for example: a relationship, a hobby, a friend, your family dynamic. Protect it for your own head first.

Funnily enough, it makes the content better too. People can feel the difference between someone performing their whole life and someone choosing what to share.

You don’t have to walk off the set like Truman. You just need to know where the door is and go through it often.

If any of this feels uncomfortably close to home, it’s worth talking to someone you trust or a professional. Being watched for a living takes a toll, and plenty of creators feel it.

Sources

• Psychosis in the spotlight: Truman show delusion and social media vulnerability, a case report — Future Health, Feb 2026

• The “Truman Show” delusion: life as a reality TV show — ScienceDaily, 2012

• APA Monitor on the Truman Show delusion, 2009

• Erving Goffman, The Presentation of Self in Everyday Life (1959)

A few weeks ago I wrote about the social casino. The argument was simple. The platforms are slot machines, creators are pulling the lever, and the house always wins. I ended it with a promise: next time, I’d write about the house building AI influencers to replace the players.

I got one thing wrong.

It isn’t the house.

Last week I watched a man in a houndstooth suit shadowbox in front of the Eiffel Tower.

Blonde bob. Handlebar moustache, curled at the ends. Bow tie. Not a flicker of expression. He threw a few silent jabs at nothing, the caption said “she call me night she call you cabbage patch”, and the video ended. I watched it again. Obviously.

By the next morning my feed was him. Jean Philanthrope in a café. ‘Jean Phil’ in a gallery. Jean Phil under a tree in a Paris park. And then a second man, an Englishman called Archibald Brown, who apparently wanted to fight him.

It took me about three hours to go from “who is this AI guy?” to “why is he everywhere?” to the link in his bio, and the realisation that I’d been watching a pump and dump take over social and media (media had done no research when sharing it but that is a whole other topic of convo, I digress).

Then the full picture came into focus, and it was worse than a scam. It was efficient.

For a decade, the creator economy has run on a chain of conversions. Attention became followers, followers became a media kit, and the media kit eventually became brand deals. Each link in that chain took a cut, and each one took time. Whoever built Jean Phil (and ‘friends’) has simply removed the chain. There’s no brand, no product, no audience to nurture, not even a person. The attention is the asset, and it trades on pump fun by the minute.

And the two feed each other. Every view sends someone to the coin. Every jump in the coin’s price gives the internet a new reason to talk about him, and every conversation brings more views. The content is the engine.

This is where ten years of chasing virality was always going to end. We spent a decade telling ourselves views would turn into value one day. Someone figured out jumping straight to “one day.”

Two Things Are True Right Now

Thing One: Jean Phil is one of the most effective pieces of character content I’ve seen all year. A repeatable visual joke, instantly recognisable, absurd enough to rewatch and simple enough to remix because no one ‘owns it’. He went from his first post on 17 September to more than 228,000 Instagram followers and tens of millions of TikTok views in under two weeks. One video alone passed 35 million views. If a brand had briefed that, the agency would already be framing the case study.

Thing Two: Jean Phil doesn’t exist. He’s AI-generated. Three days after his first video, an X account appeared promoting “the official JEANPHIL Solana token”, linked straight from his Instagram bio. The coin climbed to a reported value of around $12 million. On 21 September it lost roughly 99% of its value in a single day.

Both of these things are true. And, once again, they’re true because of each other.

Two Slot Machines, Stacked

Here’s how I’ve come to understand it.

In the casino I wrote about last time, there’s one machine. You post, the algorithm decides, you win or you don’t.

Jean Phil runs on two.

The first machine is the feed. You make a character, you post, and the algorithm decides whether he travels. Most don’t. But you only need one to pop off. And now that it is possible at scale to products thousands of videos and accounts at once, the chances of one ‘popping off’ becoming much higher and easier to attain.

The second machine is the coin. It lives on a platform called pump.fun, on the Solana blockchain, where anyone can launch a token in about a minute for almost nothing. The price runs on an automatic curve, so every purchase makes the next one more expensive. Get enough people buying and the coin “graduates” onto a proper exchange, with charts and trading bots and all the theatre of a real market.

The character is the chip you put into the first machine to win a pull on the second.

And AI makes the chips nearly free. You don’t have to bet on one character. You can make ten. Jean, Archibald, a Jade, a Nigel, a Derek: all of them have turned up in the same orbit, several with coins of their own. Most barely registered. Archibald’s coin passed $1 million and now sits more than 90% below its peak.

So Who Actually Gets Paid?

This was the question I couldn’t let go of, so I went and found out.

Whoever launches the coin. They can buy in the first seconds, at the very bottom of the curve, before anyone else knows the coin exists. Then they can sell into the wave of buyers the videos send. On top of that, pump.fun pays a coin’s creator a cut of every single trade, buys and sells, of up to nearly 1% on smaller coins. A crash is still trading. Trading still pays.

(For the record, on the day of the crash the Jean Phil account posted that it had sold no tokens and that anyone could check the creator’s wallet on-chain. That may well be true of that wallet. The fees get paid either way...)

The snipers. Bots that buy automatically the moment a coin appears. When Archibald’s coin launched, one crypto alert account flagged it to traders when it was about eight minutes old. Earlier this year, CoinDesk reported on a sniper who turned $285 into $627,000 in a single day on a viral coin, while regular buyers lost out.

The insiders and promoters. People who know a launch is coming, or are paid in tokens to talk about it, and then sell to their own audiences. Now i’ve seen the pattern, I can’t un see it, and a lot of very strange content pillars are now making a lot more sense to me (perhaps more on that in another article).

A handful of lucky early buyers. They become the story, and every casino needs a jackpot story.

The houses. Pump.fun takes 1% of every trade on the curve whether the coin lives or dies, and it has made more than $800 million doing so. Instagram and TikTok collect tens of millions of views of engagement. Neither needs anyone to get rich. They get paid either way.

And who loses? The people who arrive late. Which, as it happens, is exactly when the algorithm is showing them the character most.

How a Pump and Dump Actually Works

I want to be careful here, because “pump and dump” gets thrown around a lot. So I’ll just describe the pattern and let you decide what to call it.

Someone creates a coin and buys a big chunk of it, cheaply and early.

Then comes the pump, which is attention. Viral videos. An “official coin” link in a bio. A promotional account tweeting six times on launch day. New buyers pile in, and every purchase pushes the price further up the curve. Rising numbers attract more buyers, each one betting there’ll be someone after them willing to pay more.

Then the dump. The people who bought cheap sell into the crowd. Because they hold so much and paid so little, their selling collapses the price, and the late arrivals are left holding tokens worth a fraction of what they paid.

What makes it a pump and dump, rather than just a bad investment, is that the people doing the promoting are the people who bought first. They aren’t hyping the coin so it’ll succeed. They’re hyping it so there’s a crowd of buyers to sell to. Every new buyer is someone they can offload their own tokens onto. So the marketing isn’t selling a product. It’s building their way out.

Crypto people have a name for those late buyers: exit liquidity.

Next round you introduce a rival. Or you get “the gang back together”. A fresh character, a fresh coin, another spin.

Across the platform, the numbers are brutal. An analysis by Solidus Labs found that 98.6% of tokens launched on pump.fun showed signs of rug pulls or fraudulent trading. Only around 3% of people trading there have ever made more than $1,000.

If that last number sounds familiar, it should. The median creator in the creator economy earns about $3,000 a year. It’s a different casino with the same odds.

The Algorithm Is the Accelerant

What struck me most wasn’t the coin. It was how fast my feed changed.

One full watch of one strange video, and within twelve hours this whole cast of characters was all I was seeing. Feeds like TikTok’s For You page and Instagram Reels don’t much care who you follow. They care what you watch, and how many times you watch it. A ten-second loop of a man silently punching the air in Paris is practically engineered to be rewatched.

For a coin that needs attention to peak within hours rather than weeks, that is perfect. The algorithm squeezes the hype into exactly the window the trade needs. And because my feed was saturated, it felt like everyone was talking about him. That feeling, the sense that everyone’s in on this, is precisely the feeling that makes people buy.

Except everyone wasn’t. My feed was. Yours might never have seen him once. Personalisation makes a narrow thing feel universal.

It’s a casino with no clocks and no windows, and the machine you were playing follows you home.

Even the Media Picked It Up

I ran social across fifteen luxury titles, so I know exactly how this happens. A social desk sees something trending, the numbers are enormous, the character is charming, and there’s pressure to be on it now. You post first and ask questions later. Or not at all.

Over the past fortnight I’ve watched media outlets, including ones I respect, pick these characters up as a fun viral moment before anyone realised what they were actually for. Nobody set out to promote a coin. But every “who is this mysterious Frenchman?” post pointed more people towards an account with a token in its bio.

I’m not pointing fingers. I’ve sat in that seat. But it’s the clearest proof I’ve found of something I keep coming back to. Social didn’t just become media. Now media is being gamed by social. The people whose job is to check are being fed the same hype, by the same algorithm, and rewarded by the same traffic for not checking.

The Influencer Doesn’t Have to Exist

The promise of the creator economy was: be yourself. Show up. Build something real with the people who follow you.

Jean Phil is that promise turned inside out. As fitness creator Colin Smith put it while watching all this unfold: “You just don’t even have to exist on the internet… make your quick hundred grand, get off, fly into the sunset.”

This is where everything I wrote in The Casino was heading, and the shift is easy to miss.

• In The Casino, creators were the gamblers. They put in time, money and energy, the algorithm decided whether they got paid, and most of them didn’t. The platform won regardless.
  With Jean Phil, the person behind the character has stopped gambling. AI makes characters almost free, so a flop costs next to nothing. When one does go viral, they make money from the coin whether it rises or crashes.

• The risk hasn’t gone away. It’s moved. The people now putting real money in and hoping to win are the audience, the ones buying the coin.

• So the creator has switched sides. They’re no longer a player. They’re making money from the players, the way the house does.

One anonymous account now does what used to take a studio, a distributor and a bank. Production, distribution and finance, all collapsed into a man in a houndstooth suit, punching the air.

So Who’s Winning?

There’s one more player I haven’t mentioned, and it might be the smartest one at the table.

After all this, the AI video company Higgsfield posted on X: “Officially, every viral AI influencer was made on Higgsfield.” Then the pitch: make one with your face or from scratch, then “jump on viral trends with Genjutsu.” The credit underneath went to the Jean Phil account.

If you haven’t come across Genjutsu, it takes an existing video, keeps the movement, the camera and the timing, and rebuilds everything else around your character. Its sister tool, Recast, does the simpler version: upload any video and a photo, and it swaps the person in the clip for whoever you’ve uploaded. Put plainly, the quickest way to make a viral AI character isn’t to invent anything. You take something that has already gone viral, a dance or a skit or a walk that a real person filmed and timed and sweated over, and you put a new face on it. That might also explain why Jean Phil and Archibald sometimes move in perfect sync.

And once you know the tools are marketing themselves through the content, you start looking at every “trend” a little differently. Who started it? Who benefits when a million people join in?

So who owns any of it? Nobody really knows. The person who made the original video rarely holds any meaningful right over twenty seconds of themselves on a phone screen. In the US, work made purely by AI can’t be copyrighted at all. The platforms’ rules were written for a world where a person posted a video of themselves. Law moves in years, and this moves in hours. By the time anyone decides whether a Frenchman who doesn’t exist can borrow your choreography, he’ll have had his three weeks, the coin will have crashed, and the account will have changed its name again.

So who’s winning?

The tool, which sells the shovels and then puts the gold rush in its advert. The bank, a launchpad that takes its cut of every trade whether the coin goes up or down. And the platform, which gets the views ‘real or no real’. None of the three needs anyone to get rich, and none of them carries any of the risk.

Then there’s everyone else. The creators whose videos get recast. The brands paying real people to make real things. Anyone who still believes the deal was be yourself, show up, build something. We’re all working very hard to make an honest living, in a casino where the house now sells the chips, lends you the money and owns the machines.

What I Know For Sure

AI characters are going to be one of the most creative formats on social, and some of what’s being made right now is brilliant.

But when something is suddenly everywhere, overnight, with a link in the bio, the first question isn’t “who is he?”

It’s “who’s selling?”

Two things can be true. He doesn’t exist. The money does.

Further Reading & Sources

On Jean Phil and the characters:

• Who is Jean Phil? — Dexerto

• Jean Philanthrope, the viral sensation with a $12M crypto behind it — Cointribune

• Dark truth behind mysterious influencer Jean Phil — LADbible

• Jean Philanthrope / Jean Phil — Know Your Meme

• Disturbing truth behind mysterious influencer — Bored Panda

On pump.fun and how memecoins work:

• 98% of pump.fun tokens flagged as scams (Solidus Labs) — BeInCrypto

• Just 3% of pump.fun traders have made over $1,000 — The Defiant

• Pump.fun launched 42,000 tokens in one day — Solana Compass

• Pump.fun’s creator fee model — Blockworks

A crypto sniper turned $285 into $627,000 in a day — CoinDesk

On the tools and copyright:

• Machina on X: claim that Higgsfield seeded the influencers

• Higgsfield on X: “every viral AI influencer was made on Higgsfield”

• Higgsfield Genjutsu

• Higgsfield Recast

• Inside AI: ChatGPT 80s photo trend

• Oimi: Who is Jean Phil? (no tool confirmed)

• Yahoo Tech/BGR: Higgsfield Steal and the copyright concerns

• Artlist: AI copyright and licensing in 2026

Part of the Two Things Can Be True series, exploring the contradictions that define how we live and work now.

Previously: The Social Casino

The Creator Economy Is the Future of Work // Most Creators Will Never Earn Minimum Wage From It

I have a confession to make before we get into this one.

I built a social agency. I knocked it down. And I am currently sitting in the exact career crisis this piece is about — trying to figure out whether my future is in the industry I have spent years inside, or whether it’s time to walk into the casino and pull the lever.

I still have fractional clients. I’m still doing the work I know how to do, building out marketing departments, helping brands figure out who they are and how to talk to people. But I’m also doing this — writing, creating, putting my thinking out into the world — because somewhere in the back of my mind I’ve started to wonder if the thing I’m helping other people build is actually the thing I should be building for myself.

Which makes me an extremely unreliable narrator for this piece. Or possibly the most reliable one.

You decide.

Two Things Are True Right Now

Thing One: The creator economy is one of the fastest-growing industries in the world. It is currently valued at $234 billion. Brands now allocate up to 25% of their digital marketing budgets directly to creators. Creator ad spend is growing at four times the rate of the overall media industry. The influencer marketing sector alone is on track to hit $34 billion in 2026. The money is real. The opportunity is real.

Thing Two: 73% of creators earn less than $30,000 a year. The median creator income is $3,000 annually — and it declined between 2023 and 2025, while the market boomed. The top 1% of creators receive 21% of all ad payments. The top 10% receive 62%.

Both of these things are true. And they are true because of each other.

What’s Happening to the Old Industry

Let me give you some numbers from the world I came from first, because context matters here.

The traditional advertising industry is not quietly restructuring. It is — to quote the trade press directly — holding “a funeral for advertising.”

Since 2024: IPG has laid off thousands. Omnicom has laid off thousands more. P&G cut 7,000 roles. Between August and December 2024, the American ad industry lost 4,600 roles. Job openings in advertising and marketing fell 7.5% between 2022 and 2025. The holding companies that have dominated the industry for decades are consolidating, automating, and cutting — in that order.

And where is that budget going? To creators. The same brand dollars that used to pay a full creative team’s salaries now pay a creator’s sponsorship fee. The transaction is real. The redistribution is happening.

On paper, this sounds like a revolution. The little guys winning. The monolithic agencies finally getting what they deserve. The creative class taking back the means of production.

On paper.

The Casino

Here’s what nobody in the creator economy wants to say out loud: the platforms are, functionally, casinos.

Every creator showing up consistently, posting into the void, chasing the algorithm — is pulling a lever on a slot machine. Some will win big. Most will walk away empty-handed after years of trying. A tiny few hit the jackpot and become the story everyone else points to as proof it’s possible — because that story is essential to keeping everyone else playing.

But the house — Instagram, TikTok, YouTube — always wins. It doesn’t matter whether you make it or not. Your content, your data, your audience, your attention, and your hope are all value flowing upward to the platform, regardless of what flows back down to you. They need you to believe you are one viral post away, because the moment you stop believing, you stop producing — and the machine stops running.

The dream isn’t a side effect of the creator economy. The dream is the product.

This is why the median creator income has fallen while the market size has grown. The money is real. It’s just not going to most of the people creating the content that makes the platforms valuable enough for brands to spend it.

You Can’t Be Made Redundant (But You Can Be Deleted)

Here’s the thing that the “become a creator” conversation usually offers as its killer argument: you can’t be made redundant from your own audience.

And that’s true. Your boss can’t call you in on a Friday afternoon. No holding company can restructure you out of existence. No client conflict can end your contract.

But the algorithm can demonetise you overnight. A platform update can halve your reach with no warning and no appeal process. A single controversy — deserved or not — can unravel years of work in 48 hours. Your account can be suspended. Your account can be deleted. The platform can simply change what kind of content it wants to reward, and if you built your identity around the old version, you start again.

You don’t have a boss. You have an algorithm. And the algorithm doesn’t offer severance.

The difference between employment and platform dependency is not as large as the creator economy would like you to think. You’ve just traded one form of precocity for another — one with better aesthetics and worse labour protections.

Who Actually Makes It

There’s a truth running underneath all of this that I find both uncomfortable and clarifying, because I can see it from where I’m sitting.

The people who tend to break through in the creator economy are not, by and large, people who emerged from nowhere. They are people who already had something: skills developed inside the industry they’re now disrupting. Networks built over years of doing the work. Financial runway — savings, a partner’s income, client work on the side — that allowed them to build an audience before the audience paid anything back.

Emma Chamberlain’s runway wasn’t family money. It was being 16 and living at home — which is its own form of financial insulation that most adults contemplating the same leap simply don’t have.

The creator economy dismantled the gatekeepers of the old industry. But the skills, connections, and stability required to succeed in it? Still flows disproportionately to the people the old industry already trained.

The revolution redistributed the format. It did not redistribute the advantage.

I say this as someone who is quite consciously using the decade I spent inside marketing to inform the content I’m creating outside it. I am not outside this observation. I am a data point in it.

Where That Leaves Me (And Maybe You)

I am sitting in a strange place right now. I still have clients. I’m still the person helping brands think through how to build, what to say, who to say it to. That work funds the time I need to figure out whether I can build something for myself — whether my thinking is interesting enough, consistent enough, me enough, to build an audience around.

Which means I am simultaneously: the person helping brands understand the creator economy, and the person tentatively walking into the casino, chip in hand, wondering whether to play.

I think a lot of people in marketing and advertising are in this exact moment and not saying it out loud. The industry is changing fast enough that staying still feels risky, but the creator economy is not the sure thing it’s being sold as. The ground between those two positions is uncomfortable. There’s no map for it.

What I keep coming back to is this: the creator economy is genuinely worth something to the people who approach it with honesty — about what they actually think, what they actually know, and what they actually can’t see yet. The content that tends to cut through isn’t polished performance. It’s the specific perspective that nobody else has, shared without too much dressing.

Which might be the most reassuring thing I can say — and also, I’m aware, the thing every casino tells you about why your particular hand is worth betting on.

What I Know For Sure

The money is moving from the old industry to the new one. That’s real and it’s not reversing.

Most people playing the creator game will not make a living from it. That’s also real and also not reversing.

The platforms extract value from everyone, winners and losers alike. The algorithm is not your friend. You are building on land you do not own.

And the people best positioned to navigate all of this are the ones who can see the system clearly enough to make an honest decision about where to place their chips — rather than the ones who got swept up in the dream without reading the odds.

Two things can be true. The casino is real. The house always wins. And sometimes, knowing that, you still decide to play.

I’m still deciding.

Further Reading & Sources

On the creator economy and income inequality:

• Creator Economy Statistics 2026 — inBeat Agency — comprehensive data

• Creator Economy Income Distribution — Archive.com — the distribution breakdown

• Creator Economy Market Size — Precedence Research

On the advertising industry’s collapse:

• A Funeral for Advertising — Storyboard18

• Large Agency Layoffs — CM Galaxy

• AI or Accountability? The Real Story Behind 2025’s Layoffs — Advertising Week

Next up: The creator economy gave real people a seat at the table. Now the house is building AI influencers to replace them.

I have to preface this by saying, I wrote and worked on this piece pre 28th of February, when US and Isreal attacked Iran. I do however think it is more relevant than ever so I am sharing it now. I hope you are safe, brave and blessed <3

There is something quietly disorienting happening right now that almost nobody is naming directly.

We are all technically living in March 2026. But we are not living in the same year.

Depending on when you first paid serious attention to artificial intelligence — and how deeply you followed it — you may be operating with a mental model that is anywhere from three years out of date to current. The gap between those positions is not a small one. It is the difference between thinking AI is a clever search engine and understanding that autonomous systems are already making decisions without human input. It is the difference between protecting your position inside an existing structure and quietly questioning whether that structure itself can hold, and wondering what will come next.

This is not a piece about who is paying attention and who is not. It is about the speed of change, the unevenness of access, and the unfamiliar feeling of realising it is March 2026 and that perhaps time has now fundamentally shifted.

All responses to experiencing the rise of AI and the changing of time — curiosity, anxiety, denial — are human. None of them are wrong.

Let’s have a look at the timeline shall we…

The Timeline

November 2022 — The starting gun

ChatGPT launched on 30 November 2022. According to a UBS analysis cited by TIME magazine, it reached 100 million monthly active users within two months — faster than any consumer application in human history. TikTok required nine months to reach the same milestone. Instagram required two and a half years.

Most people experienced it as a curiosity. A useful toy. Something that wrote emails and summarised documents. Almost nobody — including, by their own subsequent admission, many of the people who built it — understood the scale of what had just been released into the world.

March 2023 — The people building it start to worry publicly

GPT-4 launched on 14 March 2023. It was multimodal — able to process both images and text — and demonstrated human-level performance on several professional benchmarks, including passing a simulated bar exam in approximately the top 10 percent of test takers.

In the same month, over 1,000 researchers and technologists signed an open letter calling for a six-month pause on AI development, citing societal and humanitarian risks. On 30 May 2023, hundreds of AI scientists and industry leaders — including Geoffrey Hinton, widely described as the godfather of modern AI, who had just resigned from Google to speak more freely — published a formal Statement on AI Risk: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”

That statement was largely buried by the news cycle within a week.

2024 — The year it became structural

According to McKinsey’s State of AI report, 78 percent of organisations reported using AI in at least one business function by 2024 — up from 55 percent the year prior. OpenAI previewed Sora, a text-to-video model, demonstrating that the disruption would extend beyond images to video, voice, and human likeness. The European Union passed the world’s first comprehensive AI regulatory framework. Governments were drafting regulation approximately two years after it was needed.

2025 — The year of agents

This is where most people’s understanding currently stops. They still think AI is the thing that writes their emails. It is not, or not only.

McKinsey’s 2025 State of AI report recorded that 88 percent of organisations now use AI in at least one business function, and that 62 percent were at minimum experimenting with AI agents — autonomous systems capable of planning, reasoning, and executing multi-step tasks without human instruction. The World Economic Forum’s Future of Jobs Report 2025, drawing on a survey of 1,000 companies across 55 economies, projected that by 2030, 92 million existing roles will be displaced while 170 million new roles are created — a net gain globally, but one requiring the restructuring of 22 percent of all jobs and the obsolescence of 39 percent of existing skill sets within five years.

March 2026 — The reckoning

Steve Yegge, former engineering lead at Amazon, Google, and Sourcegraph, published a widely circulated essay in February 2026 describing what he calls the AI Vampire: the phenomenon of AI-driven productivity gains being absorbed entirely by employers, leaving workers exhausted without corresponding compensation or reduced hours. His proposed antidote was explicit: “Go touch grass every day. Do something without AI. Close the computer. Go be a human.” He argued the sustainable new workday, once AI is genuinely integrated, is three to four hours of cognitive work.

AI is no longer an experiment running alongside normal operations. It is inside them. Caroline Clark, writing in her Substack Liftoff with Caroline in February 2026, identified a convergence of three simultaneous consolidations — economic, political, and informational — all accelerated by technology. Drawing on Yanis Varoufakis’ book Technofeudalism, she describes how the AI companies now emerging mirror the pattern of cloud computing in the 2010s: a race to claim territory, lock in customers, and charge rent for access to infrastructure rapidly becoming as essential as electricity. OpenAI is backed by Microsoft to the value of $13 billion. Anthropic by Google ($2.3 billion) and Amazon ($4 billion). The entities controlling the frontier models are a very small number of very large players. Everyone else pays for access to infrastructure they did not build and do not own.

The IMF published a working paper in April 2025 finding that AI-driven productivity gains compound existing wealth inequality: those with capital are better positioned to benefit from the returns AI generates, while those whose labour is displaced have fewer resources to retrain or reposition. The Center for Global Development has documented a parallel divergence between nations — wealthier countries positioned to use AI to enhance productivity in high-value sectors, widening the gap with economies that cannot afford equivalent adoption.

The structure is not failing dramatically. It is continuing to perform. It is asking us to perform along with it. And the gap between those performing fluency with the shift and those genuinely navigating it is widening — quietly, without announcement, every single day.

The Levels — What Year Are You Actually Living In?

This is not a hierarchy. It is a map. The same person may occupy different positions in different areas of their life simultaneously.

2022 thinking AI is a clever chatbot. Interesting as a novelty. Working and living largely as before. The shift is theoretical.

2023 thinking Using AI for drafting, summarising, basic tasks. Treating it as faster search. Has not yet understood that the disruption is structural rather than functional — that it is not a better tool for doing the same things, but a different relationship between human capability and economic value entirely.

2024 thinking Integrating AI meaningfully into daily work. Beginning to understand which roles and industries are at risk. Possibly anxious. Still operating within the existing professional structure, using AI to preserve a position within it rather than questioning the structure itself.

2025 thinking Understands the agent shift — that AI is now acting, not just responding. Building with AI rather than just using it. Beginning to question whether the structures themselves can hold: employment models, intellectual property law, the meaning of value, the social contract between people and institutions.

2026 thinking Sees the full arc. Not panicking. Not denying. Understands that the most significant questions are no longer technical or economic — they are human. What kind of communities do we want to live in? What skills matter when most cognitive tasks can be performed by a machine? What is the difference between productivity and a life well lived? Quietly investing in things that will hold regardless of what the economy does. The embodied. The local. The slow. The trusted.

The Quiet Before

The WEF Future of Jobs Report 2025 projects that 39 percent of existing skill sets will be obsolete by 2030. McKinsey found that only 7 percent of organisations have fully scaled AI across their operations. The gap between those two facts is where most of us are currently living.

Nobody knows what comes next. The people building the technology do not know. The institutions attempting to regulate it do not know. The economists modelling its impact do not know. The range of outcomes on the table, from profound human flourishing to serious civilisational disruption, is genuinely wide. We are inside it, without the distance required to see its shape clearly.

What is clear is that the shift is already underway. It does not require your awareness to proceed.

How you prepare is a personal question with no universal answer. Mine looks like this: learning as much as I can about AI, regulating my nervous system, building real life skills, and reflecting seriously on my own values and where I actually stand.

That is not a recommendation. It is just where I am.

References

Center for AI Safety. (2023, May 30). Statement on AI risk. https://www.safe.ai/work/statement-on-ai-risk

Center for Global Development. (n.d.). Three reasons why AI may widen global inequality. https://www.cgdev.org/blog/three-reasons-why-ai-may-widen-global-inequality

Clark, C. (2026, January 24). #31 — Rewriting the belief that governs the world. Liftoff with Caroline.

Liftoff with Caroline

#31 - Rewriting the belief that governs the world

At Davos this week, Mark Carney addressed a room full of the world’s elite and political leaders to talk about Canada’s new role in the international order. It’s a brilliant speech and…

Read more

10 months ago · 7 likes · Caroline Clark

Clark, C. (2026, February 14). #34 — The power of three. Liftoff with Caroline.

Liftoff with Caroline

#34 - The power of three

Part of this article references Technofeudalism by Yanis Varoufakis, which I recently finished reading as part of a book club with CPO Connect, a community for product leaders at Head of level and above. We held the book discussion yesterday, so credit for some of the ideas in this article must also go to fellow colleagues in the book club: Harry, John…

Read more

9 months ago · 8 likes · Caroline Clark

IMF. (2025, April). AI adoption and inequality (Working Paper No. 2025/068). International Monetary Fund. https://www.imf.org/en/publications/wp/issues/2025/04/04/ai-adoption-and-inequality-565729

McKinsey & Company. (2025). The state of AI: How organisations are rewiring to capture value. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

OpenAI. (2023). GPT-4 technical report. https://cdn.openai.com/papers/gpt-4.pdf

Swartz, J. (2023, February 8). Why ChatGPT is the fastest growing web platform ever. TIME. https://time.com/6253615/chatgpt-fastest-growing/

Varoufakis, Y. (2023). Technofeudalism: What killed capitalism. Penguin.

World Economic Forum. (2025, January 8). Future of Jobs Report 2025. https://www.weforum.org/press/2025/01/future-of-jobs-report-2025

Yegge, S. (2026, February). The AI Vampire. Medium. https://steve-yegge.medium.com/the-ai-vampire-eda6e4f07163

It’s all a bit fucking boring.

Yes, this is a rant… (how did you guess?!)

Let’s just say that my break from working in social media for the last 14 years has meant that when I’ve come back online to peek into the crazy world you are all still adding to and consuming, it feels a bit, well, DULL.

I get it. We are bored people spinning around on a rock, no longer hunting and gathering but instead trying to figure out how to pay our rent or mortgages and make something meaningful and valuable in a very capitalist society.

But really, coming back online after a good long break in my now flourishing and beautiful balcony garden… what on earth is going on?! I had felt the undercurrent before - the “erm, we draw the line here” feeling when it comes to my career in social and not becoming a “prompt engineer” in the “attention economy”- but now there is no stepping back into this and thinking it is anything other than pretty shit.

AI slop has absolutely ruined my enjoyment of Pinterest, my previously favourite channel. All I’m now seeing are either AI prompts for fashion shoots (indistinguishable from real life) or gardens that, frankly, no amount of pruning will get you to.

LinkedIn makes me want to throw up in my mouth every time I hear “I’m delighted/honoured/proud”. Are you really though, Janet?! You couldn’t even write this post.

If I manage to read ONE post written in someone’s actual voice and not AI, it feels like a miracle after scrolling for 10 minutes.

Instagram is now flogging me “AI influencer” courses. Trust me, very few are making money and it is not worth the water used to power it.

Honestly, it is like the emperor’s new clothes out here. Are we all OK?! And HONESTLY, if I read the word “honestly” again, I want to chuck my laptop across the room. That, or “the real rub?”. IYKYK. WE ALL KNOW.

This comes across a bit like I’ve lost my mind, but I know for a fact it is not me (spoiler, it is all of us) and is in fact us as a society. We are numb. We are in denial. We are losing touch with who we are and what matters — at speed.

2026 feels like a much better time to opt out of adding to this absolute, boring chaos. Alas i’m writing about it. However, we are going to look back at this year (and ourselves) and cringe SO hard. The irony as I write this and post it online. I’m already cringing.

I predict there will be a race to be the last online and the last on social media. Like the great Facebook exodus for anyone under 40. The last ones consuming content will be the losers.

Will no one take a stand and do the most obvious thing ever and make a platform based on the “trust economy”?! (which is the most valuable commodity of our lives now) If you do, please share the link.

Again, if you didn’t get the messages, i’m bored online. Apart from Substack, thank goodness for Substack.

And yes, I bastardised ‘After the Ball"the 1899 painting by Catalan artist Ramon Casas i Carbó with Grok. Oh the irony.

Right, back to the garden.

These images are intended to be read as proof of human creation and posting, of real life being shared. A cry out to escape AI generated content. They are meant to reassure us that what we’re seeing is real. That a person was there.

I understand the instinct. When polish becomes cheap, we reach for texture. When perfection feels suspect, we trust mess.

But here’s the uncomfortable turn.
Mess itself can now be prompted.

Not just copied, but convincingly generated. With all the right cues. The accidental composition. The casual clutter. The appearance of effortlessness. What once felt like proof of life is now just another style.

That’s where something quietly shifts.

AI generated using this prompt: A macro photograph of a messy handwritten note on lined paper, written in blue ink with a ballpoint pen. The handwriting is slightly inconsistent in size and slant, appearing rushed but with the lines. Include subtle pen-pressure variations, slight paper smudging, and a few scratched-out words. The note is filled with to-do list items, arrows pointing to circled words, and marginalia. Realistic lighting from a slight angle,, 4k, no CGI look

Two stages of response

What’s striking is that many people are still responding as if we’re in an earlier phase.

Stage one is about proof.
People are trying to demonstrate humanness, and consuming content they believe signals it. Messy becomes meaningful. Imperfect becomes trustworthy. Texture becomes truth.

This makes sense. These signals worked for a long time.

Stage two is quieter, and more unsettling.
It’s the realisation that these same signals can now be generated just as easily. That the markers we’re relying on to identify one another in online spaces are no longer reliable.

The shift has already happened, but our shared understanding hasn’t caught up. The cost that once made these signals meaningful has collapsed faster than our ability to recognise it.

In that sense, much of what we’re seeing feels less like awareness and more like delay. Not denial as refusal, but denial as lag. People are still operating with a signalling system that no longer holds.

Why these signals stopped working

In economic theory, and later in communications and organisational studies, this dynamic sits within signalling theory, first formalised by economist Michael Spence in his work on information asymmetry.

Signalling theory looks at how observable traits or behaviours are used to signal underlying qualities in situations of information asymmetry. For a signal to function, it needs to be costly or difficult to fake. That cost is what gives the signal credibility.

Online, mess and imperfection functioned as signals of humanness precisely because they implied time, presence, and lived experience. They were not easily reproducible at scale, which made them meaningful.

But once those same markers can be generated cheaply and convincingly, they lose their signalling power. When a signal no longer carries cost, it can no longer reliably distinguish what it was meant to signal.

The breakdown here is not a failure of human judgement. It’s environmental. The conditions that once made these signals meaningful have changed.

Why this feels like a threshold

When any content can be generated, imperfection can be generated, and even proof of humanness can be generated, trust can no longer sit inside the feed.

It has to move elsewhere.

Into time.
Into consistency.
Into relationships.
Into reputation.
Into voice, presence, and embodiment.
Into things that cannot be instantly reproduced.

This is why there’s a renewed pull toward IRL experiences, smaller communities, slower media, and long-form thinking. Not as nostalgia, but as a structural response to a broken signalling environment.

The hidden cost

I’m not saying AI is bad.
I’m not saying this content is fake.
I’m not saying the trend isn’t a breath of fresh air.

What I am saying is that the audience is being pushed into a role they were never meant to occupy.

We’re no longer just consuming. We’re constantly, quietly verifying.

Who made this?
How was it made?
Why should I trust it?

That level of vigilance is cognitively exhausting. Historically, systems don’t survive that unchanged.

Social platforms were built on the assumption that authorship was legible. That you didn’t need to interrogate every artefact for signs of human life. When that assumption breaks, the system doesn’t end. It mutates.

The uncanny part

We’re not failing to read the signals.
The signals are failing.

And when that happens, people don’t just lose trust in content. They lose trust in the environment that trained them to read it.

That’s why this moment feels heavier than a trend cycle. It’s not about AI images versus real ones. It’s about what happens when the markers we rely on to recognise one another stop working.

A quiet threshold

I don’t think this is the end of social media. But it may be the end of a particular assumption. That authenticity is immediately legible. That reality can be read at a glance online.

When humanness becomes something you have to signal rather than something that’s assumed, you’ve crossed into a new phase.

And we’re only just learning how to live there.