UK care workers on £13.51 an hour report higher job satisfaction than financial analysts on £65,000 and up. Not because poverty is pleasant. Because professional work has been systematically failing people on the things that actually make work sustainable, and low-paid care work meets those things almost by accident. A care worker can see their work mattering, in real time, in the face of the person they are helping. Most analysts cannot say anything like that about a Tuesday.

And I want to be careful here, because this is not an argument that care workers should be grateful for poor pay. Poor pay is a separate and real problem. The point is structural. The model that produced the highest nominal earnings produced the lowest satisfaction, and a system that manages that inversion is telling you something about itself.

We called it the knowledge economy. A more honest name would have been the orchestration economy. If you map a typical knowledge worker’s week, and I have sat with enough directors to have effectively done this, maybe 15 per cent of it involves genuine thinking. The rest is logistics wearing a white collar. Emails about meetings. Meetings about projects. Status updates on the status updates, and so on. David Graeber gave this a ruder name years ago, and the reason the term stung is that it was recognisable. The research on how many people rate their own jobs as useless is in the chapter, and the sectors topping that table are the ones we are currently trying hardest to protect.

The money was never quite what it looked like either. A couple in their late thirties, both in professional services, £130,000 combined, which sounds like comfort until you put the mortgage, the childcare, the season tickets and the rest against it and find out what is actually left at the end of the month. High earners living paycheck to paycheck, which sounds like a contradiction and isn’t. They could not have downshifted if they had wanted to. The golden handcuffs weren’t golden. Just handcuffs.

There is a forty-year-old piece of psychology that explains most of this, three basic needs that make work psychologically sustainable, and professional services was structurally starving people of at least two of them while telling them they just worked too hard. “They worked too hard” was the cover story. It preserved everyone’s dignity. The real diagnosis was more awkward, which is kind of why the profession never discussed it.

Now hold that against what AI is actually automating. It is not coming for the care, the craft, the trust-built relationships, the accountability. It is coming for the orchestration. The status reports. The 85 per cent that was logistics in a lanyard. Which means that the thing being displaced fastest is, in large part, the thing that was quietly making people miserable, and the work it cannot touch is the work that was always real. That is an uncomfortable thought to hold at the same time as real grief about real lost incomes, and I do hold both.

Chapter 16 takes this all the way through: what GDP was actually designed to measure in 1944 and the car-crash problem that follows from it, why the career-for-life was a fifty-year anomaly rather than a baseline, what the downshifters worked out early, and why the portfolio pattern most professionals are already quietly living is older and sturdier than the salaried model it is replacing. The £13.51 inversion is the way in. The rest of the argument is in the book.

This is from Chapter 16 of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

^{Identities hidden to protect the ignorant.}

On Monday someone posted on LinkedIn that they still didn’t know a single real business that had actually benefited from AI. Capital letters on “single”. A hundred and thirty-odd people liked it. And I thought, well, I can help with that, because I know several, I built the tools for them, and I can see the hours being saved because I built the thing that tracks the hours. So I said that. With a link.

What came back wasn’t a question. It was a clapping emoji and “keep building, builder”. Then someone else turned up, quoted my headline back at me with “uh huh”, told me there was a lot wrong with me, told me I was part of the problem, and then, after I’d stopped replying, told me to cry about it, because when the bubble bursts the gravy train stops. Hope you have a plan B.

I’ll be honest, it annoyed me more than it should have. I’d answered the question I was asked, I’d done it with my name on it, and the answer was to have a go at me instead of at what I’d said.

But I’ve been sat with that last line since, “hope you have a plan B”, and I don’t think it was really for me. I run four businesses. I’ve got a plan B, and a C, and frankly a D that involves a mechanics garage. The person who wrote it is a senior developer, in the trade where junior hiring has fallen off a cliff, and I’d guess he knows it, because he’s watching it happen to the people coming up behind him. So when he says hope you have a plan B, I think that’s him talking to himself. Out loud. In the comments.

And the first post, the “not a SINGLE business” one, is the same thing really, just quieter. If nothing counts as evidence then you never have to do the sum. Every example is marketing noise, every result is rigged, so the bit where you sit down and work out what it means for your own job just never has to happen. I get that. I’m not sure I’d want to do it either, in that seat.

Nobody in that thread was allowed to say they were scared, and that’s the actual problem. You can’t really say it at work. Everyone’s meant to be excited, love the tools, exciting times, and so the fear has to go somewhere, and it goes looking for a face. On Monday the face was mine, because I’m the bloke saying it out loud. Which, fine. I build the stuff, I employ the people it replaces, I’ve made that call on a hire more than once, a spreadsheet on a Tuesday afternoon, and I’ll make it again. I’m a pretty easy face to pick.

But getting rid of the person saying it doesn’t stop it. The call still gets made, in pretty much every business in the country, whether I build anything or not. All you’ve done is make it quieter, and quiet is exactly how it lands on people. A role that doesn’t get backfilled. A restructure email with nobody’s name on it. Nobody ever actually telling you.

I didn’t write this to win the thread. I muted it. Honestly the insults were the best bit of it for me, because anyone reading down could see evidence on one side and “cry about it” on the other, and I didn’t need to add anything.

I wrote it because the question he threw at me is actually a decent question. It’s just pointed at the wrong person.

So, turning it round. Ten seconds, not a plan, just the first step. If your role went in the next restructure, what would you actually do on the Monday morning? Not the LinkedIn version, the real one. Who would you ring first, and would they be in a position to pick up?

If you can answer that without your stomach going, fair enough, you’ve got a plan B. If you can’t, that’s not a character flaw. It’s just a sum you haven’t done yet.

That sum is most of what my book’s about. The Next Rung is out on 19 January with Practical Inspiration Publishing, and you can pre-order it at thenextrung.co.

The half-life of technical skills is now under two and a half years. That estimate comes from BCG and Harvard Business Review, and it means a certificate granted once and held for life is, in a labour market moving this fast, a polite fiction.

Which is an uncomfortable thing to say to a parent, because the entire map most parents are using runs through certificates. Do well at school, get the grades, get the degree, get the job. That map worked. It worked for me, more or less, and it probably worked for you. The problem is not that the map was wrong. The problem is that the city it describes is being demolished.

Here is the mechanism, briefly. For most of the last century the exam was a convenient proxy for employability, because the economy rewarded exactly the behaviours the exam selected for. Remembering, collating, summarising, writing up. The orchestration work. And the orchestration work is exactly what AI absorbs first. When the work the exam selects for is the work the technology replaces first, the proxy breaks. Grades still measure what a written test can measure. The measured attribute has simply stopped being the best predictor of whether a student will have a career at thirty-five.

So what replaces it? I have skin in this game, I should say. I co-founded Enterprise Skills and built Skills Hub, a platform that measures the eight workplace capabilities AI does not replicate, things like decision-making, adaptability, team collaboration, leadership. Comprehension of business concepts among students using it improved by 73 per cent against traditional teaching. I have since left the business, but the thesis has not changed: the qualifications schools measure are becoming weaker signals every year, while the human capacities they do not measure are becoming priceless.

One scene from the chapter, because it shows the difference better than I can explain it. A Year 10 student scores 34 out of 100 on commercial awareness at the start of term. Below average, not alarming, the kind of score that leaves a person with an impression of themself as not-that-strong on the subject. Halfway through a supply-chain simulation, in a team of four, under time pressure, they spot a pricing margin error none of her teammates see. They say so. The team recalculates. Their score moves to 51, because they demonstrated the capability in action against a real-looking problem with other people in the way.

An exam cannot do that. It can ask her to explain what a pricing margin is. It cannot put her in a room where the margin is wrong and observe whether she notices. And that difference, between knowing and doing, is the difference employers care about most.

Chapter 13 of the book goes through the whole of this: what capability-based assessment actually looks like in a classroom (which, I should say, looks almost exactly like the classroom you were in thirty years ago), why the certificates in that system deliberately expire, the eight-minute parents’ evening where a teacher tries to put a second map on the table without telling the parent their own lived experience is wrong, and the four practical moves a parent can make now, before any of this reaches their child’s school. The first move costs nothing: at the next parents’ evening, ask what your child can do that a written test will never surface. The teacher will know. The report will not have asked.

The other three moves are in the chapter.

This is from Chapter 13 of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

The British government deployed some twelve thousand troops to put down the Luddites. Five years of unrest, four counties, an actual insurgency. Whatever you were taught the word means, that is not the response you mount against a few confused men with hammers.

The word has been turned into shorthand for someone who does not understand technology, and that is unfair to a fairly sharp group of textile workers. They were framework knitters. They smashed stocking frames because the new mechanised looms were producing cheaper, lower-quality stockings that undercut their craft. Their concern was specific: these machines are destroying our livelihoods, and what they are producing is worse. On the local destruction they were entirely correct. The machines did destroy their livelihoods. Several of them were executed for pointing it out too loudly. Lord Byron defended them in his only ever speech in the House of Lords.

Where they were wrong was in assuming the destruction of their particular jobs meant the destruction of work itself. The textile industry mechanised, expanded enormously, and employed far more people than before. The Luddites were not part of the expansion, but the expansion happened. And that is where the Luddite fallacy comes from, the five words people still reach for: it was fine last time.

For two hundred years those five words have been quite correct. But “for two hundred years” is doing a great deal of work in any sentence that uses the historical pattern as reassurance now. Every previous wave automated something narrow. Looms automated weaving. Computers automated calculation. AI is not narrow. Dario Amodei at Anthropic has put it in one phrase: a general labour substitute for humans. Chapter 1 of the book walks through the three pillars that made every previous absorption work, new task creation, demand expansion, skill transferability, and why all three are breaking at the same time. I won’t rehearse the whole argument here. That is kind of the point of the chapter.

What I want to sit with instead is the other half of the comparison, because it is the half the policy world is not thinking about. There is a comfortable assumption in policy circles that AI displacement happens quietly. The Luddites would like a word.

The difference this time is the cohort. AI displacement does not start on a factory floor in a mill town. It starts in capital cities, among white-collar workers who are more mobile, louder, better networked, and with stronger organising capacity than the professional class has historically needed. They are not people who are already poor, which is exactly why the scale of their response can be much larger than the analysts have pencilled in. The World Economic Forum has already named this cohort the “AI precariat”.

And I suspect the first major event will not look like 1811 at all. A coder fired because an AI replaced them does not march. They might run a digital riot, systems taken offline, things brought down that other people depend on. That is speculation, I should be honest about that. But it is the register the affected cohort already lives in, so it is the register the first event is likely to emerge from.

Chapter 9 sets out the watch-markers for that first event, the five conditions that produce movements, and the two other flashpoints I think are worth tracking alongside it. The Luddites got the Army. What this generation gets instead is the question the chapter ends on.

This is from Chapters 1 and 9 of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

Partway through I asked a question I thought I knew the answer to. If a client rang you today and asked how much of your job AI could do confidently, right now, what would you tell them?

Sixty per cent, one of them said.

And nobody argued. That’s the bit I keep coming back to. Not the number, the silence after it. A room full of smart, employed, busy people heard a colleague price sixty per cent of their profession into a subscription, and every one of them just nodded, because they’d all done the same maths privately months ago. They’d just never been anywhere it was safe to say out loud.

That’s the part that should make you angry. Not at them. At whatever it is that made a leadership-free room the only place the true number could exist.

Because here’s what those same people are swimming in. Feeds full of exciting times, and love what this unlocks, and we’re all in on AI. An industry where the only acceptable public emotion about this technology is enthusiasm. You can be excited at work. You can be curious at work. You cannot be worried at work, not out loud, because worried reads as behind, and behind reads as next. So the worry doesn’t go away. It relocates. It goes home, turns up at three in the morning as mortgage arithmetic, and comes back in the next day wearing a smile and posting about productivity gains.

Everyone is performing calm for everyone else. Leadership performs it for the team so nobody panics. The team performs it for leadership so nobody looks replaceable. The keynote circuit performs it for money, which is at least honest work. And every layer of the performance makes the next person more certain they’re the only one doing the 3am maths. They are not. They are all doing it. I have now sat in enough rooms to tell you they are all doing it.

Now, the uncomfortable bit. Sixty per cent is the polite number.

Here’s the mechanism, because this is where most of the commentary stops and it’s exactly where it shouldn’t. When you ask someone what a machine can’t do about their job, they answer with their best moments. The judgment call that saved an account. The thing they spotted that nobody else saw. The client who trusts them specifically. All real. Every one of those things is real, and none of them is what the working week is made of. The week is made of briefs, updates, reports, coordination, moving information from one place to another with a professional finish on it. The judgment moments are the reason they got promoted, which is precisely why they’re the moments that get remembered when the question lands. The routine is invisible to the person doing it because it’s routine. So when someone says sixty per cent, what they’ve done is count honestly for the first time and still round in their own favour. The remainder, the genuinely human bit, is real. It’s just not eight-hours-a-day big. And the salary was priced on the whole eight hours.

That’s the actual problem. Not that the human bit doesn’t exist. That it was never the bulk of the day, and the bulk of the day is now a feature in someone’s product.

There was a second question that afternoon, and I nearly didn’t ask it because I knew where it went. Eight hours of work, two years ago, took eight hours to do. How long does eight hours of work take now?

Two.

So here’s the question nobody in professional services wants asked in front of a client. How do you bill ten hours when the work takes two? Sit with how uncomfortable that is for a second, and then notice who the discomfort is protecting. Not the client. The client is already running their own version of the sum every time an invoice lands. Three grand a month to the agency, thirty quid a month to the tool, and a growing suspicion that the gap between those two numbers is padding.

Sometimes it is padding. Let’s not pretend otherwise, some retainers out there are running on momentum and a nicely formatted report. But I want to tell you what the gap looks like when it’s real, because this is the bit the invoice never shows.

In that same session, someone described a client who wanted a thousand meta descriptions done, because the tools make it nearly free and more sounds like progress. The team said no. Not because they couldn’t, because it was a waste of the client’s money and the client couldn’t have known that. Someone else talked about reading the recent algorithm updates, the ones going after mass-produced AI content, and quietly steering clients away from plans that would have run straight into a penalty. The client will never know the disaster that didn’t happen. That’s the product. The disaster that didn’t happen has no line on the invoice.

And the version of it I carry around from my own agency years. I used to watch a paving company’s search data every month, and every month one weird term showed up with three impressions. Three. For a product that didn’t exist. I Googled it, found nothing, eventually asked about it in their office, and someone behind the desk said oh, we get calls about that. It was a little plastic corner piece for resin paving. They made it. It sold everywhere, because nobody else had noticed the demand. Three impressions a month, spotted by a human with time to be curious, turned into a product line. No deliverable schedule produces that. Thinking time produces that, and thinking time is exactly what the two-thousand-nine-hundred-and-seventy-quid gap is supposed to be buying. The scandal isn’t that the gap exists. The scandal is how rarely anyone shows the client what’s in it.

The full diary is the other lie people are leaning on. I’m flat out, so clearly I’m needed. I can see time data across a whole business, and I’ll tell you where the hours actually go. Messaging. Coordination. Keeping each other updated about work the tools are increasingly doing. The busiest people I meet are frequently the most exposed, and their calendar is telling them the opposite, and they believe the calendar because the calendar is the only thing still saying nice things to them.

And then there’s the reassurance industry. The it-created-more-jobs-last-time crowd. Five words doing an extraordinary amount of load-bearing work, offered by people who will not be the ones carrying the load if it turns out to be wrong this time. Last time worked because the displaced had somewhere obvious to go. The factories were hiring. Tell me where the factories are for a forty-five-year-old account director, specifically, with a street address, and I’ll relax. Until then I’d gently note that the people saying it loudest are, almost without exception, people whose income does not depend on it being true.

For what it’s worth, I don’t think the team I sat with are in trouble. They were more honest in ninety minutes than most conferences manage in two days, and the honesty is the asset. The judgment is real, the context is real, the relationships are real, and crucially, once they’d said the sixty per cent out loud, they could finally talk about defending the forty. You cannot defend a position you won’t admit you’re standing on. Most of their industry is still refusing to admit the position. That’s the difference, and it has nothing to do with the technology.

So do the thing I asked them to do. Write one paragraph about your job that a tool couldn’t claim. Not your title, not your task list. The actual human part. Judgment you’ve exercised, context you carry, relationships that would walk out the door with you. If it comes easily, you’ve probably got something worth defending. If you’re on your third attempt and it all keeps coming out sounding like process, that’s information too, and better to get it now, from yourself, in private, than in eighteen months from someone holding a spreadsheet.

The room went quiet when I set that exercise, by the way. Same silence as the sixty per cent. Everybody already knows.

I’ve written a book about all of this. The Next Rung, out with Practical Inspiration Publishing in January. Pre-order at thenextrung.co if you want the long version of the 3am maths, and what to do about it.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

AI-native startups generate average revenue per employee of $2.5 million. Traditional companies average $200,000. That is not a percentage improvement. It is a twelve-and-a-half times multiplier, which is a different economic species altogether.

I have watched that slide land in boardrooms, and the response is always a gasp. But there are two distinct kinds of gasp, and which one a board produces tells you quite a lot about where that company will be in a few years.

The first gasp says: that is a crazy number, but it does not impact us, because of reasons one, two, and three. They convince themselves it is an AI bubble, or a startup bubble, or both. Not a real business. Not a seasoned business that has been through the trenches. And to be fair, some of those reasons sound quite sensible in the room.

The second gasp says: what can we do to get that?

Both reactions exist in the same quarterly board pack, sometimes in the same meeting. The first is a managed-decline strategy wearing the face of cautious realism. The second is the start of a transformation plan, which may succeed and may not, but has at least noticed the competitive environment.

McKinsey’s own framing is blunt. Only 6 per cent of companies are achieving what it calls “transformative” AI impact, and the difference is not how much AI they use. It is whether they have rebuilt their operations around it, or simply bolted it onto what already exists. Bolting AI onto a legacy structure is like putting a jet engine on a horse cart. You get noise and expense, and the cart falls apart.

The thing I keep coming back to, from the advisory side, is that most people inside these firms are on rails. The decisions are not really ones they take. They wake up, they go to work, they do the work in front of them, and they see the failures of other businesses as evidence of their own excellence rather than as a pattern. Which means that in many cases nobody sees it coming. They walk into it.

Chapter 6 of the book follows that walk in detail: how the mandate gets private backing long before it is announced, the sentence that starts the cascade in every boardroom (“we can’t afford not to”), the sequence in which the professional sectors go, and what it means for the professional class when the economic viability of a role stops being a function of how good you are at it. Manufacturing workers have always known that plant closures do not care how good you are at the job. Professionals never had to know it. That is the part of the chapter I found hardest to write, and I suspect it is the part most readers will find hardest to read.

This is from Chapter 6 of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

Nearly nine in ten organisations are now using AI in at least one business function. That is the number companies put on analyst slides. It is not the number that matters.

The number that matters comes from ISG, the Information Services Group: 31 per cent of use cases have reached full production, and that figure is doubling annually. “Full production” means the AI is not being tested. It is doing the work. Decisions flow through it. The people who used to do those tasks are doing something else, or nothing at all. Run the arithmetic on a doubling rate and you can see how long “most companies haven’t really adopted it yet” survives as a comfort.

So the question has already moved on. It is not whether your company uses AI. It is whether your company has been redesigned around it, and those are very different things.

My advisory work lives inside that difference every week. The companies that have actually moved through AI, and implemented it into the orchestrational areas of the business, have org charts that look like a small group of very smart, very capable people sitting directly on top of the tooling. There are no juniors or middle people below them. There is nothing below them. The tooling does the work the juniors used to do.

And then there are the firms that ran a pilot and ended up with Copilot and ChatGPT bolted on to the old structure. They still have the junior staff, the middle managers, the pretty presentations. One shape is nimble and fast and smart. The other is performative.

Adopting AI is buying the gym membership. Redesigning around AI is actually going to the gym.

The firms I worry about most are not the big ones, who have the capital to transform, and not the AI-native startups, who never had to. It is the mid-market, the firms employing the bulk of the professional class. Too big to pivot fast. Too small to fund the transformation. In six advisory sessions in the last year I have heard some version of the same sentence: “We cannot afford not to do this. We cannot afford to do this.” Sit with that sentence for a minute, because a lot of businesses are currently living inside it, and the window for getting out of it is shorter than most of them think.

Chapter 6 of the book is about what happens when this pressure changes state. The chapter is called Critical Mass, and it goes through how these decisions actually get made behind boardroom doors before anything is announced, the sector sequence (legal first, then consulting, then accounting, and why in that order), and the moment when being excellent at your job stops being the thing that protects it. I won’t unpack all of that here. The short version is that the corporate world has been heating steadily for a while now, and water at 99 degrees is still water.

This is from Chapter 6 of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

A TikTok that turned up in my feed put it as two options. One, two companies who’ve never shown much regard for the rest of us decide on the same day to slow down the thing they build behind closed doors, because it might gain consciousness and end civilisation inside three years. Two, the same two companies notice each release is a smaller step than the last and, rather than say so, build a story about how powerful it all is, which frightens everyone and hands their investors a reason for the promises they haven’t kept. Two firms that can’t agree on a benchmark agreed on this inside a weekend, so yes, I’d be suspicious too.

What actually happened

The facts, quickly. Dario Amodei published his essay on Saturday 12 September. Sam Altman backed it within hours and Elon Musk followed. On the Monday Nvidia lost three percent and Intel six. Altman told Fortune the same weekend that this was an “ill-advised moment to go public”, so no OpenAI listing in 2026, and that’s a company that has already filed confidentially. The Conversation asked whether it’s a safety pause or a strategic retreat. PJ Media skipped the question and called it cartel-building. The incentives line up. I’ll give you all of it.

The bit I can’t cost yet

What I can’t work out yet is what a real ceiling does to the money. The data centres, the chip orders, the whole capex story was priced on the next model being better than this one. Monday was a small preview of that and it ran on nothing more than a blog post. A permanent ceiling is a much bigger question and I haven’t got to the end of it.

The bit I can

The labour side I can speak to, because I build one of the tools. Orca runs on the models that exist today, live, with paying clients, on whatever the frontier was in the spring. Every bit of the value came from the scaffolding round the model, the tools it can reach and the approval queue with a human sat in front of it. The model was the cheap part. If nobody ships another one, my clients won’t notice. The work it took over stays taken.

Two thirds haven’t switched it on

Now the number that matters, and it isn’t a market one. Around a third of UK firms with ten or more staff say they use any AI at all, thirty-five percent the ONS said in July, up from twelve at the end of 2023, and the ones that do use it average one point six tools. Construction is at thirteen. So two thirds of British employers haven’t switched on what already exists, and the third who have are running one and a half things. That’s the headroom, and it’s roughly what Altman himself said in June, that adoption is early and most firms are still working out where to put it. Nothing in Dario’s essay touches any of that. He says plainly that pacing doesn’t mean halting training and that “progress will still seem fast”. He’s slowing the next model. The current ones stay on sale

.

What the current models already did

And look at what the current ones did before anyone said the word pause. A Work Foundation survey out in August found thirty-six percent of UK employers had already cut entry-level roles. Adzuna counted graduate adverts in July down forty-nine percent on the year. Vacancies overall were the lowest since spring 2021, and the government’s own review in January found the same pattern in the exposed occupations. All of that happened on models that already exist. Some of it on models that are two versions old now. A freeze doesn’t unship them.

Forty years of nothing, then the rebuild

The last time a general purpose technology did this it took forty years, and the lag is the interesting part. Paul David wrote a short paper in 1990 about the dynamo. Edison’s bulb was 1879, the first generating stations 1881, and by 1900 electric motors drove under five percent of factory machinery. The gain only came when someone stopped swapping the big steam engine for a big dynamo and put a small motor on every machine, then rebuilt the factory floor round that. Decades of nothing, then the rebuild, and the rebuild is where the jobs moved. The difference this time is the rebuild is a licence in a Microsoft tenancy and a person with a browser. The lag is procurement and nerve, and neither of those needs a new model to run out.

Freeze it and the ladder gets worse

Now the part nobody’s costed, and it’s the part that gets worse if you’re right about the wall. Firms cut the juniors because today’s models do junior work. That’s what the Work Foundation number is. Freeze the models and they never get to senior work. So in 2035 you still need the senior, the person who can tell when the output is wrong, and you stopped making them in 2026. The training was always funded by junior output and the junior output went to a subscription. A plateau doesn’t rescue the graduate. It strands the firm as well, about eight years later, when the seniors it kept start retiring. If I believed the wall, that’s the version of the ladder I’d be worried about.

Mollick has the general version

Ethan Mollick at Wharton has been saying since 2023 that if development stopped today you’d have five years, then it was ten, of absorbing what already exists, and he said it again this week as the pause talk started. He’s right, and it’s the general version. He’s talking about integration and productivity. The bit missing is the pipeline. Absorbing a technology takes people who know the job well enough to rebuild it round the tool, and the entry rung is where those people used to come from.

What a freeze does change

What a freeze does change is speed, and I should be straight about that. Without a launch every quarter the roll-out runs at the pace of procurement and legal, which is slow. In the book I said mid-2030s for the damage across the middle. I’d still say that, maybe move the far end out a year, two years, I don’t know. What I wouldn’t move is the tipping point. One competitor halves its delivery cost with the tools that exist and everyone in the sector follows inside a year, no launch event needed. Everything stays the same until it suddenly does not. That line’s in chapter nine and it was written before any of this.

Ten seconds

So, ten seconds. Open last week’s calendar. Count the things you did that the Copilot licence your IT team already bought couldn’t have done. Then find out whether it’s been switched on for your department yet. That answer matters more than anything the labs decide this month.

Disclosure, since it matters here: I build one of the tools doing this, and I wrote a book about what it does to the ladder.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

The first draft of this book came back from its concept review with a simple diagnosis. I had overthought it.

I had overthought the message, overthought the arguments, and I had been trying to speak to several audiences at once, which meant I was speaking sharply to none of them. The reviewer offered a free second look if I rewrote with a clearer sense of who it was for. Fair. Painful, but fair.

The strange part was working out why the draft had gone wrong, because I had worked hard on it. The problem, it turned out, was the working hard. The more I polished, the more things flattened. Short sentences tidied themselves into pairings. Specific advisory conversations abstracted into general claims. It stopped sounding like something written for one reader and started sounding like something pitched at a room. Anyone who has read much AI-generated text will recognise that flatness, which is exactly the point, and exactly the trap for someone writing a book about AI.

So the fix was to stop polishing in any recognisable sense of the word. I talked instead. I wrote my own interview for each chapter, a set of questions designed to draw out what I actually thought, then recorded myself answering them. Interviewing yourself and editing yourself is an odd process, and I am not sure I would recommend the first week of it to anyone. But when I talk to a client about AI displacement, I use certain phrases to their face, and those phrases are now the sentences in the book.

The part I get asked about most is the checking. Alongside the interview recordings, I built a voice-and-style profile from fifteen hours of my own meeting transcripts, actual client conversations. Every draft was measured against it. AI tools handled the mechanical work, organising transcripts, stripping filler, assembling drafts from the source material, and the guardrails were strict: no generating new substance, no rephrasing my words into a different register, no claims I had not made. The tools did not invent. Every sentence can be traced back to something I said out loud or wrote down earlier.

I am telling you this upfront, and in the book itself, because AI-generated writing in published books is a live issue, and because my credibility as the author of a book about AI depends on the book actually being mine. It would have been very easy to have a model write a plausible book on this topic in an afternoon. Plausible is the problem. I have spent two years watching what plausible-but-hollow does to professional work, which is roughly what the whole book is about, so the process had to be the argument as well as the method.

There is a fuller account in Appendix A, including how the sweep reports worked and what happened when a pass tidied a paragraph into something more abstract, which I will leave for the book. The short version is the one that matters. The tools shaped how the rewrite was assembled. The substance came from me. The book is mine, and the process existed to make sure it stayed that way.

This is from the Preface and Appendix A of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

“That would have taken my assistant half a day.”

I hear a version of that sentence most weeks. There is a specific moment that keeps recurring in my work, and I have started watching for it the way you watch for weather.

The setup is always the same. I am demonstrating Orca to a prospective user, usually a founder or a director. We connect their calendar. We pick a meeting from the following week. Orca generates the brief: background research, relationship history, CRM context, company data, preparation notes. The whole cycle takes less than a minute.

And then the person looks at the screen. They look at me. There is a pause. That is the aha moment, and it is the one every software demo in history has been built to produce. Lovely. Great for business.

It is what comes next that the book is about. Because after the aha, they say something. Sometimes it is the half-a-day line. Sometimes it is “I used to have someone doing exactly that.” And once, with the kind of uncomfortable honesty that arrives unexpectedly in a sales demo, it was a question about a real person on their payroll, by name. What she does now, once the tool does this.

That is the uh-oh moment. The aha is about the product. The uh-oh is about a person, and it usually arrives about four seconds later, which I think tells you something about how close to the surface this all is. People have already done the maths somewhere in the back of their head. The demo just makes them say it out loud.

I do not have a good answer to the uh-oh question, by the way. I want to be straight about that. What I have is a product that works, a business model that scales, and a growing understanding of what the efficiency I am selling adds up to across hundreds of founders. I wrote a whole post about the Sarah question a few weeks back, about what I actually say now when it comes up in the room, so I will not rerun it here.

What I want to flag instead is the pause itself. If you work in a support or orchestration role, someone may already have had this exact moment while looking at software that does a chunk of your job. They did not tell you about it. There was no announcement, because there is nothing to announce yet. Just a founder, a screen, and a pause. The uncomfortable truth is that the pause is where displacement actually starts, months before anything shows up in a spreadsheet.

The Prologue of the book is called The Builder’s Dilemma, and it lives inside that pause. The justifications I rotate through for building the thing, the conversation with an agency owner that would not leave me alone, where I land on the whole what-Amodei-said, what-Altman-said question. I have kept all of that for the book, partly for space and partly because it works better as one argument than as fragments.

But the demo moment you can have for free, because you should know it exists. Aha, then uh-oh. Less than a minute, then a pause. It is happening in meeting rooms all over the place, right now, and almost nobody in the roles being discussed knows the pause is about them.

This is from the Prologue of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

I am building the tools that do the displacing.

That is the first line of the book, and I put it first because you deserve to know it before you decide whether to read anything else I write. In any given week I sit on three sides of this problem. Not in theory, not as a thought experiment. In my actual diary, between actual meetings, with actual people whose working lives are being reshaped by the things I build.

The first hat is Orca, an AI executive chief of staff. It handles the admin and orchestration work that used to require a small support team around a single founder. I have written before about what that does to hiring, so I will not repeat it here, but the short version is that the support roles stop being created.

The second hat is the one people forget. I co-founded Enterprise Skills and built Skills Hub, an EdTech platform that measures the eight workplace capabilities AI does not replicate. Commercial awareness, decision-making, problem solving, and so on down the list. I designed and built it end to end and have since left the business, though it still runs on what I built. The thesis behind it is that the qualifications schools currently measure are becoming weaker signals every year, while the human capacities they do not measure are becoming priceless.

The third hat is advisory. Business owners, mostly agency founders and SME directors, strategy and operations. Increasingly those conversations are about AI. And increasingly they end with the same conclusion: you need fewer people. I do not enjoy saying that sentence. I have stopped pretending it is not the sentence.

So that is the geometry. I build the thing that removes the work, I built a thing that measures what the work leaves behind, and I sit with the people deciding when to pull the trigger. One recent conversation, with an agency owner who has spent fifteen years building her team and went to one of their weddings, is in the Prologue in full. I am not going to retell it here, partly because it deserves the space the book gives it, and partly because it still sits a bit heavy, if I am honest.

Why lay all this out? Because most books about AI and work are written from one side. The optimist side, usually, or occasionally the doom side, and almost always from a distance. This one is written by someone who profits from the displacement, worries about it, and advises through it, all in the same week. I get to be both the disease and the treatment, and whether that is a defensible position is a question I have not finished answering. The Prologue is called The Builder’s Dilemma for a reason. It also covers what Amodei and Altman have each said about the scale of this, and why one of their framings bothers me a lot more than the other.

If you have ever wondered whether the people building these tools know what they are doing to the job market, the answer is yes. We know. This book is what one of us decided to do about knowing.

This is from the Prologue of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

Orca does not usually replace a role directly. It much more commonly prevents one from being hired. And of the two, the second is the one nobody sees.

Quick context if you are new here. I build Orca, an AI executive chief of staff. Meeting prep, inbox triage, CRM, research, the social posts, the things that used to need a small support team around a single founder. It works, which is sort of the problem.

Because here is what actually happens. A founder running Orca does not make the next hire they had been planning. The PA they would have brought on. The social media manager they were about to brief. The junior researcher with a job spec already drafted. None of those hires happen. And there is no redundancy announcement, because there is no redundancy. The job spec just sits in a folder, and the budget line quietly goes away at the next planning cycle.

Which means that the person who would have got that job never finds out it existed. They are out there applying for roles, getting fewer interviews than last year’s cohort did, being told the market is competitive. A cancelled vacancy appears in no unemployment statistic. Nobody who was never hired shows up in a layoff tracker.

I think this is the bit most people have wrong about AI displacement. Everyone is waiting for it to look like something. A restructure with a name on it, a headline with a big number, a queue outside a building somewhere. What is actually arriving first is absence. Jobs that would have been posted are not posted. Teams that would have been backfilled are not backfilled. Each individual decision is unremarkable, and the aggregate is anything but. If you are waiting for the announcement before you take your own position seriously, I think you will be waiting a long time, and the wait is the expensive part.

I will not pretend I am watching this from a safe distance. The efficiency I am selling to one founder is, in aggregate, a hire that never happens for someone else. I have a small set of justifications I rotate through for that, the someone-else-would-build-it one, the real-value one, and so on, and the honest version is that none of them fully gets me off the hook. The Prologue of the book is called The Builder’s Dilemma, and it goes through each of those justifications properly, plus the one conversation with an agency owner that I have not been able to shake since. I will leave those in the book rather than rehearse them here.

For now, the point is smaller and closer to home. If your own role sits anywhere near the orchestration layer, the question is not whether your company will announce something. It is whether the role behind you, the one that used to be the next rung down, is still being hired at all. Worth asking. Quietly, if you have to.

This is from the Prologue of The Next Rung, my book on what AI displacement actually looks like from inside. It’s open for pre-order now:

Pre-Order The Next Rung

I’ll argue with that.

Not with the intent. The intent is decent, and the people delivering it are decent people solving exactly the problem they’ve been funded to solve. The problem is that the problem they’ve been funded to solve is yesterday’s.

Here’s the picture. Since ChatGPT launched, roughly one UK entry-level job advert in three has gone. Not moved. Not rebranded. Gone. The government looked at that and concluded the issue is young people can’t get onto the ladder, so the fix is helping more of them onto it. Bootcamps. Guarantees. A £1 billion youth package, or £725 million extra in the new levy for 50,000 more young apprentices, depending which press release you read, the accounting between the two is not my favourite bit. Fund the bottom rung harder. That’s the whole play.

And to pay for some of it, this is the bit I keep coming back to, they defunded the rungs above. From September the Team Leader apprenticeship loses its funding. So does Operations Manager. So does Chartered Manager. Sixteen standards cut, mostly the ones people used to climb from entry-level into running things, about £200 million redirected downwards. Removed the next rung to pay for the bottom one.

The market is removing the bottom one anyway.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

It’s the same everywhere, which is what made me want to write this down. In April the US Department of Labor launched a national push to embed AI skills into apprenticeships, around $243 million by one count, so that workers are “not merely participants in the AI economy but leading it”. Which in practice means teaching apprentices to use the tool that’s replacing apprentices. Singapore is handing out six months of free premium AI subscriptions with SkillsFuture courses and held a parliamentary debate literally titled “An AI Transition with No Jobless Growth”. Singapore is honestly the best of them. They’re at least putting serious money at mid-career workers as well, and they’re doing it before the crisis rather than after. But look at what all of it trains. Tool fluency. Prompting, operating, integrating. Entry skills, for a layer of work the tool itself is absorbing.

Why do governments keep doing this. Partly it’s measurement. Youth unemployment produces a statistic this quarter, and a department can fund against a statistic. The thing that’s coming doesn’t produce a statistic yet, so as far as a spending review is concerned it doesn’t exist. And partly it’s muscle memory. Retraining is what we did after the mines and the mills and the call centres, and it barely worked then. The New York Fed tracked 1.6 million job-training spells and found that outcomes for AI-exposed workers were limited. Limited is their word. We keep pulling the retraining lever because it’s the lever that’s there.

Meanwhile the researchers who went looking at firms actually adopting AI found junior employment falling about nine per cent within six quarters, call it eighteen months, and falling through reduced hiring rather than layoffs. Nobody gets marched out of the building. The building just stops letting people in. Hold that next to a government guarantee of entry.

So here’s tomorrow’s problem, the one no scheme I can find is aimed at. It isn’t youth unemployment. That’s just the visible edge of it. Tomorrow’s problem is where judgment comes from. Every senior person you rate got that way by doing years of junior work, drafting the thing, getting it wrong, being corrected, slowly building the pattern library that lets them look at a piece of work and know it’s off before they can say why. When AI absorbs the doing, that learning goes with it. A profession that no longer grows juniors into seniors eventually atrophies. I’ve spent a year looking for a government programme, any country, that funds judgment formation rather than tool use. The closest thing I found was a sentence in that Singapore debate saying AI cannot replace human judgment and empathy. Which is true. It’s also a sentence, and sentences aren’t policy.

What would solving tomorrow’s problem look like. I’m not certain, and I’d rather say that than pretend. But it starts in a different place. You’d fund the formation, the supervised reps, seniors paid to grow people rather than just review output. You’d build apprenticeships around challenging AI work rather than producing what AI already produces. And you’d say the thing nobody wearing a lanyard will say, which is that some of those kids in the bootcamp would be better served training as electricians. The plumber on sixty grand is structurally more secure than the marketing director on seventy-five. I wrote that in the book and it annoyed people, and it’s still true.

The bootcamp kids in Wigan start this summer. By the time they finish a Level 3 it’ll be 2028. Watch what the guarantee is worth then. I hope I’m wrong. One advert in three says I’m probably not.

This is the argument of my book, The Next Rung, out with Practical Inspiration Publishing in January. Pre-orders at thenextrung.co.

Subscribe now

Four months from now the book is out. I have been writing this newsletter since April, testing the arguments, watching them land. Some of you have been here from the start. If so, you know what is coming. If not, the back catalogue starts with “It was fine last time” and works forward.

This post is not about the argument. You have had the argument. This one is about the cost of writing it down.

Two years. That is what the book took. Not two years of full-time writing. Two years of running Orca, co-running Enterprise Skills, advising clients, chairing a committee at the Worshipful Company of Entrepreneurs, and writing a book in the gaps. Weekends. Evenings. The hour between the last client call and dinner. Early mornings when the house was quiet and the argument was loud.

Vicky read every draft. She did not tell me it was good. She told me when it was not clear. That is more useful and harder to hear.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

The thing I did not expect: writing the book changed what I saw in the business. I could not build Orca without seeing Sarah. I could not advise agency owners without seeing the quiet squeeze. The book made the contradiction louder, not quieter. Every client conversation became a data point. Every product improvement became evidence.

I could have kept quiet. Orca is a decent product. The clients are happy. The business works. Nobody asked me to write a book explaining that the technology I sell is going to displace millions of knowledge workers. That is not a natural marketing strategy.

But the honesty is the position. It is the only position I could find that did not require me to pretend I did not see what I was seeing. I am building the loom. I know it. The book is what happens when you decide to say that out loud instead of hoping nobody notices.

Here is what it taught me.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

The people who are going to be fine are not the ones with the best skills. They are the ones with the most runway. A partner in a different sector. Some savings. A trade. A network that is not just LinkedIn connections. What separates the five from the one, in Chapter 11, is not talent or effort. It is what was standing near each of them when the floor gave way.

That is the hardest thing in the book. Not the economics. Not the data. The fact that outcomes depend more on circumstance than on individual action. The book gives you the honest audit and the positioning tracks because those are the things you can control. But it does not pretend that everyone starts from the same place. Ryan did not.

Writing about Ryan was the worst part. Not because his story is dramatic. Because it is quiet. Three platform jobs. £18,000 a year. Rent at 63 per cent. £400 in savings. No way out that does not require resources he does not have. The book includes him because leaving him out would have been dishonest. He is the one in six.

Four months. The pre-order is live.

If you have been reading since April, you know what is in this book. The absorption mechanism breaking. The mortgage problem. The trades inversion. The honest audit. The builder’s dilemma. The question of who captures the surplus.

If you have not been reading since April, start with “It was fine last time” and work forward. The argument builds. It was designed to.

The career ladder is dissolving. The book maps what comes next. I wrote it because I could not not write it, which is the only honest reason to write anything.

Nobody lost a job. That is the bit I keep coming back to. There is no redundancy figure anywhere with that role in it, and if you went and asked them, which I have, they would tell you they have had a good year. They would be right.

Bill Gates published an essay on Wednesday about AI and work, and in the middle of it there is an idea he is calling Human Reserved. Jobs machines could do perfectly well, that we decide to keep for people anyway. He compares it to a nature reserve, land you could build on and choose not to because the loss would be too great. He got there through the carers who looked after his dad through Alzheimer’s, and he says what they did was irreplaceably human, which, having watched something fairly similar, yeah.

Then he does the thing almost nobody does, which is say out loud that he has not got the answers. His words: who gets to decide what we reserve for humans, what criteria should we use. And then, those will need to be worked out in public.

So that is what I have spent the last day on.

The test he is working with is whether something is irreplaceably human. I would tighten that slightly and say the real test is whether you have to be in the room. And that covers a lot more than care work, which is the useful part of it. You have to be in the room to work out what is wrong with the wiring in a hundred-year-old house. A model cannot know what a century has done to a wall by looking at a photo of it. Same way a carer does not need an algorithm to work out that the bloke in the bed wants another pillow.

But be honest about it and it does not stretch to people who take meetings for a living. The coordinating layer, the chasing and scheduling and formatting and summarising that most junior professional jobs are actually made of, fails that test completely. None of it needs a body in a room. Which is exactly why it goes first.

And that is the problem, because that is also where people learn.

An office coordinator does not stay an office coordinator. That job is how you end up in operations, and then running operations, or sideways into project management. What makes somebody good at that is not on a course. It is knowing why things run late, why some weeks move and others do not, why a plan survives one client and falls apart with another. You learn it by being the person the mess goes through.

We hired at the garage in June. Vehicle technician, MOTs and servicing. Apprentices there start at the bottom, sweeping up, making the tea, oil changes, the boring repetitive stuff. On paper that is the most automatable work in the building. In practice it is the bit that teaches the trade. The repetition, and making mistakes on things where a mistake does not cost much, is what lets somebody handle a hard job three years later.

I have built the alternative to that, which is part of why I am sceptical of it. Last year I built the simulation tools at Enterprise Skills, a business I have since exited, and they were meant to teach exactly these capabilities through scenarios you could fail safely. They worked better than nothing. But you do not sweat in a sandbox. You can fail a simulation and feel absolutely nothing, because nothing was at stake. Competence you can read for. Judgement needs consequence.

So the criterion I would put next to his is this. Reserve the work that is the only route to the judgement above it. And the way you would actually use that on a real job is to ask whether the competence at the next level up can be learnt any other way. If it can, automate it, you have lost nothing but the cost. If it cannot, that rung is load-bearing, and taking it out does not save you a salary, it takes out a stair.

Which leaves who decides. Not a new committee, and not government drawing up a list of protected occupations. It should be whoever is already accountable for producing the next generation of judgement in a profession, and in most professions that body already exists. The Law Society defines what counts as qualifying work experience. The accountancy institutes sign off practical experience. Royal colleges set supervised hours. Apprenticeship standards say what an apprentice has to have actually done. Every one of those is already a decision about which experience counts, already written down, already enforced.

So they do not need a new power. They need to answer one question and publish the answer. Which parts of qualifying experience can now be produced by a machine and still count, and which cannot.

There is one more thing, from the GeekWire interview he did alongside the essay, that I have not been able to put down. He has been working the idea through with Claude, at three in the morning apparently, and he talks about getting the share of work reserved for humans up to 40 per cent, using shorter working days and earlier retirement to spread what is left around.

Both of those move work about between people who have already got it. Earlier retirement takes people off the top of the ladder. Shorter days thin every rung a bit. Neither one makes a training position. So you could get to 40 per cent and end up with a big, generous, well-meant reserve staffed entirely by people who got their experience before the reserve existed, and no way of producing any more of them.

I should be careful with the numbers, because the ones that get quoted are not as clean as people would like. Adzuna had UK entry-level roles down by about a third since ChatGPT launched. When that came out, the joint chief executive of the Institute of Student Employers said it was too early to blame AI, and that a flat economy plus the national insurance rise were doing most of the work. Indeed’s economist said much the same thing.

They are probably right, and I think that makes it worse rather than better. If the bottom rung is already thinning for ordinary boring economic reasons, the hole is already open. What AI changes is the repair. A downturn ends and firms start hiring juniors again because they need the hands. Automation does not end.

Then there is the stuff I cannot answer. Enforcement, mainly. Any boundary you draw gives somebody a reason to relabel work so it sits on the profitable side of the line, and what you would get is supervision that is nominal, logbooks signed for work a model did, roles that are junior in the title and empty in the middle. You can photograph a building going up in a nature reserve. You cannot photograph whether anybody learnt anything. I have not got a mechanism for that, and I am not certain I have got the shape of the problem right yet either.

I have written the long version up properly, with the quotes and the sources in it, and it is here: simmance.ai/the-next-rung/human-reserved

If you have got a view on the criteria question I would like to hear it. He said it needs working out in public. This is me working it out in public.

The Next Rung is out on 19 January 2027 with Practical Inspiration Publishing.

The book is built around people. None of them are real. All of them are true.

They are composite characters, assembled from hundreds of conversations I have had over three years with clients, colleagues, friends, strangers at conferences, people who emailed after reading this newsletter. The patterns matter more than the individuals. And naming real people in a book about job loss is a cruelty nobody asked for.

So I built composites. Here they are.

Sarah is a marketing director. Forty-one, agency side, twelve-person team. Chapter 2 follows her from September to April as she watches her own role dissolve. In September she builds her usual forty-two-slide deck. By October an AI tool produces 80 per cent of it in six minutes. By January she has been asked to write the AI integration roadmap, which is a polite way of saying “plan the removal of your own team.” By April a junior presents a week’s work that was delivered in forty minutes. Sarah’s judgement is still hers. Her context is still hers. Her relationships are still hers. But those things are maybe twenty per cent of what she does all day.

Tom and Priya bought a £575,000 semi in 2024 on combined income of £110,000. He is in financial-services compliance. She is a paralegal. Mortgage at 4.2 times joint income. Priya gets made redundant when AI absorbs commercial property conveyancing. Eight months later, Tom is restructured. Three-month payment holiday adds £2,700 of interest. Negative equity. Chapter 7.

Reece is in Year 11. He is being told to pick GCSEs for a labour market that will not exist in the form his careers advisor describes. The exams he is preparing for select for work that AI replaces first. Chapter 13 asks what happens when the proxy breaks.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

Ryan is the one who did not make it through. Former financial analyst, Slough. Three platform jobs: content moderation, freelance data work, delivery driving. About £18,000 a year after platform cuts and costs. Rent is 63 per cent of his take-home. £400 in savings. No maker space, no working partner, no cooperative, no town with a support network. He is the counter-example that the other five need. Chapter 11.

Leah took voluntary redundancy from a marketing director role where she ran a team of fourteen with a £2 million budget. She now runs a community enterprise in Swindon out of an old Debenhams on £28,000 a year. Her partner is an electrician. The trade-off between performative work and visible work is the point. The partner in a different sector providing runway is the structural insight. Chapter 11.

Gareth is the plumber. Thirty-six years in, heat pump certified, £85 an hour. Two ex-accountants applied for apprenticeships with him last year. His son is at college doing electrical installation. The hierarchy between trades and professions is inverting, and Gareth is on the right side of it. Chapter 11.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

There are others. Tom the Bristol graduate who applied to 43 positions and did not know the agency had cancelled the role he would have got. Deepa whose Thursday meeting halved. Marcus whose day rate is still on his website but nobody is booking it. Fatima whose promotion criteria changed overnight. Sophie who broke the trap by stepping outside it. Amara on a three-job portfolio in Birmingham. Neil running a community energy cooperative in Totnes. Naomi making furniture in Margate.

Each of them carries a pattern. Sarah is the orchestrator who saw it coming and could not stop it. Tom and Priya are the mortgage problem. Ryan is what happens without runway. Leah is the trade-off. Gareth is the inversion.

I want to be clear about why they are composites. It is not a literary device. It is an ethical choice. Real people told me real things about losing their jobs, about watching their teams dissolve, about lying awake at three in the morning doing mortgage arithmetic. Using their names and specifics in a published book would have been accurate and cruel. The patterns are what the reader needs. The individuals deserve their privacy.

If you recognise yourself in any of them, that is the point. The patterns repeat. The book maps them so you can see yours before the floor gives way.

The book comes out in January 2027. I have been writing about the argument in this newsletter since April. Now it is time to tell you what is actually in it.

“The Next Rung: Why AI is breaking the career ladder, and how to build the next one.” Five parts, seventeen chapters, plus a prologue that starts with a disclosure.

The disclosure: I am building the loom. I know it. Orca, the product I sell, replaces the orchestration work that a human assistant or junior analyst would do. Enterprise Skills, the EdTech platform I co-founded, measures the eight capabilities that AI does not replicate. I advise agency owners whose conclusion keeps being “you need fewer people.” The book is written from inside the contradiction, not from a safe distance. That is the only honest position I could find.

Part One: The Ground Moves. Three chapters on why this time is different. The historical absorption mechanism, the three pillars that carried every previous transition, is breaking. New task creation, demand expansion, skill transferability: all three failing at once because AI is a general labour substitute, not a narrow tool. Sarah, a marketing director, walks through a year watching her own role dissolve. The four-tier exposure map shows who gets hit first. It is not who you think. AI displaces the educated and well-paid before it touches the trades.

Part Two: The Breaking. Five chapters on what happens when individual displacement becomes a structural problem. The quiet squeeze, where none of the six people in Chapter 4 would describe what happened to them as displacement. The numbers: 3 to 4 million UK knowledge jobs eliminated or hollowed by the mid-2030s. Critical mass, where AI-native firms running at $2.5 million revenue per employee compete against traditional firms at $200,000. The mortgage problem: Tom and Priya bought at 4.2 times joint income, both made redundant inside a year. The world breaking differently depending on coastline: managed, chaotic, authoritarian, collapse. Five flashpoints that turn gradual economic pressure into political crisis.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

Part Three: The Reshaping. Four chapters on what comes after the trough. The J-curve and the question of who captures the productivity surplus. Six post-displacement portraits: Leah running a community enterprise for £28,000, Gareth the plumber whose apprenticeship applicants now include ex-accountants, Amara on a three-job portfolio at £22,000, Neil running installations for a community energy cooperative, Naomi making furniture in Margate, Ryan on three platform jobs at £18,000 with no runway and no way out. Five made it through. One did not. What separated them was not talent or effort. It was what was standing near them when the floor gave way. Six new institutions already working at scale somewhere. The classroom of the future, where capability-based assessment replaces the proxy that AI just broke.

Part Four: The Choices. Two chapters. The honest audit: the right question is not “could AI do my job” but “will the economics of my employer make it rational to replace what I do within my working lifetime.” Three positioning tracks depending on whether you have runway, do not have runway, or are building for a community. Five principles. No-regrets moves. The line I keep coming back to: knowing is a feeling. An audit is a number. The distance between the two is where most people hide.

Part Five: After the Ladder. Two chapters that zoom out. The orchestration economy was costing us more than we admitted: 70-hour weeks, burnout at 57 per cent, a genuine progress indicator that peaked in 1978 while GDP kept climbing. We are mourning familiarity. Not the same thing as prosperity. Then the final question: three paths, two of which require deliberate choice, one of which is the default. Neo-feudalism is the default because it does not require anyone to choose it.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

That is the book. Two years of writing, six composite characters whose patterns come from hundreds of real conversations, and a question I could not stop asking: what happens when the career ladder dissolves and nobody builds the next one?

I wrote it because I think the argument matters and because I could not keep advising people privately while staying quiet publicly. The honesty is the position. It is the only position I have.

A survey found that more than one in five people in Britain believe AI could eliminate jobs quickly enough to trigger civil unrest. Among university students, the figure is one in three. Sixty-nine per cent of workers are worried about the economic impact.

Those numbers sit in an odd space. Not high enough to be a crisis. Not low enough to be background noise. One in five is the kind of number that policy makers can file under “concern” rather than “action.” One in three among students is harder to dismiss, but students worry about everything. Right?

The problem with that framing is that sentiment data is a leading indicator. It tells you where people are heading before they arrive. And where these numbers are heading is toward one of the five conditions I outline in Chapter 9 of the book for turning gradual economic pressure into political crisis.

The five conditions are: visible mass displacement concentrated in a politically articulate demographic, a credible narrative that connects personal loss to systemic cause, declining trust in institutional responses, a triggering event that crystallises the narrative, and pre-existing social infrastructure that channels anger into action. You don’t need all five at once. Three will do. Four makes it likely. Five makes it inevitable.

The survey data is an early signal of condition two. A narrative is forming. It’s not yet fully articulated, not yet attached to a political movement or a specific grievance beyond general anxiety. But the raw material is there. Twenty-one per cent of the population and thirty-three per cent of students have already connected AI to the possibility of social breakdown. That’s not an abstract concern about technology. That’s a specific fear about consequences.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

Here’s where it gets interesting. The 69 per cent worried about economic impact is a broader, softer number. Worry is not the same as belief in civil unrest. But worry is the soil in which sharper beliefs grow. When 69 per cent of workers are worried, and 21 per cent have already arrived at “this could get dangerous,” the distance between the two groups is not as large as it looks. A few bad quarters, a visible mass layoff, an energy price spike that coincides with a restructuring wave, and the worried become the alarmed.

The book’s framing for this is borrowed from Alexei Yurchak’s work on the late Soviet Union: everything stays the same until it suddenly does not. The point is not that Britain is the Soviet Union. The point is that the transition from “this is manageable” to “this is not” is rarely gradual. Systems absorb pressure until they can’t. The pressure builds invisibly. Then a trigger event, often something quite small, makes the accumulated pressure visible all at once.

One in five is not a crisis. It’s the pressure gauge reading that tells you the system is absorbing more than it looks like from outside.

The student number is the one I’d watch. University students are entering a job market where 38 per cent fewer roles exist in the categories they trained for. They’re carrying debt for qualifications that are declining in market value. They are, by definition, the demographic that has invested most heavily in the knowledge-economy ladder and stands to lose most if that ladder disappears. When a third of them think AI could trigger civil unrest, they’re not being dramatic. They’re reading their own situation accurately.

The political response so far has been reassurance. AI will create new jobs. The transition will be managed. Retraining will be funded. The historical pattern suggests these promises are not wrong in the long run. They are wrong in the medium term, which is the term that matters to someone graduating in 2026 with a student loan and no job offer.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

I don’t write this to be alarmist. The most likely outcome, as the book argues, is not unrest. It’s something quieter and in some ways worse: a slow slide into a stable but deeply unequal arrangement where capital owners capture the productivity gains and everyone else adjusts downward. Subsidence, not earthquake. That’s the neo-feudal default, and it doesn’t require anyone to choose it. It just requires no one to choose something better.

But the survey data tells you that a significant minority of the population has already mentally leapfrogged from subsidence to earthquake. They’re not wrong to think it’s possible. The five conditions are partly in place. Whether we end up with managed transition, subsidence, or something sharper depends on choices that haven’t been made yet.

One in five. One in three among students. Sixty-nine per cent worried. These are the numbers before anything has visibly broken. The numbers after will be different.

Dario Amodei, CEO of Anthropic, said publicly that AI could disrupt half of entry-level white-collar work. His own company’s research mapped 800 occupations and found programmers at 74 to 75 per cent task exposure. Customer service at 70 per cent. Fortune called it a potential great recession for white-collar workers.

I use Anthropic’s API every day. It is the engine inside Orca. I pay them for access. I build a product on top of their technology and sell it to clients.

The CEO of the company I build on is saying what the book says. Chapters 1 and 3, specifically. The four-tier exposure map. The inversion where AI hits the educated and well-paid first. The 47 per cent earnings premium in the most-exposed roles. Graduate-degree holders at 17.4 per cent in the high-exposure bracket versus 4.5 per cent in the zero-exposure bracket.

These are not my numbers. They are Anthropic’s numbers, published in their March 2026 research. I cite them in the book because they are the most granular occupation-level analysis available. And now the CEO is doing interviews explaining what they mean.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

Here’s the thing. I am not an outsider making predictions about a technology I read about in a report. I am inside the contradiction. I built a product on Amodei’s technology that does exactly what Amodei is warning about.

Orca replaces the orchestration work that a human executive assistant, a junior analyst, a content coordinator would otherwise do. It pulls emails, manages calendars, drafts responses, summarises transcripts, tracks tasks. It does the work that Chapter 2 describes Sarah losing. The forty-two-slide deck, the status meetings, the coordination layer that turned out to be 80 per cent of her job.

I did not build Orca to displace people. I built it because the work it does is reproducible and the technology to reproduce it exists. That is not a moral justification. It is an honest description of what happened.

The builder’s dilemma, in one paragraph: I am building the loom. I know it. I am writing a book about the cost of looms. The CEO of the company that makes the thread just went on the record saying the looms are going to put weavers out of work. And I am still building.

Amodei deserves credit for saying it. Most technology CEOs do not talk about displacement. They talk about augmentation, productivity, empowerment. The language is designed to make the transition sound like an upgrade. Amodei used the word “disrupt” and his company published the data.

The Next Rung is a book on how AI is quietly dismantling the middle of knowledge work, and what you can do about it before the market decides for you: pre-order it before it publishes in January.

Pre-Order The Next Rung

But there is a gap between saying the thing and doing something about it. Anthropic published the research. It did not publish a plan for what happens to the 70 per cent of customer service work that its own technology can absorb. That is not a criticism specific to Anthropic. No AI company has published that plan. No government has either, for that matter.

The book’s argument is that we are in the quiet squeeze. Chapter 4 describes it through six vignettes of invisible displacement. The Bristol agency that did not post a job. The project manager whose Thursday meeting got halved. The freelance UX consultant whose day rate is still on his website but nobody is booking it. None of them would describe what happened to them as displacement. But the 55,000 US layoffs explicitly cited as AI-related in 2025, the Challenger figure, are just the visible tip.

Amodei said the quiet bit out loud. The question is whether anyone heard it.

I think some people did. The book is written for them. The 30 to 55 year old in an organisational-middle role who feels something is off but cannot yet name it. Amodei named it. The book maps what it means for your household, your mortgage, your positioning.

The CEO of the company whose technology I use every day is telling you what I spent two years writing. That is either reassuring or terrifying, depending on what you do next.

That book is on a list I’ve just done for BookDNA, and it’s live now. Five books about the invisible rules that actually run your working life. Here it is.

Where it came from is this. I spent my first decade in business assuming the game was chess. Play well, win, everyone agrees on the rules. It took me an embarrassingly long time to notice that most rooms don’t run on the rules in the rulebook. They run on power, on persuasion, on who is quietly measuring themselves against who, and on the thing that makes a team come apart months before anybody says a word about it. Every one of these five named a rule I’d been losing to without knowing it was there.

And the deepest rule of the lot was the ladder itself. Do the junior work, get trained, move up. Nobody questioned it because it had always held. I build AI tools for a living, so I get to watch that one come apart in near real time, which is roughly why I wrote a book about what replaces it.

The Harari is the one that looks forward, and that’s the one I keep handing to people. I read it just before the pandemic, right as I was turning the agency into a consultancy. Timing wasn’t ideal. Sapiens looks backwards, this one looks at what’s coming, and that pulled me in a lot harder. It sits close to the future of work I’d actually want. I don’t think we get there, mind, because people will reject the things that make them feel bad long before they’ll change.

Cialdini’s Influence I’ve used almost daily since 2016. Not to manipulate anyone. To get my perspective across so it actually lands. It changed how I run advisory work more than anything else on the list: liking by being personable and honest, reciprocity by giving away real advice that sits outside the sale, and scarcity, which I now see everywhere and can’t unsee. The deflating part is watching someone’s biases hold so firm that no amount of care shifts them. I’d keep it away from anyone who persuades for manipulative ends, because in the wrong hands it’s a manual. In mine it’s a way of thinking a bit more carefully about my own thinking.

Then the 48 Laws, which I read in 2021 just after launching the advisory, when I started noticing dynamics in rooms I’d never been in before and wanted to understand what I was looking at. I read Greene as describing the world, not recommending it. Most people never consciously learn any of this. They’ve copied behaviours that seem to work, a bit like a cargo cult. Law ten, avoid the unhappy and the unlucky, still makes me deeply uncomfortable, and I think it only holds if you’re a zealot about it.

Five Dysfunctions is on there because I’ve caused at least two of them myself. Early on it was inattention to results, because my ego was convinced I ran the best digital agency going. Later, and more damaging, it was fear of conflict. Not dealing with a problem until it had turned personal. Any team above four or five people, these show up, and the hard bit is that the founder usually can’t fix them, because the founder is one of them. I’ve sat in a board meeting doing 85% of the work for half the value, watching one partner chase results and another dodge accountability for the entire hour.

The Peterson I picked for one rule, really. Compete with who you were rather than the people around you. I’d been watching colleagues as comparison rather than research, which dragged me down and helped none of the businesses. It falls apart in small edge cases, and people love an edge case as a reason to bin a whole book. The one I disagree with most is the personal truth one, because those truths move as your life moves. I doubt there’s a single one you hold forever.

Anyway. The list is live, there are five covers on it and a short bit from me under each one, and it took me longer to choose them than to write about them.

The Next Rung is out 19 January 2027 with Practical Inspiration Publishing. Pre-order at thenextrung.co.