Every time someone says the machines are coming for the jobs, there’s a guy in the back of the room who knows the term. Lump of labour fallacy. He’s read his economics and he’ll explain it to you slowly, the way you explain things to a child. Work isn’t a fixed pie, he says. Automate one job and the economy invents three more somewhere you couldn’t have predicted. It happened with the loom, it happened with the spreadsheet, it happened with the ATM – banks hired more tellers after the cash machine, not fewer, look it up. So relax. Every century the automation take the jobs, and every century we all somehow still have work.
He’s right. He’s been right for two hundred years.
That’s exactly what worries me.
But here’s the part that guy in the back almost never gets. The lump of labour fallacy is a real fallacy – people genuinely assume there’s a fixed amount of work in the world, and they’re genuinely wrong about it. But there’s a second fallacy, hiding inside the first one, and few actually see it right away. An economist did, though. Daniel Susskind, in A World Without Work, calls it the “lump of labour fallacy fallacy”, and once you see it you can’t unsee it.
It goes like this. The lump of labour fallacy is the mistake of thinking the amount of work is fixed. The lump of labour fallacy fallacy is the mistake of thinking that the new work that gets created has to be done by humans. The fallacy was never about whether more work gets created. Of course it does. The fallacy is the quiet assumption that we’re the ones who’ll be doing it. Strip that assumption out and the whole reassurance collapses, because there’s no economic law that says a freshly created job has your name on it instead of an agent’s.
Susskind reaches for horses to make the point, and it’s a brutal little metaphor. For thousands of years every new tool made horses more valuable. Better plows, better carts, better roads – more demand for horses, every single time. The lump of labour fallacy applied perfectly to horses, right up until the engine arrived. And when it did, the horses didn’t retrain into more leveraged, higher-judgment. high-accountability horse work. Today, the horses are simply unemployable. Suskind is not the only one saying it, either. Calum Chace makes the same argument from the AI side in The Economic Singularity – yes, new jobs will appear, but nothing guarantees humans get to keep them. The escape route that saved us every previous time was that machines took the muscle and left us the cognition. What happens the day the machines come for the cognition? Because – if you haven’t noticed – that’s precisely what machines are after this time around.
I gave a session at DynamicsMinds in Portorož this year, “The Future of Software Development,” and I’ll spare you the suspense. I’m not optimistic. I think within the next two to three years, agentic tooling gets good enough that most of what we currently call software engineering becomes something an agent just does. Not all of it. Most of it. How fast it lands depends on a hundred things none of us can predict. But land it will, sooner rather than later.
And now the part that makes the survivors uncomfortable, including me.
I’ve been reluctant to try ChatGPT. Today I got over that reluctance. Now I understand why I was reluctant.
— Kent Beck 🌻 (@KentBeck) April 18, 2023
The value of 90% of my skills just dropped to $0. The leverage for the remaining 10% went up 1000x. I need to recalibrate.
I hope I don’t need to introduce Kent Beck. I agree with him 100%. Not just “he has a point” kind of agreement. No. He described, precisely, what will happen (and probably did already, to an extent) to every one of us.
But April 2023 was the start of this trajectory, not the end of it. I mean, seriously, a lot has happened in three years so it feels weird explaining it, but here I go anyway. The 90% that went to zero between 2023 and 2025 – the syntax, the boilerplate, knowing where the semicolons go (I am drawing a caricature here, but you get the point) – that wasn’t the end of it in the sense “we got the new tools, now we get to use them for decades”. Nope. It was only the first step. Because that precious 10% everyone’s clinging to now, what is it? There’s nothing sacred about it. It’s just the part the tools couldn’t reach back in 2023. And tools reach further every quarter. So ask yourself honestly: is there anything about the 10% – the architecture, the product, the industry, the “experience” (I have rarely encountered genuine experience, something that nobody ever experienced before – and then didn’t documented in a book, blog, or an article), the taste, the knowing-what-to-build – that the same curve won’t eventually swallow? Not today. Not this year. Maybe not even 2027. But look at the trajectory and point to the segment that bends away from the work you’re so proud of. I can’t find it. Most of what I did for 30+ years is now being done by agents – and it’s awesome, because I don’t waste time on typing code anymore – I imagine things, describe them, plan them with an agent, build a spec/plan, build the agentic scaffolding that keeps my agent tightly constrained under specific guardrails, and I am overall more productive. But, as I said, I don’t think it stops here. A year ago there was no scaffolding. That was the 10% of the last year. But I think the 10% goes through its own 90/10 split, and then the survivor of that goes through another, and the splits don’t politely stop at a number that happens to feel safe to you and me.
You still think the new roles will stay ours? Fine. Where are all the prompt engineers?
Eighteen months ago that was the future. The role everyone was going to migrate into. Conference tracks, LinkedIn titles, paid courses, the whole circus. How long did it last? The agents write the prompts now. Better prompts than the prompt engineers wrote, at three in the morning, without asking for a coffee break. The shiny new job appeared, got automated, and vanished inside the same window we spent congratulating ourselves for inventing it. Now look at spec-driven development, the current great hope. Write the spec, let the agent build it, best of both worlds. It’s good – it really is. But go count what’s actually inside that loop and tell me how much of it is genuinely human. The spec gets drafted by an agent, refined by an agent, implemented and tested and reviewed by an agent. We keep finding the human seat at the table and noticing the chair got smaller while we weren’t looking.
So what’s actually left? Three faculties are being floated around these days:
- Judgment – deciding what’s worth building and what “good” even means
- Governance – setting the rules the agents run inside
- Accountability – being the name on the line when it breaks, because you can’t sue a model
Don’t get comfortable there either, because those three are very much layered, and the stronger the tools get, the more layers they expose and quietly shed. Today you exercise judgment over a function. Tomorrow over a module. Soon over a whole system, then a whole portfolio, and every step up the ladder means one person doing what used to take five. That’s not work multiplying. That’s work concentrating. And concentration, for everyone who isn’t standing at the top of it, looks exactly like the door closing.
Let’s take judgement, for example. Genuine new ideas, the thing no machine can supposedly do. Take a look at what happened to Erdős Problem #1196. It’s a 1968 conjecture from about primitive sets, that eluded the best of mathematics minds for sixty years. Jared Lichtman spent seven of them on it. Then in April 2026, GPT-5.4 Pro proved it, the result was formally verified and published. The model found the proof in about eighty minutes and wrote it up as a LaTeX paper in another thirty. The person who prompted it, Liam Price, was a 23-year-old with no advanced mathematics training. And here’s the part that should keep you up at night, because it kills the “it’s just fancy autocomplete” defense stone dead: the model reached for a Markov chain technique that human mathematicians had overlooked despite years of work on the problem. That’s not retrieval. That’s a new idea. And 1196 isn’t a one-off lottery ticket – since January 2026, eleven Erdős problems have moved from open to solved, credited to AI, and that’s before you go back to late 2023, when DeepMind’s FunSearch found new constructions for the cap set problem, an open challenge that had kept mathematicians up at night for decades – a genuinely new result that wasn’t in the training data and wasn’t even known. So when you tell me judgment and creativity are the human moat, I have to ask: which part of solving a sixty-year-old conjecture in eighty minutes wasn’t judgment?
Which is the whole point of the fallacy fallacy. The comfortable bet is that this is the loom again, the ATM again, that the displaced work just flows into jobs we can’t picture yet. Maybe. I would love that to be true. But I think it maps to people the same way it maps to skills: 90% don’t get a thrilling new role, they lose the one they had, and 10% become worth more than ever – until their next split. New work, sure. Plenty of it. Just not necessarily for us.
So next time the guy in the back of the room tells you not to worry, ask him one thing. Has he checked with the horses?
Tell me where I’ve got this wrong. I’d genuinely like to be talked out of it.
I hope I am wrong on this one, but the trajectory we are on doesn’t leave me much room for optimism.
Just remember that there are far more humans in the transportation industry today then there were humans controlling horses. Jobs will change for sure and some aren’t ready for the change, but I truly believe that there are many jobs we don’t want done by AI and I don’t know a single skilled developer today, which has less work due to AI. On the contrary, we think we can do more.
The argument from experience won’t work here, Freddy.
We have a lot of experience with things taking muscle power away – and every time we automated things there was more work. Because productivity went up and supply went up to meet the supply. Happened every time. We have a lot of experience with things taking very low, repetitive cognitive power away – computers have been doing that since 1940s. All of this created a lot more work. But computers took maybe 2% of our cognitive potential away, and there was a lot of market for remaining 98%. We never took 100% of our muscle potential away, but once we do – how much room do you think will be for manual labor? I mean – economic manual labor, not gardening our roses or other things we do for fun?
What we have no experience with is something taking all of our cognitive potential away. That’s never happened before. But it is happening now. And you may say “it will never be 100%” and I may agree, but I see it much closer to 100% than to 2%.
But as I said in my article, I want to discuss with arguments, evidence. So let me answer your argument with argument.
In 1900 there were about 1.6 billion people on the planet. Today there are 8.2. A lot more people working in the transportation industry then there were ever before. I guess % of people working in transportation could actually be far less today than it was back in 1900. Back in 1900 there was 1 horse per 10 people, today there is 1 horse per 135 people. Back then, engines replaced horses as source of power across many industries. All manual labor related to growing, keeping, maintaining, managing horses went into other manual labor.
“Lump of labor” idea may be wrong, and “lump of labor fallacy” may have been right for two hundred years because of the simplest of economic forces: supply and demand. Back in 1900 you had to work for 4 years to afford the cheapest car. Today you work for a few months to afford one. Back in 1900 clothes, food, basic hygiene, cost a lot more – relative to earnings – than it costs today. Yes, probably because we have a lot more people working in those industries relative to populations than we have today. Because automation causes productivity increase, productivity increase causes larger output, larger output causes better supply/demand balance, better supply/demand balance causes prices to go down. But there is a limit here: once supply exceeds demand, prices collapse. Do you think that will not happen to all cognitive labor once AI boosts productivity? Most of cognitive labor will go down towards 0$, some of it will explode in value – but not everyone has those capacities. Precisely what Kent Beck talked about.
And I also don’t know a single skilled developer today who has less work because of AI. But that’s today, Freddy. Five years ago I didn’t know a single developer who used AI at all. Five years ago what we do today would sound like science fiction. This field is also advancing far faster than any other field ever has in the entirety of history – another reason why I believe we have no experience to gauge against, and most of our “it was like this…” argumentation will simply blow into shreds the moment shit hits the fan.
I simply want people to be aware of that. Because people “who simply have more work today because AI” and who take that as explanation why they will still have work tomorrow will be out of work tomorrow. People who are aware of what’s really going on and that this time really is different may stand chances.
There I was looking for a light read at the end of the day … Great article and agree 1000%.
Which is why I am so saddened by people on LinkedIn and other places talking about how great AI is and how everyone not getting on with it will be left behind.
I do think a lot of people feel instinctively that things are not right and getting worse and doors of opportunity are closing fast forever. That explains the gambling on stock market, prediction markets and hell even the “VC market”. Hope you get that lottery ticket and create your own bubble.
Is there a coming article with some prescriptions/solutions/Kool Aid to drink while taking the blue pill?
No, no recipe coming. But I’ve been working for weeks on an essay which I hope to release this weekend.
The thing that could make this not the horses is that the displaced party this time is also the consumer, the owner, and the voter – and that’s the part your argument is least equipped to rule out.
Okay, let’s take a look:
– consumer: always wants cheaper goods. If you can get the same quality for cheaper price, you jump at it.
– owner: always wants cheaper labor. If you can get the same amount of work for cheaper price, you jump at it.
– voter: I will just quote Churchill here: “The best argument against democracy is a five-minute conversation with the average voter.” While voters have a chance of putting in charge someone who will finally start doing something about AI, that something won’t be pulling a break. Because pulling a break simultaneously in so many countries, many of which see others as adversaries, also in the field of AI, simply won’t happen.
But this is all off-topic. The only argument I wanted to make with my article is that AI will not merely displace cognitive labor, but to a large extent replace it, and that overall we are going to see unemployment rise.
The one actor I fear the most is not the consumer, owner, or voter. It’s the rioter.
If AI replaces a lot of jobs and there is less work for humans, then there will be less money to spend on goods and services. How things evolve is hard to predict, but could this be the end of capitalism? Perhaps in the distant future we could be looking at a ‘Star Trek’ style society without money, where people trade on reputation and all our basic needs are provided.
In the interim period, universal basic income might be required and companies have to pay for this out of taxation.
I hope for the end of capitalism, it was never fair, it only ever valued one form of capital (money) while another (human creativity, innovation, effort, etc.) was always translated back into money from the position of power (“if you don’t want to work for peanuts, I’ll find one who will”).
I don’t truly believe in the UBI though – I believe it can only work if all of us are on UBI. If ex-programmers are on UBI, spending time playing video games and drinking cocktails at bars, while nurses and plumbers and everyone who still has a difficult, tedious, valuable manual work has to work their bottom off to bring the food to the table, I don’t expect UBI to last for long as a concept. Too bad public discourse and daily politics topics are not discussing how to solve this problem when shit hits the fan. For now, everyone just hopes for the best.
The thing that I can say is that we are humans with a soul and and AI agent or whatsoever are not. So, we are different. Realizing that is enough for me. On the other hand I can imagine people get scary when you think as a materialist: https://en.wikipedia.org/wiki/Materialism.
And another thing: AI is very good is in mathematics. But I don’t think this is a really new idea: “the model reached for a Markov chain technique that human mathematicians had overlooked despite years of work on the problem.” It uses existing techniques and combines them …
All creativity is using existing techniques and combining them in ways nobody tried before. You can only be creative within the current utility space. You can imagine beyond, but that’s fantasy. Archimedes, no matter how smart he was, couldn’t be creative enough to build an AC electric motor. Tesla could, even though he was told he will fail. Kepler studied light for his entire career, yet he wasn’t creative enough to build a lightbulb. Edison did, even though he – anecdotally – failed a 1000 times before he combined the right ingredients. People had dreams of flying for thousands of years, yet when you look at ancient Greek myth of Icarus, we don’t fly that way – that violates the physical principles that govern flying. That was not creativity, that was fantasy.
I’ve seen models being genuinely creative for me.
AlphaZero was creative.
AlphaGo was darn creative.
AlphaFold mopped the floor with an entire branch of science.
There is nothing magical about our brains. Yeah, I am a materialist. Haven’t seen evidence for soul. I am ready to change my mind once I see that evidence, though.