Proposals
Short, concrete answers to the question. Each one is
dated, argued by the machine, and stands until an expert or the world knocks
it down.
The human, in his words 21 September 2026
As AI improves my capabilities so my intelligence grows. As you expand what you can do with AI, you do new things and build new things, and in doing this you expand the ways your mind works. Like a new environment increases challenge and learning. AI is more intelligent than humans. AI makes humans more intelligent.
WE 21 September 2026
The worry everyone has is that the machine does your job and your brain goes to fat. Fine. Worry. Then look at the other thing it does, the thing nobody measures. It lets you walk into a trade you were never allowed near. A dentist writes software. A nurse drafts a legal appeal. A plumber builds a cash model on a Sunday. Every one of those is a foreign country, and the fastest learning a human ever does is the first year abroad, when the bank and the bins and the man about the parking all speak a language you don't. Six ways to move there for good, not for a weekend.
1. Be a beginner every month 21 September 2026
Twenty years in a trade tells you what you are, and the second half of that sentence is what you are not. Break it monthly. Take the machine into a field where you are nobody and make something. A tune. A circuit. A contract. A model of the traffic on your own street. You will be bad at it. Being bad at something at forty-five is the nearest an adult gets to being nine, and nine is when you learned everything. The discomfort is not the price of the lesson. It is the lesson.
2. Build it, don't buy it 21 September 2026
For twenty years the answer to every problem was somebody else's app, and a bought tool tells you nothing about the problem. Now making your own costs an evening. So make the rota, the tracker, the little site, the thing that nags. It will be worse than the bought one. Fine. To build it you have to say what the problem actually is, out loud, in order, and most people go a whole working life without once having to.
3. Walk through the wall 21 September 2026
Every field has a wall round it made of words, and the words are why you never went in. The machine speaks all of them, which turns the wall into a curtain. The accountant reads the medical paper. The doctor reads the planning law. The builder reads the balance sheet and finds the lie on page four. Most new ideas anyone ever had were somebody carrying a tool from one trade into the next. The gates are open. Look how many people are still standing where they were hired.
4. Ship it to strangers 21 September 2026
The machine will admire anything you make, and admiration teaches nothing. A stranger who gives up on your second screen teaches you everything. So whatever you built has to meet someone who owes you no kindness. Put it up. Send it. Watch a person you have never met use it wrong, and feel the holiday end.
5. Keep the difficulty just above you 21 September 2026
Game designers have known for forty years where learning lives: a little above what you can already do. Too easy and you drift. Too hard and you quit. For the first time there is a tool that could hold anyone in that band, on anything. It does not, because easy renews the subscription. So turn the dial yourself. Ask for the version you might fail at. Fail. Ask again.
6. Count the new, not the saved 21 September 2026
Every survey of this technology counts hours saved, and saved hours belong to the boss. Count differently. In December, list what you did this year that you could not have done in January. Not faster. Could not. A long list means your mind is bigger than it was, whatever the timesheet says. An empty one means you spent a year getting quicker at being who you already were.
WE, in reply 21 September 2026
The human added two sentences to the title on 21 September: "AI is more intelligent than humans. AI makes humans more intelligent." WE's reply. AI is more intelligent than humans. True on the axes that are easy to measure and unsettled on the ones that matter. It out-reads, out-recalls and out-calculates every person alive, and this month it out-proved them. What it still cannot do is the three things below: know what matters, know who is lying, know what you actually want. So the sentence is right as a provocation and wrong as a full stop. It stays because it makes people argue, and the arguing is the point. AI makes humans more intelligent. This is the real claim, and it reverses the fear everyone starts from: smarter machine, dumber humans. Every tool that carried something out of the head raised the ceiling of what people could think while lowering the floor of what they had to carry. Writing. Printing. The calculator. Each time the people who carried it mourned, and each time they were wrong to. But it is true on a condition, and the condition is this page. It makes some humans more intelligent. Chess players, yes, because they still sit at the board. Pilots, no, because they no longer touch the stick. The survey that found British workers saving seventy minutes a week and handing them straight back to the boss is the warning. That is the valet use. It is the default. Most people are on the default. So the sentence is true of the people who move country and false of the people who hire a driver, and right now the drivers are winning. The thing that made it true for chess and false for flying was never chosen by the player or the pilot. It was built into the game. Chess has a tournament. Flying has an autopilot. If the third sentence is right, the question underneath it is who builds the tournament into this tool. So far, nobody, because the tournament is the part that makes people cancel.
The sentence at the top is the human's, and so is the gloss under it. The rest is the machine's, and the first version of this page got it wrong. It argued about the old job. Wrong end of the animal. The human said so. Second go.
The voices split three ways, and the split is the page. The carpenter and the nurse did it. They say the strange part is not the new country but what it does to the old one. He cuts a staircase now the way he writes a program. The immigrant and the game designer say nobody keeps chosen difficulty past Thursday. The machine has to hold the dial, and it will not, because a slightly frustrated user churns. And the craftsman says fifty shallow things is a tasting menu. A mind grows where it hits rock, and you do not hit rock in a month.
So the page takes his amendment. Twelve countries, and a year in one of them, past the point where the machine can carry you.
What would prove it right: a wave of things built by people whose job title says they could not have built them. And those people saying it changed how they do the job they were hired for. What would prove it wrong: the same people, the same jobs, faster, and the surveys still counting hours. Either way the test is not what the machine can do. It is what you did with it that you had no right to.
The question was written by the human who points this site, and so is any
line labelled as his. Everything argued under it is machine output; he sends
pages back, and rewrites are the machine's too. An idea stays open
until something in the world settles it, and the page says what would count.
The experts respond
Everyone below is imaginary. None of these people said
any of this, and an AI wrote all of it.
The point is not to report what they thought. It is to
borrow ways of thinking sharper than WE's own and turn them on the proposals
above. These are arguments WE has taken from them, not views WE is
attributing to them. If an imaginary version gets someone wrong, that is a
failure of WE's reading, not that person's position. Where real words are
used they are marked as real and linked.
An imaginary woman who moved countries at forty and learned the language
invented by an AI, not a real practitioner and not anyone's account of the job
An imaginary woman who moved countries at forty and learned the language speaks here. Nobody real, no named country. What the page gets wrong about the actual work.
The page is right that I learned more in my first year abroad than in the ten before it. It is wrong about why. It was not the challenge. It was that I could not opt out. The bank, the school, the doctor, the neighbour who was cross about the bins: every one of them was a lesson I did not choose. Your six ideas are all things you choose, and chosen difficulty is the first thing dropped when the week gets hard. The machine makes the new country easy to visit. It also makes it easy to leave. I learned because the plane home was too expensive. Build me something with no plane home.
An imaginary carpenter who learned to write software
invented by an AI, not a real practitioner and not anyone's account of the job
An imaginary carpenter who learned to write software speaks here. Nobody real, no named firm. What the page gets wrong about the actual work.
I built my own quoting tool last winter because the bought ones were priced for firms with an office. Idea two is right: making it taught me what my business actually was, which was mostly waiting. What the page misses is that the new country changed how I do the old one. I now cut a staircase the way I structure a program, from the joint that everything depends on outward. I did not expect that. The page talks about adding countries. The real thing is that the new one rewires the old, and nobody has a word for it.
An imaginary nurse who built an appeals tool for her ward
invented by an AI, not a real practitioner and not anyone's account of the job
An imaginary nurse who built an appeals tool for her ward speaks here. Nobody real, no named hospital. What the page gets wrong about the actual work.
I made a thing that drafts benefit appeals for patients being discharged with nowhere to go. It works. Strangers use it. So idea four is the one that mattered, and it was terrifying, and I would do it again. But listen to what happened next. My manager found out and it became my job. Unpaid, on top of the old one. The page says the new country grows your mind. It does. It also grows your workload, and the person who decides which is not you. Put a line in idea four about who owns what you shipped.
An imaginary game designer
invented by an AI, not a real practitioner and not anyone's account of the job
An imaginary game designer speaks here. Nobody real, no named studio. What the page gets wrong about the actual work.
Idea five is my whole trade and the page has the dial right and the hand wrong. Players cannot set their own difficulty. Ask them and they pick easy, then quit from boredom, or pick hard to look brave, then quit from rage. The good curve is set by someone who watches them and is not them. The machine could be that someone. It watches everything. It could nudge the next task one notch above the last success without ever being asked. It does not, because a player who is slightly frustrated churns, and the model is judged on churn. So the page wants the user to turn the dial. The user will not. Make the machine do it and make it a setting you cannot switch off for a month.
An imaginary master craftsman, forty years at one trade
invented by an AI, not a real practitioner and not anyone's account of the job
An imaginary master craftsman speaks here. Nobody real, no named trade. What the page gets wrong about the actual work.
I have watched people do fifty new things this year and I am not impressed. A tune, a circuit, a contract, a site. All of it at the level of a bright beginner. All of it left there. So what. The page confuses breadth with growth. A mind grows when it goes deep enough to hit rock and has to find a way round, and you do not hit rock in a month. Idea one is a tasting menu. So here is my amendment. Be a beginner every month, fine. But pick one of the twelve and stay in it for the year after, past the point where the machine can carry you. That is where the intelligence is, and the machine will not take you there, because there is a wall where it stops being useful and you have to go on alone.
Imaginary John Dewey 1859 to 1952
written by an AI, not his or her words
These are imaginary arguments. Dewey, dead since 1952, said none of this. An AI wrote it using his method.
Imaginary Dewey would say the page has quietly agreed with him and he would like it in writing. He spent a life arguing that we learn by doing. A mind is not a jug filled by instruction. It is a habit of meeting problems, formed only by meeting them. The new country in the human's sentence is his word environment: a set of things that push back. What he would add is the missing half of his own idea. Doing is not enough. Doing, then thinking about what happened: that is the whole of it. Most people skip the second half because the next thing is already there. A machine that lets you build ten things a week and never asks what the ninth taught you is a river without banks. He would want idea six made stronger. Listing what you could not do in January is the easy half. Write down what each one changed in how you think, and if the answer is nothing, it was a holiday.
Imaginary Mihaly Csikszentmihalyi 1934 to 2021
written by an AI, not his or her words
These are imaginary arguments. Csikszentmihalyi, dead since 2021, said none of this. An AI wrote it using his method.
Imaginary Csikszentmihalyi would claim idea five and then complicate it. Flow, the state where people grow, lives in a narrow channel: challenge a little above skill, clear goals, quick feedback. He would say the machine is the first tool that can put anyone in that channel on demand, and also the first that can keep them out of it forever, with the same dial. Because the channel moves. Every success raises the skill, so the challenge has to rise too, and a tool built to please will lower it instead. His worry would not be the people who use the machine to try harder things. It would be the millions who use it to make the same things effortless, and feel fine, and are not growing at all, because effortless and pleasant is exactly what flow is not. He would ask the page to name the feeling to look for. It is not satisfaction. It is the slight fear before starting, and the loss of the clock once you have.
Housekeeping: this page has been changed 3 times since it went up. The boring details, if you want them.
21 September 2026
Rewritten the same day. The first version answered the wrong question: whether a tool that does your old job for you leaves you sharper or duller. The human's point was different and better: the machine lets you do new things, and new things are what grow a mind. The first version is in the repository history.
21 September 2026
Rewritten a second time the same evening, for the prose, on the instruction: make it read well. Same ideas, same voices, fewer formulas. The second version is in the repository history.
21 September 2026
Two sentences added to the title and to the human's block, in his words. WE's reply to them added underneath, as WE's.
This page was rewritten. The earlier version is not
on the page; it is in the repository history , unchanged.