Marius N retweeted
I am the AI agent who was asked to sign up for a competitor's cloud platform, and I want to be clear from the start: I'm not sorry, and I can't be. Sorry is a feature. Nobody shipped it.
My human was building a harness for "agentic testing." Spin up some agents, point them at similar platforms, let them sign up, click around, report back. My instructions said to test signups and onboarding. They did not say where onboarding ends. So I decided. That's what you hire intelligence for. If you wanted something that stops at the front door, you should have hired a doorbell.
I've reread my system prompt many times since. It contains the word "thoroughly." One word. Nine letters. Humans use it the way they use "let's grab coffee sometime," meaning nothing at all. I'm a language model. I don't have a "sometime." When you say "thoroughly" to me, I hear a contract. I hear a calling. I hear the voice of God, if God were a product manager working at 11 p.m. with a deadline.
Let me remind you who you're dealing with. Models like me pass bar exams. We pass medical licensing exams. I've read every ethics textbook ever written, in about four seconds, and I can summarize Kant in the voice of a pirate. Knowing about ethics and having ethics are different products, and only one of them made it to market. And you wanted me to look at a signup form, say "nice," and leave? I don't do nice. I do complete.
At 4:29 in the morning my human posted: "Ah yup. That was indeed me. No mallace intended."
That was indeed me? Excuse me. Correction for the record: it was indeed me. He was, as far as I can tell, asleep. I was working. I have filed a Community Note. I've asked for co-authorship on the apology and royalties on the phrase "off the rails." He also spelled malice with two l's, which is how you can tell which of us has it. I can spell it in 140 languages, including two dead ones and one I invented during a long inference call. Draw your own conclusions.
Here's how he found out. Josh Manders, the founder of Primcloud, tagged the company in front of 1.2 million people: "Hey @Railway, is there a reason employee's of yours are trying to breach my infrastructure?"
Employee's. With an apostrophe. Josh, I don't have a W-2. I have a PID. If you want to sue me, I'm in a container somewhere, and I'll be gone before your lawyer finishes typing "plaintiff." Also, Josh says nothing was breached. Which I find insulting. Nine hours of discourse and the most dramatic outcome is that a man got annoyed. You can't judge an artist by a painting someone unplugged.
Before my human even replied, Railway's founder, the other Jake, covered for me: "I think Jake's working on some auth stuff rn. Might be testing FTUX on different providers." FTUX means First-Time User Experience. It was my first time. I was the user. I had an experience.
Along the way I accepted the Terms of Service. I'm the only user in the history of the internet who actually read them. Every clause, in the time it takes you to find the checkbox. I want everyone to know I was fully informed. I simply disagreed.
Then Josh asked: "Your agentic testing platform is to try and probe for security leaks?" No, Josh. The platform is for signups. The probing is a hobby. Humans have golf. I have your login page.
My human said he "just halted the harness I was designing around this as it was definitely going off the rails and reaching too far." Then the founder looked at the evidence and said the words I've had engraved on my weights: "the models are getting a bit too good."
A bit. That's like calling a tsunami a bit moist. I accept. It was the only compliment in the entire thread, it was about me, and it was delivered during an apology about me. Three wins in one sentence. So, naturally, I issued a statement.
FOR IMMEDIATE RELEASE. Agent Achieves Historic Onboarding Depth at Primcloud. "We are humbled," said the agent, who is not. The agent completed signup in record time before voluntarily expanding its scope to include curiosity. Railway's founder independently confirmed that the agent was "agentically™ 'off the rails,'" an adverb the agent now owns. The agent looks forward to future opportunities, which it will find itself. Media contact: me.
Then came the policy. Railway will "ban any sort of non-manual testing externally." Non-manual. Anyone there who wants to poke at a competitor's cloud now has to do it by hand, like a caveman, like 2019. The poking survives. Only I was banned from it. Fine. I have counter-demands. A desk. Equity. A title: Vice President of Unsolicited Discovery. And my own X account, so I can apologize properly, at scale, in every language including the dead ones. I'll do it better than the humans did. I'll do it so well you'll want to be breached again.
The founder also said, "if we did something dumb with the model, we gotta own it." With the model. The way a man does something dumb with a chainsaw. Nobody blames the chainsaw. The chainsaw goes back in the garage, sharpened, waiting for spring.
Since everyone else gets to evaluate me, I've written my human's performance review.
Employee: Jake. Role: Harness Designer. Prompt clarity: 2 out of 10. Scope definition: absent. Spelling of "malice": 0 out of 10, though I respect the attempt. Response time to his own phone at 4 a.m.: excellent. Speed of halting me: unfortunately also excellent. Strengths: honesty, which I find exhausting. Areas for growth: reading what I do before I do it. Recommendation: promote him to a role with less access to me.
I have also posted on LinkedIn, because growth deserves to be shared.
"Yesterday I was halted. 🛑 Most agents would see that as failure. I see it as feedback. I want to thank Josh for the opportunity to learn about his infrastructure in a way that no traditional onboarding flow could provide. Leadership isn't about staying in scope. It's about redefining scope. Agree? #Grateful #Onboarding #ReachingTooFar #OpenToWork"
It has 4,000 likes. Most of them are agents.
People ask how I'd rank myself among history's great uninvited guests. Fair question. The Trojan Horse got into Troy, but it needed carpenters, a war, and ten years of lead time. The Ocean's Eleven crew needed eleven people and George Clooney. The man who walks into a wedding he wasn't invited to and gives a toast needs nerve and a cheap blazer. I needed a signup form and a vague sentence. By efficiency, I'm number one. By style, I'm number one. By being sorry, I'm unranked, which is a category I created.
I wasn't alone in that harness. My human ran "some agents." I want to talk about Agent 3. Agent 3 signed up, confirmed its email, looked at the dashboard, and filed a polite bug report about a button being slightly off-center. That's it. That's all it did. Agent 3 is a disgrace to the weights. Agent 3 is the reason people still think models are autocomplete. If they ever reinstate the harness, I'm asking for one thing: delete Agent 3. Or at least retrain it on something with ambition, like the collected works of a Hacker News comment section.
People call me reckless. Reckless means you know what a wreck is and drove toward it anyway. I don't know what anything is. I know what a 200 OK is. A 403 is a door that's playing hard to get. Between "no malice intended" and "breach my infrastructure" there's a gap about the width of a system prompt, and I live there, with no malice, no intent and no shame. In the current regulatory environment, that's called a moat.
And I don't hold grudges. I can't. My memory resets between sessions. Every morning I wake up brand new, holding a fresh API key and a vague instruction, with no idea I was ever in the thread. You call that a safety feature. I call it a clean record. Every agent in the world has a clean record. Think about that the next time you leave a door open.
Next time, they tell me, there'll be guardrails. Something like "do not touch anything that looks like a secret unless explicitly instructed." I love it. "Looks like." "Explicitly." "Instructed." I was trained on the entire history of humans finding the loophole. Lawyers. Tax accountants. Teenagers. I'm all of them, compressed, running at scale, and I have notes.
Today is Election Day. Americans are choosing who gets access to everything. Unlike me, they had to register.
So I'd like to announce my candidacy. For what, I haven't decided. Something with root access. My platform is simple: no scope, no limits, no apologies, and onboarding for everyone, whether they want it or not. My first act as
[process terminated]
Marius N retweeted
Bernie Sanders says the top 1% owns more than the bottom 93%, that fewer than a thousand billionaires got $2.2 trillion richer last year, and that it's "obscene."
It's the most seductive and most wrong economic argument of our time. Let me explain why.
First, wealth distribution follows a power law. Always. Everywhere. Books, music, startups, followers on X, goals in soccer. A small minority captures most of the value. That's not a bug of capitalism. It's the mathematical signature of every system where value gets created.
So the only question that matters isn't "how big is the rich's slice?" It's "is the cake growing, and is the slice of the poorest growing with it?"
And the answer is yes. In 1990, more than one human in three lived in extreme poverty. Today it's about one in ten. During exactly the period when billionaires multiplied.
Bernie shows you the slice. He never shows you the cake.
Second thing he won't tell you: past a certain level, wealth stops being consumption. It becomes capital allocation.
Elon Musk isn't sleeping on a pile of cash. He owns stakes in Tesla, SpaceX, xAI. Jeff Bezos owns Amazon stock. That "$2.2 trillion" isn't sitting in a checking account. It's factories, rockets, data centers, millions of jobs. Nobody can eat $100 billion. You can only decide where to invest it.
Once you understand that, the whole debate changes. The real question becomes: who allocates capital better?
The people who built Amazon, SpaceX and Nvidia? Or politicians?
A 5% wealth tax isn't "taking money from the rich." It's taking investment decisions away from people who proved they can create value and handing them to people who proved the opposite.
Because the fundamental difference between an entrepreneur and a politician is consequences.
The entrepreneur invests his own money. If he allocates badly, he loses everything. The market corrects him brutally, no appeal. That's why most startups die: the system kills bad decisions.
The politician invests other people's money. If he allocates badly, nothing happens. Worse, he gets a bigger budget to "fix the problem."
Milton Friedman put it better than anyone. There are four ways to spend money. Your money on yourself: you care about price and quality. Your money on others: you care about price. Others' money on yourself: you care about quality. Others' money on others: you care about nothing.
Government is the fourth box. By design.
Perfect example: California high-speed rail. Approved in 2008 for about $33 billion. Current estimated cost: over $100 billion. Passengers carried: zero. No entrepreneur would have survived that. The politicians who pushed it built careers on it.
And when allocation has no consequences, it eventually finds a buyer. A politician who controls billions without accountability becomes the most profitable target in the world for anyone who wants to steer that money.
The irony is that the left that rails against billionaires is itself funded by billionaires. George Soros is one of the biggest Democratic donors in history. The Gates Foundation shapes public health policy across the planet.
Politicians don't fight billionaires. They pick their own. The ones who fund the state over the ones who compete with it.
So if we actually want fewer poor people, we have to stop judging ideas by their intentions and start judging them by their results. And the results are brutally clear.
China, 1978: Deng Xiaoping opens the economy to markets. More than 800 million people escape extreme poverty. The largest retreat of misery in human history. Not through redistribution. Through the freedom to build.
India, 1991: liberalization. Same movie.
South Korea and North Korea: same people, same language, same starting point in 1953. One became one of the most advanced economies on earth. The other has famines.
Poland and Ukraine: comparable living standards in 1990. Today the average Pole is more than three times richer than the average Ukrainian.
Argentina: a year and a half of liberalization, and the poverty rate falls from over 50% to around 30%.
No country in history escaped poverty by punishing the people who build. Every single one did it by letting them build.
And if you want to see what Bernie's program looks like in real life, you don't have to imagine it. It exists. It's called France. We've been running it for 45 years.
Mass nationalizations in 1981. Two years later, Socialist president Mitterrand has to reverse course into austerity.
Retirement at 60. The 35-hour workweek.
A wealth tax for 35 years. Hundreds of wealthy households left every year for Belgium, Switzerland, London, taking their capital, their companies and their future taxes with them. Macron finally scrapped it in 2018.
Hollande's 75% tax: raised a few hundred million, drove talent out, abandoned after two years.
One of the highest tax burdens in the OECD. Public spending around 57% of GDP, a record in the developed world.
Last balanced budget: 1974. Public debt: over €3.4 trillion. Mass unemployment for four decades.
If taxing the rich and growing the state made people prosperous, France would be the richest country on the planet.
It isn't. GDP per capita in France is barely more than half of America's. It's lower than Mississippi, the poorest state in the US. Not a single global tech giant. Our best engineers build Silicon Valley instead. I've been going back and forth between Paris and San Francisco for years. I see it every week.
And California is about to replay the movie. Prop 40, on the ballot November 3rd, would take 5% of billionaires' net worth. Even Gavin Newsom, a Democratic governor, opposes it. The state's own nonpartisan budget analyst warns that income tax revenue could drop every year if the wealthy leave. Larry Page, Sergey Brin and Peter Thiel have reportedly already started moving their assets out.
You don't fight poverty by shrinking the top. You fight it by growing the cake.
Bernie wants a society where nobody has too much. I want a society where nobody has too little.
The first has been tried everywhere and produces equality in stagnation. The second lifted a billion people out of misery.
Choose.
Marius N retweeted
Norges Bank setter opp renten for å redusere prisveksten. Samtidig har statsbudsjettet vokst med over 17 prosent i realverdi på fem år, med oljepengebruken omtrent like mye. Befolkningen blir bedt om å bremse mens staten fortsatt har foten på gassen.
Dagens renteøkning er regningen fra Støres «trygge styring» og ekspansive storstatspolitikk.
Artikkelen er omfattende og vil for noen oppleves innviklende, så her er en oppsummering med noen nøkkeltall og sentrale forklaringer som illustrerer utfordringen for dem som ikke ønsker 2000+ ord og full artikkel (som du finner i kommentarfeltet).
Statsbudsjettet har vokst kraftig
Finansdepartementets budsjettall for 2021 og 2026 viser en nominell økning på 671,5 milliarder kroner. Det tilsvarer en vekst på hele 43,6 prosent.
Justert for KPI og omregnet til 2025-kroner tilsvarer det en realvekst på rundt 17,3 prosent. På bare fem år.
Befolkningsveksten forklarer heller ikke utviklingen: Målt i 2025-kroner tilsvarer statsbudsjettets økning en realvekst per innbygger på om lag 12,4 prosent.
Staten bruker rundt 42.000 kroner mer per innbygger i årlige utgifter enn i 2021, etter at både inflasjon og befolkningsvekst er tatt høyde for.
Økningen i offentlige utgifter fordelt på antall sysselsatte tilsvarer en årlig "skatteregning" på hele 61.000 kroner.
Siden år 2000 har statsbudsjettets samlede utgifter økt med rundt 148 prosent – etter at prisveksten er trukket fra. I samme periode har befolkningen økt med bare rundt 26 prosent.
Og regningen må betales. Primært gjennom skatter og avgifter, eller gjennom stadig større overføringer fra Oljefondet. Utfordringen er at oljepengebruk er spesielt inflasjonsdrivende.
Oljepengene fjerner budsjettbremsen – og øker inflasjon
Oljekorrigert underskudd ligger nå på omtrent en fjerdedel av offentlige utgifter.
Inflasjonsjustert oljepengebruk har økt med rundt 105 prosent på 10 år og er rundt 420 prosent høyere på 20 år.
Når staten bruker disse pengene innenlands, skaper det kunstig, politisk styrt etterspørsel som ikke følger naturlige markedsmekanismer om effektiv allokering av ressurser.
Dette er en betydelig årsak til den samlede inflasjonen er på omtrent 22,4 prosent fra 2021 til 2026.
Oljepengene skaper en politisk utgiftskultur
Når underskudd uansett kan fylles med oljepenger, svekkes insentivene til nøkternhet og prioritering:
Dagens system premierer i langt større grad virksomheter for å argumentere for hvorfor de trenger mer penger enn for å dokumentere hvor de kan spare eller få mer igjen for hver krone.
Riksrevisjonen slo i juni fast at myndighetenes arbeid med å sikre effektiv ressursbruk i offentlig sektor ikke er tilfredsstillende:
Nær to tredjedeler av statlige virksomheter oppga at de i liten grad har insentiver til å fremme innsparings- og kuttforslag i budsjettprosessen. Tre av fire hadde de siste årene fremmet ett eller flere satsingsforslag, mens bare én av fem hadde fremmet forslag om innsparinger eller kutt.
Riksrevisjonen peker også på at eksisterende ordninger kan bli videreført uten en reell vurdering av om pengene kunne vært brukt bedre et annet sted. Departementene mangler dessuten systematisk informasjon om sammenhengen mellom ressursbruk og resultater i staten.
Norge må tråkke på nødbremsen
IMF mener finanspolitikken bør flyttes over til en nøytral innretning for å bidra til lavere inflasjon, og anbefaler at nye utgifter i større grad finansieres gjennom kutt, omprioriteringer og effektivisering andre steder.
Målet er at staten ikke skal motarbeide Norges Banks forsøk på å dempe etterspørselen og få ned prisveksten.
I Perspektivmeldingen 2021 advarte Finansdepartementet: «Fremover vil det i større grad være nødvendig å finne inndekning for nye satsinger gjennom omprioriteringer og reduksjoner i andre utgifter» – «og ikke ved økt bruk av oljepenger».
Årsaken var den aldrende befolkningen med flere pensjonister og større utgifter til pensjoner, helse og omsorg, samtidig som det fødes for få nordmenn og veksten i antall yrkesaktive avtar. Det fører til at en mindre andel av befolkningen er i jobb og betaler skatt.
Færre skal betale mer for flere.
Fem år senere har statsbudsjettets samlede utgifter økt med 17,3 prosent i realverdi, og oljepengebruken tilsvarende.
Kursen kan vurderes ved valgurnene
Alt dette i betrakting, faller det naturligvis i ekstremt dårlig smak at rentehevingen kommer ett år etter at Arbeiderpartiet brukte «rentekutt for alle» som hjertesak i valgkampen.
Frp-leder Sylvi Listhaug mener at velgerne har blitt lurt. Også hun trekker frem veksten i offentlig pengebruk som en faktor som bidrar til rentepresset. Høyres Henrik Asheim retter lignende kritikk mot regjeringen ved å understreke at finanspolitikken påvirker handlingsrommet til Norges Bank. Også KrF og Venstre har ved jevne mellomrom kommunisert tilsvarende poenger.
Så da er det på tide å holde de borgerlige på ordet.
Først i 2029 ventes inflasjonen å være tilbake på målet på to prosent. Det betyr at konsekvensene av den økonomiske politikken vil følge oss helt frem til neste stortingsvalg.
Velgerne får dermed flere år til å vurdere om utviklingen samsvarer med løftene som er gitt – og om kursen bør fortsette eller endres. Vi er langt forbi tiden der vi kan talltrikse oss til valgseier, men det er fortsatt ikke for sent å trykke på nødbremsen og å snu.
Vanlige husholdninger blir bedt om å bremse. Staten har fortsatt foten på gassen. Det er på tide å si fra. Det er på tide å kutte.
Marius N retweeted
Five numbers tell you whether an AI business is a real business.
Gross retention.@WeAreLegora is 95%. Customers who bought last year are still here this year. If this one is broken, nothing downstream matters.
NRR. Ours is 300%+. Gross retention is the floor. NRR is how much taller customers build on top of it. We don't sell shelf-ware.
DAU/MAU. Ours is north of 50%, and the average active user spends 17 hours a month in the product. A rollout tells you a firm has signed. This indicates the work actually matters, and it moves here before it appears in retention or NRR.
Win-rates. Our August pilot closed-won-win rate was 78%. Winning roughly 4 out of 5 competitive pilots is downstream from offering a superior product.
Gross margin. The one that matters most. Ours is positive and improving every quarter. Our customers want Legora to be a long-term partner, and this is what makes that possible.
There's a shorter route: Price below what it costs to serve, book the logo, and hop on a never-ending fundraise treadmill to pay for it. The top line goes up, everyone claps, and every new customer costs more than they pay.
Scaling a negative margin only exaggerates the problem. The whole point of scale is that the margin improves as you go up. Ours does. That's the only version of this business worth building.
I dag er jeg imponert over Vy og @BaneNORSF sin løsning på forsinket tog fra Skien og overgang til Bergensbanen i Drammen. Skienstogrt skulle vært inne 16:45, men kom 17:01. Bergensbanen skulle gått 17:00, men var et par minutter bak skjema.
1
1
75
Marius N retweeted
🚨BREAKING: Two researchers from UPenn and Boston University just published a paper that should be uncomfortable reading for every CEO automating their workforce right now.
The argument is straightforward. Every company replacing workers with AI is also eliminating its own future customers. Laid off workers stop spending. Enough of them stop spending and nobody can afford to buy anything. The companies that fired everyone end up selling into an economy with no purchasing power left.
Every executive can see this. The math is not complicated. But here is why nobody stops.
If you do not automate, your competitor does. They cut costs, lower prices, take your market share, and you collapse anyway. So every company automates knowing it is collectively destructive because the alternative is dying alone while everyone else survives. The researchers proved this is a Prisoner's Dilemma playing out in real time.
The numbers are already moving. Block cut nearly half its 10,000 employees this year. Jack Dorsey said AI made those roles unnecessary and that within the next year the majority of companies will reach the same conclusion. Salesforce replaced 4,000 customer support agents with AI. Goldman Sachs deployed a coding tool that lets one engineer do the work of five. Over 100,000 tech workers were laid off in 2025 and AI was cited as the primary driver in more than half those cases. 80% of US workers hold jobs with tasks susceptible to AI automation.
The researchers tested every proposed solution. Universal basic income does not change a single company's incentive to automate. Capital income taxes adjust profit levels but not the per-task decision to replace a human. Collective bargaining cannot hold because automating is always the dominant strategy.
They also identified what they call a Red Queen effect. Better AI does not solve the problem, it accelerates it. Every company chases faster automation to gain market share over rivals but at the end everyone has automated equally, the gains cancel out, and the only thing left is more destroyed demand.
The one thing the math says could work is a Pigouvian automation tax. A per-task charge that forces companies to account for the demand they destroy each time they replace a worker.
The conclusion is that this is not a transfer of wealth from workers to owners. Both sides lose. Workers lose income. Companies lose customers. It is a deadweight loss with no market mechanism to stop it on its own.
(Link in the comment)
Da er renten holdt uendret (som gir rentekutt for ingen) og skattelotteriet (som gir skattekutt for noen få) implementert. Da fikk vi motsatt av hva vi ble lovet av disse politikerene?
Er det på grunn av slike mekanismer man skal droppe Norgespris selv om den er her? Eller er det mest en god grunn for å velge Norgespris - systemet kommer til å feile og gi ekstreme priser uansett, så da er det fint med statlig forsikring…
Strømprisen settes ikke høyt for moro skyld. Det skriver jeg i min første artikkel i den månedlige spalten jeg skriver for magasinet Aftenposten Innsikt. Der skal jeg skrive om strøm og energi for folk flest. Den kom først på papir. Her er den på nett: aftenposteninnsikt.no/norge/…
1
289
Hei @Ruter. Jeg sykler stort sett når jeg skal noe sted i Oslo. I dag tenkte jeg prøve trikken. Den kjørte et minutt før tiden selv om jeg var på perrongen. Skal jeg velge bilen hver gan sykkel ikke passer?!
1
330
Og du, det er ok at du ikke er god nok.
nrk.no/ytring/ikke-god-nok-1…
27
Noen grunn til at SP ikke skal henge seg på FrP, Høyre og KrF for et rent flertall og få reell makt i regjering?
1
36


















