Don't learn formal methods. Writing proofs by hand is what kept formal methods niche for over 75 years. seL4 took about 20 person-years to prove 8,700 lines of C. Agents take that part now. What's left for you is the spec, and knowing which language fits the job. That's why the history matters. Formal methods means stating what a program or system must do in precise math, then using a tool to check that the code or design meets it. Here's how we got here, what each language is good for, and where I'd start. FOUNDATIONS (1947 to 1979) 1947: Goldstine and von Neumann annotate program flow diagrams with "assertion boxes" stating what holds at each point. 1949: Alan Turing's "Checking a Large Routine" proves a short program correct using assertions. The idea mostly sits unused for over a decade. 1962: Carl Adam Petri's dissertation introduces Petri nets for modeling concurrent systems. 1963: John McCarthy, "A Basis for a Mathematical Theory of Computation". 1966-67: Peter Naur and Robert Floyd independently show how to prove programs with assertions attached to points in the code. 1967: N. G. de Bruijn starts Automath, the first language for computer-checked mathematical proofs. 1969: Tony Hoare publishes Hoare logic, {P} C {Q}. Nearly every contract language since builds on it. 1972: Robin Milner's Stanford LCF. Its Edinburgh successor (1979) introduces ML as its scripting language, the ancestor of Standard ML, OCaml and F#. 1973: Mizar starts in Poland. Formalized math that reads close to a textbook. 1975: Dijkstra's guarded commands and weakest preconditions. 1977: Amir Pnueli brings temporal logic to programs. Patrick and Radhia Cousot publish abstract interpretation, the theory behind sound static analyzers. 1978: Hoare's CSP for concurrent processes. Milner's CCS (1980) and the pi-calculus (1992) follow. 1970s: VDM at IBM's Vienna lab. Jean-Raymond Abrial proposes Z, later developed at Oxford. Boyer and Moore build their theorem prover. TOOLS PEOPLE USED (1980 to 1999) 1981-82: Model checking, invented independently by Clarke and Emerson and by Queille and Sifakis. Turing Award in 2007. 1980s: Synchronous languages Esterel and Lustre in France. Lustre becomes SCADE, which Airbus uses to design flight control software. 1984: Nuprl at Cornell. 1985: Coquand and Huet's Calculus of Constructions, the logic under Coq. 1986: Isabelle (Larry Paulson). Late 1980s: SPARK, a provable subset of Ada. 1988: HOL (Mike Gordon). 1989: Coq's first public release. Renamed Rocq in 2025. 1990: Symbolic model checking with BDDs handles 10^20 states. McMillan's SMV follows. 1991: SPIN model checker goes public (Gerard Holzmann, Bell Labs). 1992: PVS at SRI. Oxford and IBM win the Queen's Award for using Z on IBM's CICS transaction system. 1993: Metamath, a proof language simple enough to check with a few hundred lines of code. 1994: Lamport's TLA. TLA+ follows in 1999. 1994-95: Intel's Pentium division bug costs it $475 million. AMD uses ACL2, Kaufmann and Moore's successor to the Boyer-Moore prover, to prove its K5 floating-point division correct. Intel hires John Harrison in 1998 and uses his HOL Light to verify floating-point algorithms. 1996: Abrial's B-Book. 1997: Alloy (Daniel Jackson, MIT). 1998: Paris Metro line 14 opens driverless, with safety software built in the B method. The KeY project for Java verification starts. 1999: Agda (Catarina Coquand). JML, a contract language for Java. VERIFIERS MEET REAL CODE (2000 to 2019) 2000: Microsoft's SLAM project starts. It ships as the Static Driver Verifier in the Windows driver kit. 2001-02: Separation logic (Reynolds, O'Hearn, Yang) makes pointer-heavy code provable. 2002: ESC/Java. Spec# (2004) and the Boogie verifier (2005) follow at Microsoft Research. 2003: Astrée proves the Airbus A340's primary flight control software free of run-time errors. 2004: CBMC, bounded model checking for C. 2005: Georges Gonthier finishes a Coq proof of the four color theorem. 2006: CompCert, a C compiler proved correct in Coq (Xavier Leroy). Event-B gets its Rodin toolset around the same time. 2007: Agda 2. Idris first appears (Edwin Brady). 2008: Z3 SMT solver (Microsoft Research). Frama-C. VeriFast. 2009: Dafny (Rustan Leino). seL4, the first general-purpose OS kernel with a machine-checked proof that its C code matches its spec. 2010: Why3. 2011: F*. 2012: A Coq proof of the Feit-Thompson odd order theorem is completed. 2013: Lean (Leonardo de Moura, Microsoft Research). P language at Microsoft. 2014: CakeML, a verified ML compiler. Liquid Haskell. Flyspeck finishes a formal proof of the Kepler conjecture. 2014-15: AWS reports TLA+ found bugs in S3 and DynamoDB that testing missed. IronFleet proves distributed systems correct in Dafny. 2015: Iris, a separation logic framework in Coq. Facebook open-sources Infer. 2016: Project Everest starts, aiming for a verified HTTPS stack. Viper. 2017: HACL*, verified crypto written in F*, ships in Firefox 57. Lean's mathlib starts. 2018: AWS Zelkova checks IAM and S3 policies with SMT solvers. 2019: Fiat Cryptography's Coq-generated code lands in BoringSSL, used by Chrome. Prusti for Rust. LLMS MEET PROOF CHECKERS (2020 to 2025) 2020: OpenAI's GPT-f finds shorter Metamath proofs that get accepted into the main library, a first for deep learning. 2021: Lean 4. 2022: Kani (AWS) and Creusot for Rust. Aeneas translates Rust into Lean. The Liquid Tensor Experiment finishes the Lean proof Peter Scholze asked for. 2023: Verus for Rust. AWS releases Cedar, a policy language built with proofs in Dafny and later Lean. LeanDojo gives models a programmatic interface to Lean. Terence Tao's team formalizes the PFR conjecture proof in Lean in three weeks. July 2024: DeepMind's AlphaProof reaches IMO silver-medal standard with proofs in Lean. Sept 2024: Quint 1.0, TLA+ ideas with ordinary programming syntax. Mar 2025: Imandra launches CodeLogician, an agent that turns code into its IML logic and checks it. Apr 2025: AWS publishes how it rewrote its authorization engine, called a billion times per second, in Dafny with proofs. July 2025: Harmonic's Aristotle reaches IMO gold-medal level with formal Lean proofs. Sept 2025: The vericoding benchmark tests LLMs writing verified code from 12,504 specs. Off-the-shelf models pass 82% in Dafny, 44% in Verus, 27% in Lean. LANGUAGES BUILT FOR AGENTS (2026) Feb 22, 2026: Vera. "A programming language designed for LLMs to write, not humans." Feb 25, 2026: Vow. "A Programming Language for Agentic Coding." Mar 9, 2026: Kōdo. "Designed for AI agents to write." Z3-checked contracts, JSON errors with patches. Sept 17, 2026: Bend 2. Dependent types and Lean-style proofs, C-like speed, CPU and GPU parallelism, Python-like syntax. WHAT EACH LANGUAGE IS GOOD FOR Specs and models Z: precise requirements docs. Little tooling today. VDM, B, Event-B: refining a spec step by step into code. Railways and safety-critical control. CSP (with the FDR checker): concurrent protocols. Petri nets: workflows and concurrent processes. TLA+ and Quint: distributed system designs, consensus, race conditions before you write code. Checks the design, not the code. Alloy: fast counterexamples in data models and access rules. Bounded, so finding no counterexample isn't a proof. P: modeling and testing event-driven distributed services at AWS and Microsoft. Model checkers SPIN: concurrency protocols, as abstract models. NuSMV and other SMV descendants: hardware and finite-state control logic. CBMC: bounded checking of C code. Kani: bounded checking of Rust, including unsafe code. Static analysis Astrée: proving embedded C has no run-time errors. Avionics. Static Driver Verifier: Windows drivers using kernel APIs correctly. Infer: memory and null bugs across large codebases. Proof assistants ACL2: hardware and microcode. HOL4, HOL Light: hardware and floating-point proofs. Isabelle/HOL: seL4-scale systems proofs, strong automation. Rocq: verified compilers (CompCert), crypto (Fiat), big math proofs. Slow and hard to learn. PVS: aerospace, much of NASA's formal methods work. Agda: type theory research and teaching. Lean 4: math (mathlib) and AI theorem proving. Hardest of the big three for LLMs off the shelf (27%). Mizar, Metamath: large formal math libraries with simple checkers. Verifying real code SPARK: avionics, defense, rail. Dafny: verified business logic, compiles to C#, Java, Go, Python, JS. AWS uses it for its authorization engine. Easiest for LLMs off the shelf (82%). F*: verified crypto (HACL*, in Firefox and the Linux kernel). Frama-C, Why3: C in safety-critical settings. KeY, JML: Java. VeriFast, Viper: separation-logic proofs about heap-heavy code. Liquid Haskell: proofs added to existing Haskell. Idris: general-purpose programming with dependent types. Verus, Prusti, Creusot, Aeneas: proofs about Rust code. WHY IT NEVER SCALED Formal methods worked. They didn't scale, because a person had to write every proof. Back to seL4. Its 8,700 lines of C needed roughly 200,000 lines of Isabelle proof, more than 20 lines of proof per line of code. IronFleet's verified distributed systems took 3.7 person-years, including inventing the method. So formal methods stayed where a bug costs more than the proof: flight control, railways, crypto, kernels, cloud authorization. Agents change that math. The proof is the tedious part, and an agent will grind on it all night. You also don't have to trust the agent, because the verifier accepts the proof or rejects it. What stays human is the spec. A green proof of the wrong spec is still wrong, so that's where your review time goes. WHERE I'D START I would start with the Verus formal language. The agent is the new part, so I'd pair it with a verifier that isn't. For agents writing verified code. It checks ordinary Rust against specs and proofs written in Rust, then erases them before compiling, so what you prove is what ships. It has peer-reviewed results since 2023, real systems built on it (Kubernetes controllers, OS kernel, etc), and use at Microsoft and Amazon. For other jobs: - Checking a distributed design: TLA+ - Math: Lean - Bend 2, Vera, Vow and Kōdo: worth watching, still very early though. Follow me for more on verification and agentic engineering.

Why institutions should build on Morpho? (recurring questions) Some element of answers: 1. Building secure and scalable smart contracts is underestimated in terms of how hard it is. Our protocol team is literally spending months or years to ship contracts that are just a few hundred lines of code. The reason why Morpho is still here today is because we've made security our number 1 priority. The last thing we want to do to enable a new integration is write contracts. Recreating this expertise is not an easy endeavor. 2. Innovation does not happen in silos but on open platforms. It's not intranets that gave us Airbnb or Uber. It's the internet, an open protocol that lets everyone build websites and apps. The Minitel had a great start, but only the surface of the network was decentralized; the core was not. It hindered what could be built on top. It could not compete with the World Wide Web. The same principles will apply to financial services. 3. Institutions already own the distribution, trust, and relationships. DeFi does not compete for this. In the long term, DeFi might not even be called as such. It will just be part of the financial infrastructures that have replaced spreadsheets and COBOL, which only a few could read and update. This will make institutions' backends much more efficient, allowing them to stay relevant in a space that is more competitive than ever. 4. Neutrality is key: just as Dee Hock formed Visa out of a consortium of competitive entities, Morpho provides a neutral infrastructure for credit. At the time, BankAmericard had too many limitations while owned and operated by Bank of America. The fact that Visa was an independent consortium of competitors brought the required neutrality for those entities to cooperate and maximize the value of card payments. Morpho is providing that truly neutral infrastructure for credit: no governance, immutable code, etc. This is what will make entities that wouldn’t otherwise trust each other actually cooperate or at least share markets. For end users it means deeper liquidity and better rates. An entity that tries to do everything on its own won’t benefit from Morpho’s network effects, which means worse rates and a harder time competing. At the end of the day, the end customer is the one who benefits.

heard a great line today. "the smallest of margins make the biggest of differences." true in trading more than most places. 2% in costs, a slightly better fill, one position sized slightly wrong. none of it feels like much on the day. compound it over a few hundred trades and it's the whole result.

The Information reports that Tesla managers are targeting an automated production line capable of producing more than 1,000 robots per week by the end of 2026. Everyone has been wondering where is Optimus v3. Jason Calacanis - weeks ago said Elon has been sending him videos of the new robot, that he’s “not showing publicly” but would win “half of these things” (referring to the Chinese humanoid Olympics) According to The Information, Tesla increased Optimus production at its Fremont, California, factory from a few dozen robots weekly during the second quarter of 2026 to several hundred in August. If this true that would mean Tesla has thousands of humanoid robots currently.

[ Marginalia ] Two Hundred and Thirty-Seven. Her name in the file is Mei. Not her name, the one they gave her afterward, so she could be spoken about without being found. At four she was behind in language, in writing, in drawing. Her crayon houses had no doors. The doctors put a word on it that fewer than two hundred and thirty-seven people in the world have ever been given. A gene called CHD3, misspelled by one letter in the earliest draft of her brain. Two hundred and thirty-six others, somewhere, in rooms her parents would never enter. That was the number they held on to. A number is a kind of company. Jason kept the medical records in a folder on his desktop. He opened it more often than he needed to. In the WeChat group, forty parents, all of them awake at two in the morning, all of them typing, a name surfaced in 2023 the way a name surfaces in a city: from someone who knew someone. Shanghai. Xinhua Hospital. A laboratory that had made mice with his daughter's exact mistake inside them. He wrote the email at night. He attached everything: the scans, the reports, the sentence saying he was willing and able to bear the costs. Then he sat there, the way you sit after posting a letter you can't take back. Thirteen minutes. He has counted those thirteen minutes many times since. He has divided them, turned them over, held them up to the light. In thirteen minutes a man can boil water. In thirteen minutes a man on the other side of a country can read a child's whole life and decide what to do with it. Later Jason would understand what the speed meant, that the answer had been waiting long before the question. That somewhere a laboratory had been standing at a window for two years, watching the road, waiting for a father to walk up it carrying a folder and a checkbook. They paid. First for the mice. Then for the monkeys. Then one hundred and thirty thousand dollars into a researcher's personal account, the way you'd pay a man to fix a roof. Here is what nobody said out loud in the consent conversation, though everything else was said, carefully, at length: the worst case is that she dies. They spoke of risks the way a hotel speaks of the weather. The dose went in through the spinal fluid, adenovirus, thousands of times a vaccine's load, because you cannot edit one cell when what you need is a million. Twenty-seven years ago in Arizona there was an eighteen-year-old named Jesse Gelsinger, and on the fourth day his organs stopped, one after another, and the whole world was supposed to have learned something. Somebody did learn something. They learned where the reviews are looser. Because there are two doors in that country. One goes through the drug regulators. The other, for research the hospital calls non-commercial, curiosity's own door, goes through a committee down the hall. The committee that approved Mei never read the primate safety report. This is not an oversight. Oversights get corrected. A door that stays open for thirty years is not a door somebody forgot to close. Mei was not dying. That is the part her mother cannot put down. She was six, she was behind in drawing, she was not dying. A year after she died, the paper came out in Nature. Beautiful mice. Beautiful monkeys. A breakthrough. What had been quietly removed: the human data, and the line thanking a family in the acknowledgments. Their daughter had been edited out of the sentence that thanked them for her. So they went and asked for the paper to be pulled. Not for her. For the parents reading it that morning, in forty group chats, at two in the morning, thinking: there's hope now. Two hundred and thirty-seven, minus one. Nobody has updated the number. Nobody was ever going to. — Aric

Ten years ago, one developer pushed 598 lines of code the day after Ethereum came alive, on the bet that people should be able to hold and use their money directly. A hundred million downloads later, the bet is the same. Here's what's coming next 👇

Every day for 47 years, Mr. Hamano rowed a wooden ferry across a river in Japan. Back and forth. The same line across the same water. About two hundred thousand crossings. He is 81 now. Retired. This July, he boarded his first airplane, bound for America, to see one thing. A young soldier who walks back and forth. Mr. Hamano is the grandfather of Daigo, who grew up in the apartment above mine. Daigo is 26 now, an engineer in Virginia. He told me this story over a beer last month, and had to stop twice. His grandfather ran one of the last hand-rowed ferries in Japan. A flat wooden boat, one long oar, a little hut on the bank with a bell. You rang the bell, and he came. Rain, snow, typhoons. He never once left someone waiting on the other side. When he finished a crossing, he always said one word. "Yoshi." It means something like "Right. Good. Done properly." He has said it about two hundred thousand times. In 2011, Daigo's grandmother was dying in the hospital. On her last morning, it was the day of the university entrance exams. A high school girl from across the river rang the bell at 6 a.m. His grandmother told him: "Go. Somebody's waiting on the other side." He went. He rowed the girl across. Then he hurried back to the hospital and held his wife's hand until she was gone. He has never talked about that morning. Daigo heard it from his mother. In September 2003, a hurricane hit the American East Coast. Mr. Hamano saw it on the little TV in his ferry hut. A cemetery in Virginia. Wind tearing the trees sideways. Everyone told to take shelter. And one young American soldier, in full uniform, still walking. Twenty-one steps. Turn. Twenty-one seconds. Turn. Twenty-one steps. The Tomb of the Unknown Soldier, at Arlington. Guarded every minute of every day since 1937. By soldiers who volunteer for it. The news said the guards had been given permission to take shelter. They refused. Mr. Hamano watched that soldier walk through a hurricane, and he said one sentence to an empty ferry hut. "That one understands." He waited 23 years to go and see him. Daigo had a whole trip planned. New York. Washington. A ballgame. His grandfather said no. "Only the soldier." It was July. 96 degrees. Virginia humid, which Daigo says feels like being licked by a very large dog. His grandfather stood at the rope for two hours. He took out his old wristwatch and timed the steps. Twenty-one. He timed the pause. Twenty-one. And every single time the soldier turned, Mr. Hamano nodded and whispered, "Yoshi." Every turn. For two hours. Daigo says a family from Ohio started watching his grandfather instead of the soldier. Then the sky turned black. A summer storm. Thunder close enough to feel in your teeth. Rain like a waterfall. In about a minute, the crowd was gone. Running for the buses. Umbrellas everywhere. Daigo opened his umbrella over his grandfather. His grandfather closed it. Daigo opened it again. His grandfather closed it again. "Grandpa, you're 81. You're going to get sick. He's a soldier. It's his job." His grandfather didn't take his eyes off the mat. "The ferry was my job," he said. "Did I stay home when it rained?" So Daigo stood there, holding a closed umbrella, getting soaked next to his grandfather, who was wearing a canvas hat and would not move. They were not the only ones. A teenage boy, maybe fifteen, had stopped halfway to the bus. He saw the old Japanese man standing in the rain. He turned around and stood at the rope. Then his father came back with an umbrella to get him. He saw his son. He saw the old man. He closed the umbrella. By the time the guard changed, there were nine people standing in the rain. Nine closed umbrellas. Not one of them spoke. The new sentinel came down the steps, and the inspection began. The rifle. The uniform. Every crease. In pouring rain, exactly the same as on a sunny day. When it was over, and the old guard marched away, Mr. Hamano took off his canvas hat and bowed as low as his back would go. The same bow he gave the river after his last crossing, the day he retired. Then they walked to the bus, wet to the bone. About forty minutes later, at the visitor center, a young soldier in uniform walked straight up to them. "Sir. Excuse me. Were you the gentleman at the Tomb, in the rain?" Daigo said yes. "One of our sentinels just came off the walk. He asked me to find you. He saw you close the umbrella. He saw all of them close." The soldier stood up a little straighter. "He said to tell you: thank you for standing the post with him." Daigo translated it. His grandfather looked down at the floor. After a while, he asked Daigo to translate something back. Exactly. "Tell him: 47 years. A ferry. Back and forth. Every day. Somebody always waiting on the other side." He nodded and left. Ten minutes later, he came back. "Sir. He says: 'Then you know.'" Daigo had never seen his grandfather cry. Not at his grandmother's funeral. Not ever. The old man put his canvas hat over his face, right there in the visitor center, and cried like a boy. The young soldier stayed exactly where he was, at attention, until he was finished. Mr. Hamano went home to Japan. The ferry runs with a younger man at the oar. But Daigo's mother says that every morning at six, his grandfather still walks down to the river. If it's raining, he does not open his umbrella. He stands on the bank for one minute, facing east, toward the morning. Toward a soldier who is walking, right now, in Virginia. He doesn't talk about America much. When people ask what it was like, he gives them the same one-word answer he gave the river for 47 years. "Yoshi."

Framework: the verdict listicle. Cold open from a hair loss ad that has run for one hundred and twenty-six days: "Popular hair loss treatments ranked by a dermatologist." It works for three reasons: Every option gets one verdict, "Leave it," so you keep watching to see which one gets the other answer. The options are ones buyers already use (Minoxidil, Nutrafol). Each verdict comes with a one-line flaw, so every "Leave it" is a reason to switch. The click goes to a "comparison report" instead of a product page, so the ranking still looks neutral up to the sale. To make one, rank what your buyer has already tried and give each a verdict and one flaw. Put yours last. This one and sixty-one more listicle ads in one swipe file:

Elon Musk told the world what SpaceX is building next. A sentient sun. He wrote it last night, quoting SpaceX's own announcement. "SpaceX has ordered 100 GW/year of HJT solar manufacturing equipment from Singulus in Germany. This is for Starmind to create a sentient sun." One hundred gigawatts a year of solar manufacturing equipment, ordered from Germany, for a factory in Bastrop, Texas. That is where SpaceX is building the Starmind satellites. SINGULUS, the German company making the machines, confirmed the first full production line is already under contract, worth more than 60 million euros, and arrives in the first half of 2027. More lines are planned as production scales. SpaceX has been funding that company since March, and again in October. They are building the supply chain themselves. Starmind is SpaceX's name for computers that live in orbit. They catch sunlight up there and send their results back to Earth through Starlink's laser links. The first satellite, AI1, has a 75 meter wingspan and a compute payload of up to 250 kilowatts. SpaceX says that in sun-synchronous orbit those satellites capture continuous solar power. No weather, and no loss through the atmosphere. The light just keeps arriving. Mike Nicolls, SpaceX's senior vice president of Starlink, said why the deal exists: "Partnering with SINGULUS gives us a high-performance manufacturing path for solar cells for our Starmind satellites at our production facility in Texas." The sun has been pouring energy into empty space for billions of years. Elon Musk ordered the equipment to put a mind in that light. The first line leaves Germany in the first half of 2027.

On the morning of October 9th, forty-one combines came over the hill at sunrise. In a line. Headlights still on. Engines so loud the ground hummed. Like an army that had decided to be kind. My coworker's son, Hiroto, was standing in the field when they came. He was sixteen. He didn't know what he was looking at. Hiroto had spent the year as an exchange student with the Lindgren family in eastern Nebraska. Corn and soybeans. Nine hundred acres. The same family had worked that ground since 1911. The first month was a comedy. Tom, the father, woke him at 5:30 to feed the cattle. Hiroto had never seen a cow up close. The cow was not impressed. He asked Tom how many kernels of corn were in the field. Tom thought about it and said, "Oh, about forty million." Hiroto stopped asking questions about corn. The farm dog, a scruffy old heeler named Biscuit, decided on the second day that Hiroto belonged to her. Nobody was consulted. And there was Cody. Tom's son. Seventeen. Quiet. Arms like fence posts. For the first two weeks, he didn't say a single word to the Japanese kid sleeping across the hall. Then one night, in the truck, they fought over the radio. Country music versus Japanese pop. It got loud. It got serious. Tom laughed so hard he had to pull over. After that, they were brothers. Nobody decided it. It just was. In July, Tom collapsed in the north field. Heart attack. He was 52. He was gone before the ambulance found the road. Grief in a farmhouse is quiet in a specific way. His boots stayed by the back door. Laced. Like they were waiting for him to come in for supper. Nobody could move them. Ellie, nine years old, started sleeping in her father's seed cap. It came down over her eyes. She wouldn't take it off. And Cody, at the funeral, said exactly one sentence. "I'll bring in the harvest." Nine hundred acres. A job for a whole crew. Tom had brought in every harvest for thirty years. He had never missed one. Not for a broken arm. Not for a blizzard. Not once. Cody was not going to be the Lindgren who missed one. So in September, he started. Alone. Up at four. In the combine before the sun. Back after midnight. Hiroto couldn't drive a combine. So he did the only thing he could. He rode in the cab next to his brother. All day. All night. He held the thermos. He talked. In bad English, about anything. Baseball. Japan. Girls. Biscuit. Anything to keep Cody's eyes open. Some nights Cody fell asleep at the wheel for a second, and Hiroto would shake his arm, and Cody would say, "I'm good, I'm good." He was not good. One night, around two in the morning, in the middle of a dark field, Cody stopped the combine. He put his forehead on the steering wheel. And he broke. All of it. Everything he hadn't let anyone see since July. Seventeen years old. Alone in the middle of nine hundred acres, in his father's machine, in the dark. Hiroto told my coworker he didn't know a single English word that could help. So he put his hand on Cody's back, and said the only word that came. In Japanese. "Niichan." Big brother. Cody didn't know what it meant. He knew exactly what it meant. They sat there until the sun came up. Two weeks later, on the morning of October 9th, Hiroto heard the sound first. He walked out of the house, still in his socks, and looked up at the hill. And they came. Forty-one combines. Grain trucks behind them. Semis behind those. Neighbors from three counties. Men who argued with each other at the co-op every Monday. Men who hadn't spoken to each other in years. Old men. Young men. The high school football team, driving grain carts. Church women unloading folding tables and coolers of sandwiches at the edge of the field. Nobody had called them. Nobody had to. Cody came out of the house and just stood there, staring. He couldn't move. An old neighbor named Arlen, eighty years old, climbed down out of the lead combine, walked across the dirt, and put his hand on Cody's shoulder. "Your daddy helped every single one of us, son. Every one." Then he pointed at Tom's combine. "Get in. You're leading." And Cody climbed into his father's combine, and pulled out in front, and led forty-one machines across his father's land. They brought in nine hundred acres in one day. At sunset, there was one row left. And without anyone saying a word, every combine stopped. They pulled off to the edge of the field. Forty-one machines in a long line, side by side, headlights on in the gold light. Two hundred people stood by the trucks. And they let Cody take the last row. Alone. With Ellie in the cab beside him, in her father's cap, pushing it up out of her eyes so she could see. When he reached the end of that last row, forty-one horns went off at once. You could hear them, people said, in the next town. Hiroto came home to Japan in June. His father asked him what the most important thing was that he'd learned in America. He expected English. Or baseball. Or maybe something about corn. Hiroto thought about it. Then he said: "In Japan, we say a family is the people under one roof. In Nebraska, I learned that family is whoever comes over the hill." Every October now, Hiroto gets a photo from Cody. Always the same picture. A line of combines coming over a hill at sunrise. Always the same three words. "They came again."