Jimmy Soni
Biographer and professional speechwriter
Where to hear and read about Jimmy Soni3 episodes and tests
How I WriteIs AI Killing Writing? 12 Different Answers
How I WriteHow To (Really) Write With AI — Jimmy Soni
How I WriteHow to Write Like An Athlete | with Jimmy Soni | How I Write PodcastJimmy Soni is a speechwriter, ghostwriter and biographer who wrote The Founders, the story of PayPal's early years, along with books on Claude Shannon and a Brooklyn carousel, and who co-founded the publisher Infinite Books. The most useful thing to learn from him is that deep work gets made by systems: protected early hours, timed research, raw material shaped down instead of written from nothing, and a hunt for the overlooked sources everyone else missed.
Reading over and over and over again something that you admire teaches you not the answers but the questions.
How to Write Like An Athlete | with Jimmy Soni | How I Write Podcast
Jimmy Soni: the story and the lessons
1Jimmy Soni's path from library kid to biographer
Soni grew up a nervous reader in a family that knew libraries. He grew up in suburban Chicago in the 1990s, the son of immigrant parents who knew libraries rather than summer camps. Librarians held copies of Brian Jacques's Redwall novels for him, and he chased free pizzas through Pizza Hut's Book It program.
He calls himself a kid who was "afraid of everything", and says that anxiety never left. Good English grades first made him think he had some talent.
Writing for a real audience made him work harder. He sweated his college newspaper columns more than his coursework, because friends would notice any misspelling or weak argument.
He calls this one of the most important steps in leveling up as a writer.
He got his reps by volunteering first drafts for other people. He started by offering a draft of a colleague's big speech, and at McKinsey a draft he offered got a team 70% of the way there. He went on to write op-eds, speeches and ghostwritten pieces for business and Hollywood clients.
Do it often enough, he says, and people come to you to write. He writes thousands of words a day outside his books, and ghostwriting lets him put in reps without his name on the line.
Soni writes upward, learning the field alongside the reader. He knew nothing about PayPal or Claude Shannon before writing about them, and calls himself a "better than average charlatan" who learns as he goes.
Because his reader knows nothing either, every term of art has to be made readable, and the pleasure of finding things out stays in the book.
He commits to a book only when he sees an empty space on the shelf. He parks ideas in a Google Doc called "books I should write someday" and simply buys any book that already covers a topic. If he can picture the two books on either side of an unfilled gap, he writes a proposal.
Each book takes four to six years, so he asks whether the world needs it and whether he can spend years on it. He treats this as an entrepreneur's move and refuses to write "the sequel to the sequel".
Making a physical thing from nothing is what drives him. Six years before, there was no Founders book, and then there was a 400-page object.
"That act of creation lights me up," he says.
2How Soni wrote The Founders, the story of PayPal
He saw that PayPal's founding years had never been properly told. Peter Thiel, Elon Musk and Max Levchin got constant coverage for what they did later, but the 1998 to 2002 creation story felt undercooked even after he read Zero to One, The PayPal Wars, Ashley Vance's Musk book and Reid Hoffman's books.
Every Thiel speech got a "Talmudic exegesis", he says, yet nobody had asked how they got their start.
The book took six years and a huge pile of raw material. He wrote about 470,000 words of raw material and kept about 170,000, and each chapter went through 17 or 18 revisions.
Each chapter had 13 Google Docs versions and then more than five PDF versions.
A plain spreadsheet and redundant backups kept the interviews under control. One Google Sheet tracked every interview with the name, date, pre-interview notes, audio file and post-interview notes. Audio was kept in two copies, one on a thumb drive, and transcribed with Otter AI for a flat fee.
He built the system as he went, and says nobody starts with the map figured out.
Soni records every interview, then debriefs himself right after. He brings only paper with his questions and leaves a phone recording on the table. After key interviews he types what stood out for a minute or two, then records a voice memo of his impressions.
Typing and speaking surface different memories. Only these notes preserved Thiel pulling out his phone mid-interview to check PayPal's market cap against Visa and MasterCard, a detail with no audio.
Musk thought out loud instead of reciting a script. The interview happened in Musk's living room on a Saturday afternoon, with food, and Musk paused and reworded answers he had clearly given before.
Soni says most people at that level are highly scripted because so much is at stake, which made this unusual.
PayPal's own writing became his best source. In hundreds of thousands of pages of emails he found the weekly company newsletter, where in 2001 and 2002 David Sacks wrote a 700 to 1,000 word product Q&A column answering employee questions.
Thiel prized clear writing and recruited Sacks partly for it. Founders also put their biggest moments in writing, such as Thiel's note on 9/11 and Musk's long resignation note.
He paid a fact-checker to demand a document for every claim. Ben Kallen spent six months on three rounds, and every fact or footnote needed a confirming document in a chapter's Google Drive folder. He flagged unsourced claims, wrong years and misquoted interview lines.
Publishers do not really pay for this, so Soni did. Kallen also helped decide what belonged in a footnote versus the main text.
He walked PayPal's street to feel what the founders felt. He walked back and forth between 165 and 394 University Avenue in Palo Alto at different times of day, feeling the distance, the lunch walk and the weather between rival teams.
None of it made the book directly, but it showed him how unremarkable the place was. A friend calls him a method writer.
A few words from Walter Isaacson launched the project. At an event for The Innovators, Isaacson heard the PayPal idea and called it a great book idea, and Soni left thinking "I can do this."
Later, on a walk in Central Park, Isaacson told him to be chatty in his endnotes and show how conclusions were reached. Soni's endnotes run long on purpose, down to a single YouTube video.
3Why Jimmy Soni hunts for stories at the periphery
Soni looks at the toenails: what did everyone else miss? He borrows Michael Lewis's image that in near-identical Renaissance paintings the distinctive detail sits in the toenails. For Kobe Bryant that means the interview with 634 YouTube views, not the 81-point press conference.
He admits he is not sure the painting story is true, but the image stuck.
An odd name on a cap table produced his best Thiel story. He searched an unfamiliar name, Ed Bogas, on an early PayPal cap table, found a San Francisco number and left a cold voicemail naming Confinity. Bogas, a musician and chess player, called back and said Thiel's merciless chess convinced him to write a check on the spot.
Bogas called it the only serious investment he ever made. Soni's "theory of the case" is that great stories come when you find your Ed.
Forgotten people have not rehearsed their stories on stage. He interviews people who stayed "a hot second", answered customer service calls or fought fraud, because famous founders repeat talking points they have given 15 times.
"I can't make up that story," he says of what turns up.
He opened his Musk interview with a question nobody else would ask. He asked about Greg Corey, an early adviser to Zip2 who died young, reasoning that grief casts a shadow that never leaves you. Musk paused and gave a minute-long meditation on the friendship.
People have lives outside the narrow frame we see them in, and asking about family or college often reveals the core thing.
He turns idle hours into research time. He listens to interviews about his subject while buying groceries, and during the PayPal years watched YouTube interviews with Thiel in the evenings instead of Netflix.
That lets research grow without eating into writing time.
4Jimmy Soni's rules for writing like an athlete
Soni treats writing as athletic training: reps, recovery, discipline. He draws nearly all his writing metaphors from sports, having grown up with the 1990s Bulls, and asks what Kobe Bryant would do as a writer. The answer is a huge volume of writing.
He applies the same lens to Musk and Thiel, comparing Thiel's competitiveness to Michael Jordan's.
He writes from about 4am, before the world can interfere. He writes from 4 or 4:30 until about 8 or 9am, seven days a week, with no email, admin or chores first. In book mode he will not schedule a 9am call.
He says the exact hour is not the point; it is time nobody else can claim. It also lets him be with his daughter without parent guilt.
The book gets his premium fuel. His sharpest hours are the first of the day, so the book gets them and everything else is ordered around it.
"The premium fuel is reserved for the morning."
Seven days a week turns finishing a book into math. Three to five hours a day will finish any book, he says, even if it does not guarantee a good one.
Once it is a default it stops feeling like work. On a vacation day without book time he felt off all day, in his body chemistry.
An identical ritual each morning removes decisions. He wore the same blue hoodie, flipped the lights on in the same order, drank cold sparkling water, and spent a minute picturing the finished book in his hands.
He worked every holiday and on his one vacation wrote before his daughter woke. He calls those mornings "awful and fun at the same time".
His morning runs reading, then a research timer, then a word count. He reads the model book for 10 to 15 minutes, does 30 minutes of research on a timer, then writes to a fixed word count, which defines a good day.
The reading is an on-ramp that tricks his brain into easy work before the blank page.
Love of the craft is the only fuel that lasts. Ambition burns clean for a while, then not so clean, and revenge burns worse, he says.
He points to Kobe Bryant, who journaled daily for 15 years and built his own publisher rather than accept a weaker book.
Other formats cross-train the main habit. He writes op-eds alongside books, and their roughly 700-word limit forces a word economy that carries into everything.
He compares it to a soccer player doing yoga, and thinks Twitter's character limits improved a generation of writers.
He cuts out contemporary media while inside a book. He told friends not to share news about the founders, unsubscribed from news newsletters and found out late about Musk hosting SNL. He read 300 printed pages of a 1990s blogger's posts instead.
He skipped the show Silicon Valley because it was satire, and prunes his newsletters monthly by asking whether each one delights him.
5How Soni beats the blank page and research paralysis
Soni treats endless research as procrastination dressed up as virtue. He knows someone who has spent two decades on a Plato book he never wrote. So when a section needs a fact, he digs it up and drops it straight into the draft.
"Research is an excuse not to write." He caps stand-alone research because the hunt is fun enough to need no schedule; what he schedules is making the thing.
He front-loads research in small timed daily blocks. His to-do list had 20-minute blocks: newspaper articles, PayPal emails, contacting interviewees, repeated every day.
After about 10 days on one source type you learn what it has to give. A biographer, he says, cannot know what to say until he knows what he has.
When he stalled, a coach imposed a painful penalty. Coach Lauren Zander diagnosed "analysis paralysis, research paralysis" and had him email her 1,000 words a day, with a $200 donation to a cause he hated for any miss. He hit every day for about 100 days.
He echoes Harry Lodge's line that humans are best in harness: constraints need real teeth, and a date beats a vague wish.
He never starts a chapter from nothing. Raw notes live in Scrivener folders by year; to start a chapter he pastes about 9,000 words into a Google Doc and shapes it down to roughly 5,500.
He sees writing as a sauce reduction rather than creation. A thin 2001 folder told him he had "phoned it in" and needed more research.
He dictates scenes with a manila folder over the screen. Using Google Docs voice typing with the monitor covered, he got down the McLaren F1 crash between Musk and Thiel without seeing typos or clumsy phrasing.
It works only after the research is done, when he already knows the facts.
He edits on his phone to tighten prose and lower fear. He did much of his best editing on his phone on waking, chapter four included. Padded paragraphs and ideas said 14 ways stand out on a small screen.
With notifications off, wandering to another app takes a conscious choice, and it feels like texting a friend.
Telling a story in shrinking time limits forces compression. He tells it freely, then in 2 minutes, then in 45 seconds, using voice typing.
People compress speech intuitively even when cutting five pages to one feels impossible.
He accepts that the fear never leaves. "Every day I'm still afraid of it," he says, so the job is building tricks that turn anxiety into output.
He wants writers to talk about fear more, since writing is often sold as glamorous.
6How Jimmy Soni builds taste and decides what to cut
Soni studies one model book obsessively per project. He read Brad Stone's The Everything Store 16 or 17 times and read 10 minutes of it each morning for about three and a half years. For the Kobe project his model is Naval's Almanack.
Stone traced where the name Amazon came from, so Soni dug up the origin of PayPal's name in slide decks. "I'm just Brad but for PayPal."
Every page should have a gem that makes readers smile, laugh or think. He borrows Apple's surprise and delight and Raymond Chandler's rule that something important happens on every page, so stale material 14 people had already said got cut.
A dull funding round became a joke: he recalled Musk's "funding secured" tweet and asked what Sequoia had actually invested in.
He asks what a reader in 2080 would still value. Imagining someone in 2080 or 2100 trying to understand Musk filters out petty office drama.
"Gossip has about a 5-minute shelf life." He says he gets it right about 95% of the time, and aims for books that last like The Power Broker and Seabiscuit.
Good history puts readers with people who did not know the ending. Following David McCullough, he writes as if the outcome were unknown, as with the Brooklyn Bridge builders.
He gave Musk at 19 warmth: too shy to speak in the Netscape lobby after an unanswered job letter, he went home, and then delayed Stanford by six months.
He writes the introduction last. Editor Alice Mayhew rejected an intro he sent before the book was written, saying he had not done the work to know the story's dimensions.
The intro is the reader's user guide, so it waits until the book is finished.
Writing is a team sport. He texted paragraphs to his friend Justin Richmond almost every morning for five and a half years, and a heart emoji or an honest "that's not that interesting" was a useful signal.
He tells newcomers to hire an inexpensive editor without ego, and likens his editors, fact-checker and readers to Phil Jackson around Michael Jordan.
7How Jimmy Soni uses AI as a writer
AI works middle to middle, and humans own both ends. He borrows Balaji Srinivasan's line: humans set the vision at the start and fact-check at the end, and AI helps in between.
An engineer friend put it as "the sharpest knife" but "I've still got to cook the meal." The words, research, facts and interviews stay his.
He has Claude write a savage takedown of his op-eds. He asks it to criticize the piece in 700 words and not hold back, then strengthens the draft against each objection. Nine times out of ten the counterarguments make him stop and think.
The idea comes from Tyler Cowen's point that you must be able to argue the opposite position.
He treats AI as an editor who never tires. At 4am he pastes in a flabby paragraph or a word that does not sit right and asks what is wrong or for other options.
He prefers Claude for writing because it has more personality, and uses ChatGPT for numbers and spreadsheets, fact-checking its math.
One query replaces about ten searches. Editing a grief book, he asked Claude for rabbinic sources on why Jewish burial happens within three days and got two explanations, spirit and flesh, to verify the author and share with her.
Learning it properly would have taken a month. Because hallucinations are real, he verifies everything.
Dictating a bad first draft cures his anxiety. He talks into Whisper Flow when he feels off and polishes the transcript. He used it on an apology note to a friend he had dreaded for months.
His daughter writes freely because ego is not in the way; for adults it is fear of critics and Twitter.
He sees AI as the next word processor. He respects Robert Caro writing by hand but says he does not see how a modern writer could skip AI. He calls training on books borrowing to learn, like Google Books.
He thinks many critics have not used the tools and fear for their income, and that using AI well is a skill of its own.
8Soni's case against how traditional publishing works
Soni paid for the basics publishers no longer provide. He paid for extra editing on almost every book and for his own cover on The Founders, because overloaded editors gave him table scraps of attention.
He stresses publishing is full of smart people, but consolidation leaves little room for flops.
He A/B tests every book cover. Services gather reactions from hundreds of people, because buyers decide in a snap on Amazon.
"Why wouldn't you AB test cover art?"
A two-week window kills slow-burn books. The Boys in the Boat debuted to crickets, then rowers passed it along and it hit the Times list about eight or nine months later.
He keeps his own books evergreen, so a reader discovering Claude Shannon today finds the book as new as at launch.
He co-founded Infinite Books to offer different deals. With Jim O'Shaughnessy he started Infinite Books, where some authors get 70% of royalties and no advance, against standard deals of 10 to 15% after an advance earns out.
He also wants publishers to run AI plagiarism checks before printing, calling it good, honest, responsible publishing.
A coffee table book taught him extreme word economy. After riding the Brooklyn carousel over 200 times with his daughter, he pitched Jane Walentas, who spent 30 years restoring it by hand, and Phaidon took the book. He rewrote every caption himself.
She filed away paint with X-Acto knives to find each horse's first color, and the book uses tissue overlays to show them. She died before it was done.
9What Soni learned from Kobe Bryant, Robert Caro and Michael Lewis
Kobe Bryant studied storytelling with basketball-level intensity. Soni watched Kobe interviews each morning during the pandemic and is now writing The Tao of Kobe, an almanac drawn from 400 to 450 interviews.
Kobe wrote Dear Basketball himself, recruited John Williams and Glen Keane, and became the first athlete to win an Academy Award. Asked if he had friends, he said, "What I have are fellow obsessives."
Caro shows how far immersion can go. Soni asks "What would Caro do?" of every story. Caro spends 10 years on a book and ran from the Capitol to see the dome light as Lyndon Johnson did.
Caro starts once he has the final sentence and writes toward it. Soni did not, and thinks it cost him time.
Michael Lewis shows the courage not to repeat yourself. Lewis could have written ten more finance books after Liar's Poker and wrote about coaching, parenting and baseball instead. At Salomon Brothers he wrote op-eds under his mother's maiden name, Diana Bleecker.
Soni took from Lewis's funny letters to friends the habit of asking how to make a friend laugh, and he does not want to be pigeonholed.
The PayPal story taught him to value friends who push you. Levchin and Thiel were not best buddies, but each improved the other, and Musk and Sacks disagreed productively.
In his acknowledgements he tells his daughter he loves her too much to be her coach, so she needs others who will call her out.
Is AI Killing Writing? 12 Different Answers
Twelve writers disagree about whether AI belongs anywhere near a draft, but most agree the thinking, the reading and the voice have to stay human. The most useful ideas are specific: use AI as a hostile critic, as secondary literature and as a source of a bad first draft, never as the author.
AI works middle to middle, and humans own both ends. He borrows Balaji Srinivasan's line: humans set the vision at the start and fact-check at the end, and AI helps in between.
An engineer friend put it as "the sharpest knife" but "I've still got to cook the meal." The words, research, facts and interviews stay his.
He has Claude write a savage takedown of his op-eds. He asks it to criticize the piece in 700 words and not hold back, then strengthens the draft against each objection. Nine times out of ten the counterarguments make him stop and think.
The idea comes from Tyler Cowen's point that you must be able to argue the opposite position.
He treats AI as an editor who never tires. At 4am he pastes in a flabby paragraph or a word that does not sit right and asks what is wrong or for other options.
He prefers Claude for writing because it has more personality, and uses ChatGPT for numbers and spreadsheets, fact-checking its math.
One query replaces about ten searches. Editing a grief book, he asked Claude for rabbinic sources on why Jewish burial happens within three days and got two explanations, spirit and flesh, to verify the author and share with her.
Learning it properly would have taken a month. Because hallucinations are real, he verifies everything.
Dictating a bad first draft cures his anxiety. He talks into Whisper Flow when he feels off and polishes the transcript. He used it on an apology note to a friend he had dreaded for months.
His daughter writes freely because ego is not in the way; for adults it is fear of critics and Twitter.
He sees AI as the next word processor. He respects Robert Caro writing by hand but says he does not see how a modern writer could skip AI. He calls training on books borrowing to learn, like Google Books.
He thinks many critics have not used the tools and fear for their income, and that using AI well is a skill of its own.
How To (Really) Write With AI — Jimmy Soni
Jimmy Soni explains how a deeply researched biographer protects his writing hours, keeps research from becoming procrastination, and hunts for the overlooked sources and questions that make a story. His 4am routine of a reading on-ramp, a timed research window and a daily word count is the most usable part, along with his specific ways of using Claude as an editor and opponent.
He got his reps by volunteering first drafts for other people. He started by offering a draft of a colleague's big speech, and at McKinsey a draft he offered got a team 70% of the way there. He went on to write op-eds, speeches and ghostwritten pieces for business and Hollywood clients.
Do it often enough, he says, and people come to you to write. He writes thousands of words a day outside his books, and ghostwriting lets him put in reps without his name on the line.
The book took six years and a huge pile of raw material. He wrote about 470,000 words of raw material and kept about 170,000, and each chapter went through 17 or 18 revisions.
Each chapter had 13 Google Docs versions and then more than five PDF versions.
He walked PayPal's street to feel what the founders felt. He walked back and forth between 165 and 394 University Avenue in Palo Alto at different times of day, feeling the distance, the lunch walk and the weather between rival teams.
None of it made the book directly, but it showed him how unremarkable the place was. A friend calls him a method writer.
Soni looks at the toenails: what did everyone else miss? He borrows Michael Lewis's image that in near-identical Renaissance paintings the distinctive detail sits in the toenails. For Kobe Bryant that means the interview with 634 YouTube views, not the 81-point press conference.
He admits he is not sure the painting story is true, but the image stuck.
An odd name on a cap table produced his best Thiel story. He searched an unfamiliar name, Ed Bogas, on an early PayPal cap table, found a San Francisco number and left a cold voicemail naming Confinity. Bogas, a musician and chess player, called back and said Thiel's merciless chess convinced him to write a check on the spot.
Bogas called it the only serious investment he ever made. Soni's "theory of the case" is that great stories come when you find your Ed.
He opened his Musk interview with a question nobody else would ask. He asked about Greg Corey, an early adviser to Zip2 who died young, reasoning that grief casts a shadow that never leaves you. Musk paused and gave a minute-long meditation on the friendship.
People have lives outside the narrow frame we see them in, and asking about family or college often reveals the core thing.
He turns idle hours into research time. He listens to interviews about his subject while buying groceries, and during the PayPal years watched YouTube interviews with Thiel in the evenings instead of Netflix.
That lets research grow without eating into writing time.
He writes from about 4am, before the world can interfere. He writes from 4 or 4:30 until about 8 or 9am, seven days a week, with no email, admin or chores first. In book mode he will not schedule a 9am call.
He says the exact hour is not the point; it is time nobody else can claim. It also lets him be with his daughter without parent guilt.
The book gets his premium fuel. His sharpest hours are the first of the day, so the book gets them and everything else is ordered around it.
"The premium fuel is reserved for the morning."
Seven days a week turns finishing a book into math. Three to five hours a day will finish any book, he says, even if it does not guarantee a good one.
Once it is a default it stops feeling like work. On a vacation day without book time he felt off all day, in his body chemistry.
His morning runs reading, then a research timer, then a word count. He reads the model book for 10 to 15 minutes, does 30 minutes of research on a timer, then writes to a fixed word count, which defines a good day.
The reading is an on-ramp that tricks his brain into easy work before the blank page.
Love of the craft is the only fuel that lasts. Ambition burns clean for a while, then not so clean, and revenge burns worse, he says.
He points to Kobe Bryant, who journaled daily for 15 years and built his own publisher rather than accept a weaker book.
Other formats cross-train the main habit. He writes op-eds alongside books, and their roughly 700-word limit forces a word economy that carries into everything.
He compares it to a soccer player doing yoga, and thinks Twitter's character limits improved a generation of writers.
Soni treats endless research as procrastination dressed up as virtue. He knows someone who has spent two decades on a Plato book he never wrote. So when a section needs a fact, he digs it up and drops it straight into the draft.
"Research is an excuse not to write." He caps stand-alone research because the hunt is fun enough to need no schedule; what he schedules is making the thing.
Soni studies one model book obsessively per project. He read Brad Stone's The Everything Store 16 or 17 times and read 10 minutes of it each morning for about three and a half years. For the Kobe project his model is Naval's Almanack.
Stone traced where the name Amazon came from, so Soni dug up the origin of PayPal's name in slide decks. "I'm just Brad but for PayPal."
AI works middle to middle, and humans own both ends. He borrows Balaji Srinivasan's line: humans set the vision at the start and fact-check at the end, and AI helps in between.
An engineer friend put it as "the sharpest knife" but "I've still got to cook the meal." The words, research, facts and interviews stay his.
He has Claude write a savage takedown of his op-eds. He asks it to criticize the piece in 700 words and not hold back, then strengthens the draft against each objection. Nine times out of ten the counterarguments make him stop and think.
The idea comes from Tyler Cowen's point that you must be able to argue the opposite position.
He treats AI as an editor who never tires. At 4am he pastes in a flabby paragraph or a word that does not sit right and asks what is wrong or for other options.
He prefers Claude for writing because it has more personality, and uses ChatGPT for numbers and spreadsheets, fact-checking its math.
One query replaces about ten searches. Editing a grief book, he asked Claude for rabbinic sources on why Jewish burial happens within three days and got two explanations, spirit and flesh, to verify the author and share with her.
Learning it properly would have taken a month. Because hallucinations are real, he verifies everything.
Dictating a bad first draft cures his anxiety. He talks into Whisper Flow when he feels off and polishes the transcript. He used it on an apology note to a friend he had dreaded for months.
His daughter writes freely because ego is not in the way; for adults it is fear of critics and Twitter.
He sees AI as the next word processor. He respects Robert Caro writing by hand but says he does not see how a modern writer could skip AI. He calls training on books borrowing to learn, like Google Books.
He thinks many critics have not used the tools and fear for their income, and that using AI well is a skill of its own.
Soni paid for the basics publishers no longer provide. He paid for extra editing on almost every book and for his own cover on The Founders, because overloaded editors gave him table scraps of attention.
He stresses publishing is full of smart people, but consolidation leaves little room for flops.
He A/B tests every book cover. Services gather reactions from hundreds of people, because buyers decide in a snap on Amazon.
"Why wouldn't you AB test cover art?"
A two-week window kills slow-burn books. The Boys in the Boat debuted to crickets, then rowers passed it along and it hit the Times list about eight or nine months later.
He keeps his own books evergreen, so a reader discovering Claude Shannon today finds the book as new as at launch.
He co-founded Infinite Books to offer different deals. With Jim O'Shaughnessy he started Infinite Books, where some authors get 70% of royalties and no advance, against standard deals of 10 to 15% after an advance earns out.
He also wants publishers to run AI plagiarism checks before printing, calling it good, honest, responsible publishing.
A coffee table book taught him extreme word economy. After riding the Brooklyn carousel over 200 times with his daughter, he pitched Jane Walentas, who spent 30 years restoring it by hand, and Phaidon took the book. He rewrote every caption himself.
She filed away paint with X-Acto knives to find each horse's first color, and the book uses tissue overlays to show them. She died before it was done.
Kobe Bryant studied storytelling with basketball-level intensity. Soni watched Kobe interviews each morning during the pandemic and is now writing The Tao of Kobe, an almanac drawn from 400 to 450 interviews.
Kobe wrote Dear Basketball himself, recruited John Williams and Glen Keane, and became the first athlete to win an Academy Award. Asked if he had friends, he said, "What I have are fellow obsessives."
Caro shows how far immersion can go. Soni asks "What would Caro do?" of every story. Caro spends 10 years on a book and ran from the Capitol to see the dome light as Lyndon Johnson did.
Caro starts once he has the final sentence and writes toward it. Soni did not, and thinks it cost him time.
Michael Lewis shows the courage not to repeat yourself. Lewis could have written ten more finance books after Liar's Poker and wrote about coaching, parenting and baseball instead. At Salomon Brothers he wrote op-eds under his mother's maiden name, Diana Bleecker.
Soni took from Lewis's funny letters to friends the habit of asking how to make a friend laugh, and he does not want to be pigeonholed.
How to Write Like An Athlete | with Jimmy Soni | How I Write Podcast
Jimmy Soni explains how he writes research-heavy narrative nonfiction on mornings, nights and weekends, treating the work like athletic training with systems that beat the blank page and research paralysis. The habit worth hearing is his pipeline: dump raw material into a folder, paste it into a chapter doc, shape it through many versions, and use voice typing, a phone and outside readers when fear gets in the way.
Soni grew up a nervous reader in a family that knew libraries. He grew up in suburban Chicago in the 1990s, the son of immigrant parents who knew libraries rather than summer camps. Librarians held copies of Brian Jacques's Redwall novels for him, and he chased free pizzas through Pizza Hut's Book It program.
He calls himself a kid who was "afraid of everything", and says that anxiety never left. Good English grades first made him think he had some talent.
Writing for a real audience made him work harder. He sweated his college newspaper columns more than his coursework, because friends would notice any misspelling or weak argument.
He calls this one of the most important steps in leveling up as a writer.
He got his reps by volunteering first drafts for other people. He started by offering a draft of a colleague's big speech, and at McKinsey a draft he offered got a team 70% of the way there. He went on to write op-eds, speeches and ghostwritten pieces for business and Hollywood clients.
Do it often enough, he says, and people come to you to write. He writes thousands of words a day outside his books, and ghostwriting lets him put in reps without his name on the line.
Soni writes upward, learning the field alongside the reader. He knew nothing about PayPal or Claude Shannon before writing about them, and calls himself a "better than average charlatan" who learns as he goes.
Because his reader knows nothing either, every term of art has to be made readable, and the pleasure of finding things out stays in the book.
He commits to a book only when he sees an empty space on the shelf. He parks ideas in a Google Doc called "books I should write someday" and simply buys any book that already covers a topic. If he can picture the two books on either side of an unfilled gap, he writes a proposal.
Each book takes four to six years, so he asks whether the world needs it and whether he can spend years on it. He treats this as an entrepreneur's move and refuses to write "the sequel to the sequel".
Making a physical thing from nothing is what drives him. Six years before, there was no Founders book, and then there was a 400-page object.
"That act of creation lights me up," he says.
He saw that PayPal's founding years had never been properly told. Peter Thiel, Elon Musk and Max Levchin got constant coverage for what they did later, but the 1998 to 2002 creation story felt undercooked even after he read Zero to One, The PayPal Wars, Ashley Vance's Musk book and Reid Hoffman's books.
Every Thiel speech got a "Talmudic exegesis", he says, yet nobody had asked how they got their start.
The book took six years and a huge pile of raw material. He wrote about 470,000 words of raw material and kept about 170,000, and each chapter went through 17 or 18 revisions.
Each chapter had 13 Google Docs versions and then more than five PDF versions.
A plain spreadsheet and redundant backups kept the interviews under control. One Google Sheet tracked every interview with the name, date, pre-interview notes, audio file and post-interview notes. Audio was kept in two copies, one on a thumb drive, and transcribed with Otter AI for a flat fee.
He built the system as he went, and says nobody starts with the map figured out.
Soni records every interview, then debriefs himself right after. He brings only paper with his questions and leaves a phone recording on the table. After key interviews he types what stood out for a minute or two, then records a voice memo of his impressions.
Typing and speaking surface different memories. Only these notes preserved Thiel pulling out his phone mid-interview to check PayPal's market cap against Visa and MasterCard, a detail with no audio.
Musk thought out loud instead of reciting a script. The interview happened in Musk's living room on a Saturday afternoon, with food, and Musk paused and reworded answers he had clearly given before.
Soni says most people at that level are highly scripted because so much is at stake, which made this unusual.
PayPal's own writing became his best source. In hundreds of thousands of pages of emails he found the weekly company newsletter, where in 2001 and 2002 David Sacks wrote a 700 to 1,000 word product Q&A column answering employee questions.
Thiel prized clear writing and recruited Sacks partly for it. Founders also put their biggest moments in writing, such as Thiel's note on 9/11 and Musk's long resignation note.
He paid a fact-checker to demand a document for every claim. Ben Kallen spent six months on three rounds, and every fact or footnote needed a confirming document in a chapter's Google Drive folder. He flagged unsourced claims, wrong years and misquoted interview lines.
Publishers do not really pay for this, so Soni did. Kallen also helped decide what belonged in a footnote versus the main text.
He walked PayPal's street to feel what the founders felt. He walked back and forth between 165 and 394 University Avenue in Palo Alto at different times of day, feeling the distance, the lunch walk and the weather between rival teams.
None of it made the book directly, but it showed him how unremarkable the place was. A friend calls him a method writer.
A few words from Walter Isaacson launched the project. At an event for The Innovators, Isaacson heard the PayPal idea and called it a great book idea, and Soni left thinking "I can do this."
Later, on a walk in Central Park, Isaacson told him to be chatty in his endnotes and show how conclusions were reached. Soni's endnotes run long on purpose, down to a single YouTube video.
An odd name on a cap table produced his best Thiel story. He searched an unfamiliar name, Ed Bogas, on an early PayPal cap table, found a San Francisco number and left a cold voicemail naming Confinity. Bogas, a musician and chess player, called back and said Thiel's merciless chess convinced him to write a check on the spot.
Bogas called it the only serious investment he ever made. Soni's "theory of the case" is that great stories come when you find your Ed.
Forgotten people have not rehearsed their stories on stage. He interviews people who stayed "a hot second", answered customer service calls or fought fraud, because famous founders repeat talking points they have given 15 times.
"I can't make up that story," he says of what turns up.
Soni treats writing as athletic training: reps, recovery, discipline. He draws nearly all his writing metaphors from sports, having grown up with the 1990s Bulls, and asks what Kobe Bryant would do as a writer. The answer is a huge volume of writing.
He applies the same lens to Musk and Thiel, comparing Thiel's competitiveness to Michael Jordan's.
He writes from about 4am, before the world can interfere. He writes from 4 or 4:30 until about 8 or 9am, seven days a week, with no email, admin or chores first. In book mode he will not schedule a 9am call.
He says the exact hour is not the point; it is time nobody else can claim. It also lets him be with his daughter without parent guilt.
An identical ritual each morning removes decisions. He wore the same blue hoodie, flipped the lights on in the same order, drank cold sparkling water, and spent a minute picturing the finished book in his hands.
He worked every holiday and on his one vacation wrote before his daughter woke. He calls those mornings "awful and fun at the same time".
He cuts out contemporary media while inside a book. He told friends not to share news about the founders, unsubscribed from news newsletters and found out late about Musk hosting SNL. He read 300 printed pages of a 1990s blogger's posts instead.
He skipped the show Silicon Valley because it was satire, and prunes his newsletters monthly by asking whether each one delights him.
Soni treats endless research as procrastination dressed up as virtue. He knows someone who has spent two decades on a Plato book he never wrote. So when a section needs a fact, he digs it up and drops it straight into the draft.
"Research is an excuse not to write." He caps stand-alone research because the hunt is fun enough to need no schedule; what he schedules is making the thing.
He front-loads research in small timed daily blocks. His to-do list had 20-minute blocks: newspaper articles, PayPal emails, contacting interviewees, repeated every day.
After about 10 days on one source type you learn what it has to give. A biographer, he says, cannot know what to say until he knows what he has.
When he stalled, a coach imposed a painful penalty. Coach Lauren Zander diagnosed "analysis paralysis, research paralysis" and had him email her 1,000 words a day, with a $200 donation to a cause he hated for any miss. He hit every day for about 100 days.
He echoes Harry Lodge's line that humans are best in harness: constraints need real teeth, and a date beats a vague wish.
He never starts a chapter from nothing. Raw notes live in Scrivener folders by year; to start a chapter he pastes about 9,000 words into a Google Doc and shapes it down to roughly 5,500.
He sees writing as a sauce reduction rather than creation. A thin 2001 folder told him he had "phoned it in" and needed more research.
He dictates scenes with a manila folder over the screen. Using Google Docs voice typing with the monitor covered, he got down the McLaren F1 crash between Musk and Thiel without seeing typos or clumsy phrasing.
It works only after the research is done, when he already knows the facts.
He edits on his phone to tighten prose and lower fear. He did much of his best editing on his phone on waking, chapter four included. Padded paragraphs and ideas said 14 ways stand out on a small screen.
With notifications off, wandering to another app takes a conscious choice, and it feels like texting a friend.
Telling a story in shrinking time limits forces compression. He tells it freely, then in 2 minutes, then in 45 seconds, using voice typing.
People compress speech intuitively even when cutting five pages to one feels impossible.
He accepts that the fear never leaves. "Every day I'm still afraid of it," he says, so the job is building tricks that turn anxiety into output.
He wants writers to talk about fear more, since writing is often sold as glamorous.
Soni studies one model book obsessively per project. He read Brad Stone's The Everything Store 16 or 17 times and read 10 minutes of it each morning for about three and a half years. For the Kobe project his model is Naval's Almanack.
Stone traced where the name Amazon came from, so Soni dug up the origin of PayPal's name in slide decks. "I'm just Brad but for PayPal."
Every page should have a gem that makes readers smile, laugh or think. He borrows Apple's surprise and delight and Raymond Chandler's rule that something important happens on every page, so stale material 14 people had already said got cut.
A dull funding round became a joke: he recalled Musk's "funding secured" tweet and asked what Sequoia had actually invested in.
He asks what a reader in 2080 would still value. Imagining someone in 2080 or 2100 trying to understand Musk filters out petty office drama.
"Gossip has about a 5-minute shelf life." He says he gets it right about 95% of the time, and aims for books that last like The Power Broker and Seabiscuit.
Good history puts readers with people who did not know the ending. Following David McCullough, he writes as if the outcome were unknown, as with the Brooklyn Bridge builders.
He gave Musk at 19 warmth: too shy to speak in the Netscape lobby after an unanswered job letter, he went home, and then delayed Stanford by six months.
He writes the introduction last. Editor Alice Mayhew rejected an intro he sent before the book was written, saying he had not done the work to know the story's dimensions.
The intro is the reader's user guide, so it waits until the book is finished.
Writing is a team sport. He texted paragraphs to his friend Justin Richmond almost every morning for five and a half years, and a heart emoji or an honest "that's not that interesting" was a useful signal.
He tells newcomers to hire an inexpensive editor without ego, and likens his editors, fact-checker and readers to Phil Jackson around Michael Jordan.
Caro shows how far immersion can go. Soni asks "What would Caro do?" of every story. Caro spends 10 years on a book and ran from the Capitol to see the dome light as Lyndon Johnson did.
Caro starts once he has the final sentence and writes toward it. Soni did not, and thinks it cost him time.
Michael Lewis shows the courage not to repeat yourself. Lewis could have written ten more finance books after Liar's Poker and wrote about coaching, parenting and baseball instead. At Salomon Brothers he wrote op-eds under his mother's maiden name, Diana Bleecker.
Soni took from Lewis's funny letters to friends the habit of asking how to make a friend laugh, and he does not want to be pigeonholed.
The PayPal story taught him to value friends who push you. Levchin and Thiel were not best buddies, but each improved the other, and Musk and Sacks disagreed productively.
In his acknowledgements he tells his daughter he loves her too much to be her coach, so she needs others who will call her out.