The 1997 shareholder letter is the foundational document of modern operating culture. Bezos reattached it to every annual letter for twenty-three years. It opens with a line he made into a religion: it's still Day 1. Day 2, he wrote in 2016, is stasis. Followed by irrelevance. Followed by painful decline. Followed by death.

The point is not that Day 1 is aspirational. It's a diagnostic. Are we still inventing? Are we hiring missionaries? Are we measuring outputs, or have we slipped into measuring process? Every principle below exists to answer that question with evidence, not vibes.

SECTION 01Stubborn on vision, flexible on details

The vision for Amazon — earth's most customer-centric company, vast selection, low prices, fast delivery — has been unchanged since 1997. The how has changed continuously: 1P retail to 3P marketplace, books to everything, retail to AWS, owned logistics, Prime, Kindle.

If you're not stubborn, you'll give up on experiments too soon. And if you're not flexible, you'll pound your head against the wall and you won't see a different solution to a problem you're trying to solve.— Bezos to Lex Fridman, 2023

The operating mechanism: keep a small set of unchanging tenets at the top, and let everything below be cheap to revise.

SECTION 02The six-pager and the silent study hall

Bezos banned PowerPoint from S-team meetings around 2004. The replacement is the most-copied artifact in modern business: a six-page narratively-structured memo, read silently together at the top of each meeting.

The Mechanics

• Six pages, narrative prose, complete sentences. No bullets. Appendices and data go in unlimited attachments after page six.
• Silent reading at the top of the meeting — typically 20 to 30 minutes. Nobody pre-reads, because nobody actually pre-reads. Reading together guarantees a level playing field.
• No author defense during read. Discussion begins only after everyone has finished. The author usually speaks last.
• Authorship is anonymous on the cover. Quality of thinking, not status, carries the room.

A great memo, in Colin Bryar and Bill Carr's telling, takes a week or more. The first draft is never the final draft. The artifact is the thinking. The cost is paid up front by the writer; the savings accrue forever to every future reader.

SECTION 03The customer is the primitive

Bezos draws a hard distinction between customer-focused and competitor-focused companies. The first lets you pioneer. The second forces you to wait. The mechanism that makes the difference is not posture, it's procedure.

The Working Backwards Process

Before a single line of code or spec is written, the product team writes:

1. A mock press release — written in the voice of an external announcement, dated to launch day. One page. Headline, problem, solution, quote from a delighted customer, how to get started.
2. A FAQ — internal and external. The hard questions: what could go wrong, what's the unit economics, what's the customer's alternative today, what does "done" look like.

If you can't write a press release that makes a customer's life visibly better, you don't have a product. The PR/FAQ is iterated for weeks before any engineering commitment. The process is intentionally slow at the front so that everything downstream can be fast and correct.

Customers are always beautifully, wonderfully dissatisfied. Even when they don't know it. That divine discontent is what pulls invention forward.— Bezos to Lex Fridman, 2023

SECTION 04Doors, velocity, and disagree-and-commit

The 2015 and 2016 letters formalize a doctrine of decision quality and decision speed that has aged better than almost any other framework from the era.

Type 1 vs Type 2

One-way doors are irreversible and demand methodical deliberation. Two-way doors are reversible and should be made quickly by high-judgment individuals or small groups. The failure mode Bezos names: applying Type 1 process to Type 2 decisions.

Disagree and Commit

If you have conviction even without consensus, you say: I know we disagree on this but will you gamble with me on it? Crucially, the principle flows both directions. Bezos describes greenlighting an Amazon Studios original despite his own disagreement, because his team was unanimously bullish. The most senior person in the room is the one who most needs the discipline.

SECTION 05Two-pizza teams and the API Mandate

The original two-pizza rule is famous, but the real insight wasn't team size. It was autonomy. Each team must own a discrete piece of the business with clear input metrics, and ship without coordinating with other teams.

The 2002 API Mandate memo — the seed of AWS — made this enforceable in software. The operative points were direct: all teams will expose data and functionality through service interfaces. There will be no other form of inter-process communication allowed. No direct database reads. No back-doors. All service interfaces, without exception, must be designed from the ground up to be externalizable. Anyone who doesn't do this will be fired.

Bezos turned organizational coupling into a technical contract you could measure and version. Later he refined this with single-threaded leaders: for any important initiative, one leader dedicated to it and nothing else. The best way to fail at inventing something is by making it somebody's part-time job.

SECTION 06Standards are teachable

The 2017 letter is the definitive text on standards. Bezos makes four claims: standards are teachable; they are domain-specific; recognizing the standard requires realistic expectations of effort; high standards are contagious in both directions.

Most people think if they work hard, they should be able to master a handstand in about two weeks. The reality is it takes about six months of daily practice.— from the 2017 letter, on a friend who hired a handstand coach

The hiring mechanism is the Bar Raiser: a trained interviewer from outside the hiring team, with veto power, insulated from the urgency pressure of the hiring manager. The principle: every person hired should raise the bar — the average employee should get better after each hire.

SECTION 07The seven-year horizon, and the war on proxies

If everything you do has to work in three years, you're competing against a lot of people. If you'll invest on a seven-year horizon, you're competing against a fraction of them. Most really interesting things take long horizons.

The deeper thread isn't frugality or patience. It's the war on proxies, articulated most sharply in the 2016 letter.

Process becomes a proxy for outcomes. Metrics become a proxy for truth. Surveys become a proxy for customers. The job of the leader is to keep their hand on the actual thing.

Tim Cook inherited a founder-driven company from one of history's most magnetic CEOs and, over fifteen years, quintupled revenue and built the most valuable company in the world. Not by being louder than Steve Jobs, but by being more systematic, more prepared, and more patient. The principles below are the operating substrate. They are noticeably plain. That's the point.

SECTION 01Operations excellence as the foundation

Cook arrived at Apple in 1998 from Compaq with a thesis he had been working on since IBM: inventory is fundamentally evil. His first big move was to compress Apple's inventory from months to days, close warehouses, and re-architect the supply chain around just-in-time and contract manufacturing.

The principle was simple and ruthless: find the world's best at something and partner with them; don't do everything yourself. Apple would do silicon design, software, and industrial design in-house. Almost everything else — glass, batteries, assembly, logistics — would be done by the single best vendor on earth, on Apple's terms. This is sometimes mis-summarized as outsourcing. It is closer to concentration. Cook didn't spread risk across many vendors. He picked one and squeezed.

That story is the cultural artifact at Apple. Translated: a decision is not a decision until it is moving. Distance between identifying a problem and being on the problem is measured in minutes, not days.

SECTION 02Discipline and preparation

Cook is up around 3:45 to 5:00 a.m. every day. He spends the first hour reading email — by his account 700 to 800 messages a day, the majority from customers and external "real people." He goes to the gym next. He is, famously, the last car in the parking lot.

I like to take the first hour and go through user comments and things like that. That, for me, kind of grounds me.— Cook to GQ, 2023

The morning email habit is not just vanity. It's a deliberate humility loop: the CEO of the most valuable company on earth begins every day reading complaints, edge cases, and praise from people who don't know he is reading. He occasionally forwards customer mail to the relevant executive with no commentary, which inside Apple is understood as a summons.

The discipline extends to meetings. Cook is the only Apple executive known to over-prepare relative to his own staff. He arrives at reviews with binders printed, data memorized, and follow-up questions that would not occur to the person who actually owned the project. The most senior person in the room should be the most prepared, not the least.

SECTION 03The functional org and the DRI

Apple is organized by function, not by business unit. There is no "iPhone division" with its own GM and P&L. There is hardware engineering, software engineering, services, operations, marketing, retail — each led by the best in the world at that function. Only the CFO holds a P&L. Jobs originated this. Cook institutionalized and protected it.

The general manager structure is bullshit. It creates fiefdoms.— Jobs, paraphrased in Inside Apple

Inside the functional org, accountability runs through the Directly Responsible Individual — the DRI. Every action item on every agenda has one name next to it. "Who's the DRI on that?" is the most common question in the building. No DRI, no decision. No decision, no progress.

Two implications worth stealing: specialization beats coordination — a great optics engineer works on optics across all products, she is not split. And one throat to choke, per thing — a DRI is one human being whose name is on it, not a committee, not "the team."

SECTION 04Slow on inputs, fast on outputs

Jobs made snap calls. Cook does not. Cook is famous for prolonged, often uncomfortable silence in meetings, followed by a chain of why questions that walk a problem back to first causes. Former direct reports describe the experience as forensic — almost prosecutorial.

Cook's decision style has three observable traits. He sits on questions far longer than peers think reasonable — and then, once committed, reorganizes the company behind it. He asks for the data, not the conclusion; presenters who lead with the recommendation get walked back to the inputs. And he doesn't perform certainty — he will openly say "I don't know," which raises the floor for everyone else to admit the same.

SECTION 05Quiet leadership and operationalized values

Cook deflects personal credit obsessively. He almost never says "I" in keynotes — it is always "the team." When Bloomberg Businessweek asked him about being the first openly gay Fortune 500 CEO, his response was a deliberate redirection.

If hearing that the CEO of Apple is gay can help someone struggling to come to terms with who he or she is, then it's worth the trade-off with my own privacy.— Cook, Businessweek 2014

The public values posture — on privacy especially — Cook treats as an operating decision, not a marketing one. The principle is not "have values." Every company says that. The principle is: state the values publicly, then let the operating model close off the option of betraying them. Apple's privacy stance constrains its ad business, its data infrastructure, even its acquisition targets. The constraint is the point.

SECTION 06Patience as a capital strategy

Steve Jobs hoarded cash because he remembered Apple nearly running out. Cook inherited the hoard and treated it as a tool. Over his tenure he has returned roughly $1 trillion to shareholders via buybacks — the largest share-repurchase program in corporate history. And he has refused, repeatedly and publicly, to make a marquee acquisition. Apple's largest deal under Cook is Beats at $3B — a rounding error.

Cook's framing: Apple acquires capabilities, not companies. The operating principle is that patience is a strategy. Cash on the balance sheet is optionality. Buybacks are an admission that capital cannot be deployed at Apple's internal rate of return — a confession most CEOs are too proud to make.

Jensen Huang has structured NVIDIA in a way that violates almost every textbook on management — and he has done it deliberately. The point isn't the sixty direct reports. It's that the structure is engineered so information flows fast, in writing, in public, and decisions get made by whoever is closest to the truth.

SECTION 01The flat org and the T5T

Jensen has roughly sixty direct reports. He does zero one-on-ones. He calls 1:1s "a place to hide." The reasoning is operational, not stylistic.

My direct reports are the best in the world at what they do. They need less information from me than anybody else in the company. They've reached their position because they're really, really good at what they do.— Jensen, Acquired

Two reasons for the no-1:1 doctrine: information held privately is information lost to the rest of the org; feedback given to one person is a free lesson for the other fifty-nine.

Top-5 Things

Every employee, at every level, sends Jensen — and their manager, and broadly — a short weekly note: top five things I'm working on, top five things I'm observing. Competitive moves, customer pain, project delays, market signals. Jensen reads about a hundred a day before breakfast. He pattern-matches across hundreds of T5Ts to triangulate where the company actually is.

He reaches down arbitrarily deep, skips levels constantly. The org chart is for paychecks, not for information. And he criticizes publicly — not from cruelty, but because criticism in front of the team allows the entire company to learn from one person's mistake.

SECTION 02Strategy is action

Jensen's strategy doctrine is unusually action-biased and almost anti-narrative.

The T5T system is the auditing mechanism. He literally checks: do the top-5 things people are doing add up to the stated strategy? If not, the stated strategy is fiction.

Whiteboards over slides

Almost every important meeting is on a whiteboard. "There's no place to hide on a whiteboard." Slides reward polished memory; whiteboards reward live reasoning. Everyone has to be willing to erase their own ideas. From Tae Kim: it's immediately apparent who hasn't thought things through.

Planning is a verb

NVIDIA does not have a 5-year plan. "We don't do long-range plans because the world changes. We plan constantly. Planning is a verb." Strategic re-planning is continuous because input data — T5Ts, customer signals, competitor moves — is continuous.

SECTION 03Pain, suffering, and high standards

Jensen sounds most uncompromising on the topic of standards. When he said this at the Stanford Graduate School of Business, students were visibly uncomfortable:

The principle isn't sadism. It's that resilience is the only true input to greatness. He frames high standards as kindness: "The lowest expectation in life produces the lowest result. If you want a high-performing company, the first ingredient is high expectations."

And the no-firing tradition: "I don't like giving up on people. I'd rather torture them into greatness." The standard isn't lowered. But the door isn't closed either. NVIDIA's voluntary attrition is reportedly 2 to 3% versus 15% Silicon Valley average. People stay because the bar is the bar, Jensen actually keeps you when things are hard, and the mission is unambiguous.

SECTION 04The CUDA bet, and other bet-the-company decisions

NVIDIA spent more than $10B over a decade-plus making every GPU CUDA-capable — even consumer gaming cards — when there were no customers. For ten years Wall Street asked: why are you doing this?

Because if accelerated computing is real, we have to have the installed base ready when the workload arrives.— Jensen on the CUDA bet

Jensen counts at least four moments NVIDIA "bet the company": Riva 128, the move to programmable shaders, CUDA, and the data-center pivot. His framing isn't bravery. It's necessity: "You don't bet the company because you feel brave. You bet the company because the alternative is being slowly killed by someone who did bet."

He invests against future workloads — deep learning, robotics, biology — before product-market fit exists. By the time the workload arrives, NVIDIA is the only platform that runs it. This is the inverse of customer-development orthodoxy. He calls it market making, not market sharing.

SECTION 05Teach the same thing a thousand times

Jensen is the clearest explainer in his company by design. He thinks out loud, on whiteboards, in front of sixty people. Decisions are not delivered as edicts; the derivation is shared so the org learns the model, not just the answer.

Maxims have to be repeated 1,000 times before they become culture.— Jensen, on the CEO's job

He works seven days a week. "If I'm not working, I'm thinking about working. Working is relaxing for me." The intensity sets the ceiling on the org's intensity. And the paranoia is the engine: "We are always 30 days from going out of business."

Brian Chesky's foundational claim is that the standard scaling advice he received between Airbnb's Series A and the pandemic was wrong. Well-meaning people told him he had to run the company in a certain way for it to scale. The advice could be summarized as: hire good people and give them room to do their jobs. He followed it. The results, in Paul Graham's word, were disastrous.

SECTION 01Presence, not absence

The fix came from explicitly studying Steve Jobs's second act at Apple, and ongoing conversations with Jony Ive and Hiroki Asai — the former Apple creative lead Chesky later hired.

What Founder Mode is not, in Chesky's framing: not tyranny, not refusal to delegate. He still delegates. But the borders of autonomy are narrower and earned over time. Skip-levels are normal, not exceptions. The org chart subtree is not a black box.

SECTION 02The functional org, one P&L

After the 2020 reorg, Chesky eliminated general managers, business-unit P&Ls, and the guest/host divisional split.

We are a functional organization. Functional just means expertise-based, not general-management-based. I'm the only non-functional person in the company. All functions roll up to me.— Chesky, Decoder

If we were a military, like a battalion, the cavalry general should know how to ride a horse. It's crazy that they don't. And leaders shouldn't be fungible.— Chesky, Decoder

The consequences at Airbnb: one P&L. No separate "Homes" or "Experiences" or "Services" businesses. One company, one roadmap, one functional leadership team. No separate guest and host product teams — product is organized by function (search, listing, pricing, trust) and serves both sides. Marketing reports as a leadership function, not a service function. PMs were dramatically reduced post-2020 and reframed closer to product marketing managers in the Apple sense — owning narrative and launch, not running mini-fiefdoms.

SECTION 03Where the calendar actually goes

Chesky's calendar is the most visible proof of Founder Mode. He spends the majority of his week in design, product, and marketing reviews, not 1:1s with direct reports about their teams.

• Weekly product reviews. Sunday-prep-into-Monday review of work in progress across the entire product surface. He personally critiques flows, copy, illustrations, photography, animation timing.
• Weekly marketing and creative reviews. Brand campaigns, the homepage, the annual brand film — he is in the room and the editor of last resort.
• Two release planning rooms per year that consume large blocks of his time.

His test for whether a leader belongs on his team: can they show me the work, not just the dashboard? If they only show dashboards, they're managing a black box. Chesky considers that the warning sign of organizational decay.

SECTION 04Two releases a year

Airbnb deliberately rejected continuous, A/B-tested shipping in favor of two coordinated releases a year — May (Summer Release) and November (Winter Release). Chesky's rationale:

• Coordination over local optimization. Everyone rowing in the same direction. A functional org with one P&L has to ship in synchronized waves.
• Narrative compounding. Each release is a story the company can tell at scale. Continuous shipping produces no story.
• Quality forcing function. A six-month cycle gives design, engineering, and marketing time to polish.
• Rolling two-year roadmap. Long enough to plan ambitious bets, short enough to stay accountable.

On performance marketing and A/B testing, Chesky is openly skeptical. Airbnb cut performance marketing dramatically and saw direct/organic traffic remain ~90% — the spend was largely cannibalizing demand they already had. A/B tests, in his view, optimize for local maxima and systematically under-weight bold, brand-defining bets that look bad on a two-week test but compound over years.

SECTION 05Taste as a leadership function

Chesky's decision-making is taste-driven, top-down on craft, and unembarrassed about it. He treats marketing as a leadership function, design as a first-class peer to engineering, and the CEO as the final editor.

Reviews start with the work — a mock, a flow, a video, a draft email — not a deck about the work. He personally edits launch copy, blog posts, the homepage, and earnings narratives. When a function isn't working, his instinct is to recruit a world-class operator and let that person carry the standard, not to add management layers underneath.

SECTION 06The 2020 inflection

The COVID collapse was the moment Chesky stopped trying to run Airbnb the conventional way. By his own account, the pre-2020 Airbnb was a fast-moving company becoming a slow-moving bureaucracy. The numbers: from 7,500 to 5,500 employees. Around 1,900 laid off. 2020 revenue forecast at less than half of 2019. $2B raised in debt.

The May 5, 2020 layoff memo became a Silicon Valley reference document. Guiding principles: map reductions to future strategy, not seniority or politics; do as much as possible for those impacted; be unwavering on diversity; optimize for 1:1 communication; wait to communicate decisions until all details are finalized.

In a crisis, it's all about optimism. You have to have optimism that's rooted in reality, that's believable, so people will follow you up a mountain.— Chesky, TIME

What the reorg actually changed: killed GMs and divisions; killed separate guest/host product teams; killed performance-marketing dependency; re-centered the company on hosts, homes, and experiences; installed the two-release cadence; pulled Chesky back into product, design, and marketing as the de facto CPO.

Eric Jorgenson's thesis in The Book of Elon, distilled from millions of words of Musk's own tweets, interviews, and public talks, is this: almost everything Elon does is replicable. His advantages multiply rather than add — first principles times speed times mission, compounded for thirty years. The principles below are extractable from the lifestyle. The 20-hours-a-day-or-you're-not-dedicated stuff is not, and the book treats them separately on purpose.

SECTION 01Maniacal urgency, the master principle

The operating mechanics behind the slogan: always attack the bottleneck. Travel physically to whatever the limiting factor is. He fired his personal assistant scheduling him — just responds to fires.

Lost-future-revenue math

Early SpaceX calculated roughly $10M a day in lost future revenue per day of delay. Once flew a $100,000 private jet to deliver one part because $50M of engineers and rockets were idle. Most people are penny-wise and pound-foolish — waste weeks to save thousands.

Fifty-percent deadlines

Most leaders set deadlines they expect to beat — a 100% hit rate is too conservative. Elon sets deadlines with about 50% probability of being hit. Sometimes misses. Often hits "impossible" ones because of aggressive scheduling. "If a timeline is long, it's wrong."

The daily routine: wake up, check phone for an emergency. If none, create one. Brent Beshore's analogy that Elon endorses: operating a business is a knife fight. Wake up, knife off the bedstand, go to work.

SECTION 02The Algorithm

The single most quotable operating procedure in modern business. Elon repeats it at every all-hands. Order matters.

1. Question every requirement. Every requirement must have a named human owner. No department-owned requirements. If the owner can't defend it, kill it.
2. Delete the part or process. If you're not adding back about 10% of what you delete, you weren't aggressive enough.
3. Simplify and optimize. Only after deleting. The classic failure is optimizing things that shouldn't exist.
4. Accelerate cycle time.
5. Automate. Last, never first. Elon's own Fremont mistake during the Model 3 ramp was running this in reverse — automated parts that shouldn't have existed.

Steps 1 and 2 produce roughly 80% of the value. Most companies skip straight to step 4 or 5.

SECTION 03First principles, and thinking in the limit

Decompose to physics or raw materials, rebuild from scratch. Tesla batteries: analyzed raw materials, found costs far below the industry's assumed price. SpaceX rockets: broke down to raw materials, exposed contractor markups.

"Assume you are wrong; aspire to be less wrong."

The cost test at the limit: would this still be costly at a million units per year? Separates real design flaws from low-volume effects. Ban "impossible" in critical discussions. Design for the platonic ideal, then build the tools backward.

SECTION 04Action produces information

Jorgenson's most underrated principle. Elon does not believe in slide decks.

Get to a touchable, shitty demo as fast as possible. Not a slide deck. Shitty demos work. An early Roadster that went about 10 mph got Larry Page and Sergey Brin to invest. Daimler executives flew in for what was supposed to be a boring meeting — the team retrofitted a Smart Car with EV batteries in the parking lot in days. Daimler wrote a $50M check.

SpaceX started with Grasshopper — a hopping prototype — not Falcon 9. Even "pointless" surges work: stacking Starship before it was ready to fly. Manifesting the thing physically makes people sprint to make it real.

SECTION 05Build the machine that builds the machine

"If we don't make stuff, there is no stuff." The line he delivered on Joe Rogan during the COVID era. The one that sparked Jorgenson's book.

Elon's biggest Tesla epiphany: "the machine that builds the machine" — the factory — is the product. Manufacturing has more innovation surface than product design. Designing a rocket or car is trivial compared to volume production.

And the inversion most companies miss: charge less, not more. Bezos's framing: two kinds of companies, those that work hard to charge more and those that work hard to charge less. Tesla and Amazon are the latter. Elon: "If we charge something, it's because we can't figure out how to charge less."

SECTION 06Talent, excellence, and the lieutenants

Special Forces model. Small strong team beats large mediocre team. The constraint isn't capital — it's exceptional engineers. "Excellence is the passing grade."

99.9% of firings err on keeping people too long, not letting them go too early. "Empathy is not an asset" — empathy for an individual at the cost of the team is a failure mode. Hire on hard problems overcome, not credentials. Character matters as much as intellect.

Tour-of-duty mindset: 80% turnover at each order of magnitude of growth is normal. "The reward for work is more work."

Install the operating system in 20 lieutenants

Install your operating system in 20 to 25 lieutenants who can run the playbook without you. (MrBeast does the same — new hires shadow him 24/7, live with him.) Proof: Martin Kosa flew to Starlink, war-roomed, delivered a two-order-of-magnitude improvement in 10 to 12 months on a problem industry veterans had been stuck on for years.

SECTION 07Risk, purpose, and the fuel underneath

"Failure is irrelevant unless it's catastrophic."

Thiel framing: Thiel is a risk manager; Elon is a risk taker. He's biased toward risks he shouldn't take, but mission carries him through. "Fear of failure is the biggest cause of failure." Kills more things in the crib than failure itself does. Initial Tesla plus SpaceX odds were about 10% each — roughly 1% both succeed. Sequoia passed.

Picks missions at civilizational scale: internet, then sustainable energy, then space, then AI. Rejects zero-sum thinking — grow the pie. Does not optimize for happiness. Does not celebrate past accomplishments. Always next.

Jorgenson's diagnosis: Elon burns clean fuel (mission idealism) and dirty fuel (childhood angst — "I'm a piece of shit, must do more") simultaneously.