How to build a company that withstands any era | Eric Ries, Lean Startup author
In a Nutshell
Eric Ries' *Incorruptible* warns that success breeds "financial gravity"—a structural force corrupting companies via shareholder primacy, eroding missions and ousting founders (e.g., only 20% of VC-backed founders remain CEO 3 years post-IPO). Protect via **ethos** (internal alignment on purpose/fiduciaries like Cloudflare's "better internet") + **integrity** (structures like Public Benefit Corporations, industrial foundations as in Novo Nordisk/Anthropic, director oaths, founders preferred shares). Key actions: File as PBC now with a tested purpose, add director oaths, secure mission-protected provisions—**harder is easier**, yielding trust, flow states, and superior longevity/performance.
These notes were generated by AI and may contain inaccuracies.
Famous companies destroyed not by competition but by their success becoming a liability. OpenAI vs. Anthropic story: Dario was a first-time founder, not a hot company, pre-ChatGPT boom. True believers in AI safety mission. Investor suggested they talk to Eric Ries. He warned them: if you don't get this right, here's what will happen. They wrote into their charter: Anthropic has directors on its for-profit board appointed by and accountable to an outside group of trustees who are AI safety experts without equity in Anthropic. When Anthropic refuses to release a model because it's too dangerous, consider the cost.
New book Incorruptible helps protect what you've built. The force no one controls but everyone obeys drags organizations into mediocrity, leading to loss of control. Broadstroke solution: Harder is easier. Being principled in decisions yields unexpected rewards. Most leaders can't defend principles due to ROI-based thinking and shareholder primacy—path of maximum profitability, which is flawed.
Eric Ries, author of The Lean Startup, returns with Incorruptible: Why Good Companies Go Bad and How Great Companies Stay Great (15 years later). Lean Startup about building success; this book about protecting it. Eric seen too many founders lose control, leading to disappointment and depression—a problem every successful founder faces. Shares stories, tactics, advice for building lasting companies.
Eric Ries thanks for return; congrats on podcast growth. Lean Startup (15 years ago): most impactful startup book. Endured waves of praise, backlash ("doesn't work anymore"), back to validation. Top AI companies build exactly like lean startup: e.g., head of product at Claude ships MVP as "research preview," iterates based on feedback.
AI labs didn't predict popularity of ChatGPT, Claude, etc.—small experiments, not big bets. Classic product development: can't predict future. Treat as hypotheses to use scientific method.
OpenAI, Anthropic, Cursor, Vercel, Replit, Sierra, Clay, hundreds of companies powered by WorkOS. For enterprise products: integrates SSO, SCIM, RBAC, audit logs. Turns deal blockers into APIs. Stripe for enterprise features. Used by Lenny's portfolio startups. Visit workos.com.
Incorruptible very personal. Protection from force dragging organizations to mediocrity/loss of control: getting fired, company becomes bureaucratic/evil (Frankenstein's monster). Eric's front-row seat: helped create wealth, proud, but seen darkness.
Anecdote: Dinner at restaurant recommended as good; friend checks phone after one bite—detects private equity takeover. "I could taste it." Many similar stories. Famous brands ruined. Hundreds of years of examples: success tempts to "butcher the golden goose."
Vital Farms eggs example: pasture-raised organic, now complaints of toxins, owned by BlackRock. Natural foods brand named after founder: ousted for higher growth/margins, led to lower quality, pissed customers/employees, shrinking market share. Pattern pervasive.
Calls it corruption (not legal bribery): like bridge collapse not just from gravity (financial gravity), but corroded bolts. Study load factors. Solution: stainless steel bolts. Book: organizational equivalents of stainless steel.
Book structure: mystery. Why does this happen despite markets selecting for value? Why exceptions if inevitable (greed/age/size)?
Not ethical weakness—structural. If not addressed, no other decision matters long-term; you'll lose control. Harvard Law School: among venture-backed companies with standard best practices, only 20% of founders remain CEO 3 years post-IPO.
Story: Hot pre-IPO company seeks long-term IPO structure advice. Had best practices—Eric: "you're screwed." Showed data/cases. Founder planned fixes, but bankers/lawyers/CFO/GC/VCs called Eric a "downer," said they're exception. IPO successful, then competitor acquisition causes stock collapse; founder ousted after 5 months. Critics blamed mistakes, but investors backed it 5 months prior.
Collapses everywhere, called normal. But choices: internal (cultural/management) and external (governance) structures weak.
Companies "killing it" without protections? Check math: Cloudflare, Costco do have protections (governance fortress). Many mission-driven.
Key: Timing. Best time to plant tree was 40 years ago; next best now. Always too early until too late.
Sequence founders hear "later":
- Incorporating: Lawyer says get product-market fit first; success is leverage.
- Raise money: VCs agree but later.
- Growth round: Growth VCs say might hinder fundraising.
- IPO prep: Bankers/lawyers say bundle with IPO.
- IPO roadshow: CFO: "You were serious? Too late."
Never right time. Success makes you target. IPO advisors profit from transactions up/down; onto next.
Blueprint: ethos (internal alignment, character choices) + integrity (structure to resist drift). Part 1: The Shape of the Abyss—grim stories (Whole Foods, Verra). Grapple to prepare.
Double mystery: why happens, how exceptions exist. Stories as database for principles.
Ethos meaning internal alignment, character choices; integrity meaning the structure to resist and keep ourselves aligned with human flourishing.
In 1920 Denmark, Marie Krogh was one of the first credentialed doctors and a big advocate for women's medical education. She got diagnosed with diabetes, a fatal illness with no known cure at the time, right as her husband August Krogh won the Nobel Prize. He asked if she would join his North America lecture tour; she said yes. At dinner, Marie sat next to a scientist who mentioned Canadian researchers isolating insulin as a potential cure. They detoured to Canada, understood the implications as scientists—this could save Marie's life and millions more. They asked to commercialize it; Canadians were open but concerned about exploitation and extraction instead of value creation.
Imagine depending on the only insulin maker for a life-saving cure—you'd want a fair price for moral reasons and incentives to keep producing. But temptation arises to charge anything, especially with investors whispering. Decades before Martin Shkreli, the Kroghs and Canadians worried about this. Back in Denmark, they created Nordisk Insulin Laboratorium, a for-profit with outside investors owned and governed by a nonprofit foundation—a two-tiered industrial foundation structure. This predecessor became Novo Nordisk, one of the world's largest companies. This structure endured, protecting scientific integrity for over 100 years.
Advisers push best practices younger than trees in your local park. Next time someone pushes ROI-dominated practices like stack ranking by ROI, ask if they're smarter than a Nobel laureate—August and Marie figured this out 100 years ago, withstanding attacks. Trustees once intervened to prevent selling out, creating over $500 billion in shareholder value.
This is one of the most powerful engines of value creation in the world.
Not just ethics or mission as nice-to-have. German optics company Zeiss used this structure in 1885 (lenses in glasses). Academic research shows these companies are six times more likely to reach year 50, with superior return on invested capital and better performance overall. August and Marie started as a tiny startup; now worth hundreds of billions. These techniques counter financial gravity.
Exercise: Name the most evil company—no amount of money would make you work there (examples: Halliburton, tech companies, Monsanto, private equity, Philip Morris). Imagine your company gets a bid from them for $1 more per share than current worth, to sell cigarettes to children. Most say hell no. But legal documents and charter impose fiduciary duty to say yes. Founders call lawyers, confirm it's true—"best practice" to raise money easily.
Real case: Vector Corporation, UK spinout from University of Bath, made inhaler therapeutics for asthma and COPD. Successful, raised money, went public on London Stock Exchange. Philip Morris bid to buy for diversification beyond nicotine (inhaler expertise). Bids: Philip Morris at 165p/share, US private equity at 155p/share, or stay independent. Public outraged; British Thoracic Society begged no. Board had two meetings, said hands tied by fiduciary duty to highest bid—accepted Philip Morris. Philip Morris spent 1.1 billion pounds; within three years, $900 million write-down, disposed for parts. Company ceased to exist.
Delaware charters are public—check if a company has mission protection. Stakes are existential, especially in age of powerful planet-scale technologies. Founders want products they're proud of, fortunes grandkids respect—not ruined babies.
Book out May 26. OpenAI on minds for nonprofit structures; Anthropic story in book (speaker's bit part). AI is humanity's final exam—issues amplified, but start with simpler businesses.
Harder is easier principle: Anyone with influence can adopt. Business feels hard; consider if lack of trust makes it harder. Evidence: Trustworthy companies spend less on employee communications, better alignment; customers have higher loyalty, lower acquisition costs, stick after mistakes, try new products.
Commit upfront to quality, design, ethics, integrity, safety—your values. Principled decisions yield unexpected rewards. Do it for the thing itself, not rewards—trustworthiness is most underrated asset. Right things rank low by ROI (tangible costs, intangible rewards).
Cloudflare story: Founders anti-consulting BS, no mission/values early (traumatized from Harvard Business School). Mission emerges from living superorganism, not label. Build in quality (Deming 1940s). Cloudflare defended pro-democracy protesters' sites from nation-state hackers for free (free tier, no reward).
Lunch: Engineer said, "first place where we're trying to make a better internet." Became mission statement; top value: be principled.
Junior engineer to CEO Matthew Prince: Our mission is better internet; top revenue driver is SSL encryption (upgrades from free to premium)—expensive due to certificates, compute. Wouldn't better internet be encrypted internet? Why not free? Matthew: "Let's figure it out."
Figure it out principle: Committed leaders revel in dilemmas to teach values. Couldn't bankrupt company; rallied team, hand-rolled software/assembly language, bizdev with certificate authorities, drove costs down. Could've quit anytime (too hard, free margin, board saw lower conversions)—stuck with it because
Cloudflare stood for encrypted internet, increasing top of funnel by an order of magnitude. People credit Cloudflare for the encrypted internet we take for granted, contributing to its $70 billion valuation. Most leaders cannot defend their principles due to ROI-based thinking and shareholder primacy.
Andrew Mason, founder of Groupon, shared a story. Groupon grew rapidly as a private company with one daily email featuring a cool deal, going public on that model. Executives pressured him: "We need to make the quarter, consider two emails." He resisted initially, as one email was their core. They pushed with lean startup language: experiments, data, ROI. He ran the test; two emails made more money. They escalated to three, then eight emails. This pattern repeats across CEOs with email frequency, destroying companies short-term for profit.
Harder is easier principle defends what makes a company special.
Writing mission statements or values is not valuable. Use ethos: identify fiduciaries—who would you rather die than betray? Every product needs a purpose.
Steve Jobs fought over wire layout inside computers customers wouldn't see. Yvon Chouinard of Patagonia was a quality zealot, believing quality is objective; every product deserves a certain level. Encode purpose in management system so you cannot profit by betraying it, at operational and governance levels.
Purpose counters ESG skepticism. Unilever infused purpose into products; a Wall Street investor wrote: > "look at the point that we're debating the purpose of Helman's mayonnaise, I think you've lost the plot."
Helmann's mayonnaise purpose is clear: it's food. Product managers' purpose matters. Efficiency consultants suggest saving costs by making products carcinogenic; without purpose, post-vesting exploitation happens. Johnson & Johnson put asbestos in baby powder and covered it up because actual mission became growth over patient health.
Write down purpose: > "customers first, employees second, shareholders last" (Saul Price) or > "employees first, customers second, shareholders last" (Peter Drucker).
Mission-driven means you cannot profit except by achieving the mission. Most are mission hopeful, candy coating extractive engines. Audit: can anyone profit by betraying principles like safety, performance, quality, design, innovation (five horsemen of the apocalypse)? Betrayal unnoticed immediately due to trust.
Use automated testing, AI for auditing to prevent breaches.
Write simple purpose sentence, e.g., Cloudflare: "to make a better internet"; Devoted Health: "treat every customer the way you would your own parents."
Mission drive: identify fiduciary commitments—people you're committed to, metrics/targets for them, accountability system making commitments as important as making money. Like OKRs for fiduciaries.
Study of ex-Google employees' blog posts (50 linked in book): wistful about losing "don't be evil" ethos. Removed from website, then handbook; sued twice for breaching it, settled both.
Ex-Googler (13 years): good intentions lost. Hypothetical: probability Google files quarterly report on time? 100%. Probability accidentally kills someone and covers up? 90-95%. Quarterly reporting has massive apparatus; "don't be evil" was a slogan.
Show apparatus for commitments, no exceptions.
Integrity has two meanings: keeping promises (if you say 2 o'clock, you're there; won't commit uncertainly) and structural integrity (corroded bolts vs. stainless steel).
Organizations need structural integrity to resist internal temptation, board betrayal, external pressure. Costco's governance fortress protects from pressure, e.g., criticism: > "Costco takes money that rightfully belongs to shareholders and instead invest it in improving the customer experience."
Boards as bulwark against pressure, not amplifier of financial gravity.
Founders: read your corporate charter. HBO's Silicon Valley jokes about founders not knowing theirs, losing control.
Charters say: "to pursue any lawful act or activity." Under shareholder primacy (last 40 years), means maximize shareholder returns. Not natural law; historically, corporations pursued beneficial purpose.
Pre-1980s (Adam Smith, grandparents), obvious corporations for specific public benefit, e.g., railroads/canals required legislative approval for public good.
19th century: richest person couldn't change purpose to maximize shareholder value—crime, corporate death penalty.
Rebels built alternatives like public benefit corporation (PBC)—not B Corp (confusing overlap). Two-page Delaware filing: specify purpose, e.g., "advance human flourishing by creating safe and responsible AI systems" or high-quality products.
Major AI labs (Anthropic) use PBCs. Solves fiduciary breach suits: investors agreed to purpose. Minimum armor against pressure.
If not founder, ask: "Is this a mission-driven company? Is that the legal mission? In the charter?" Forces answer up chain, creates excuse for change. Employees motivate leadership to keep them happy.
Filing a charter committing to the company's purpose has no trade-offs. Companies without it risk betrayal. If an investor objects, ask if they believe they should decide on selling the company instead of the founder. This reveals if they are the right partner. VCs claim to be founder-friendly and believe in the vision—test them.
The main downside of techniques in the book is people wasting time trying to talk you out of them. The book has sections on how to talk to investors and lawyers, with scripted responses to common objections collected from years of experience.
Lawyers may say, "Keep options open; don't prematurely commit to purpose." Respond: "Keep options open for maximum value? What about converting customers into soylent green?" If they say, "You never know what you might need," reply, "Why does nobody trust me? Because you won't take that off the table."
This has no downside. Absolute no-brainer.
Anthropic's success as a public benefit corporation (PBC) counters fears it limits growth. They are the fastest-growing company ever, breaking records, proving capitalism works with purpose.
Anthropic implements book techniques, including long-term benefit trust and two-tiered governance. People worry about raising money, but Anthropic raised funds easily.
Reasons cited for Anthropic winning: lower inference costs, faster product velocity, better focus. Tracing "why": better infrastructure from internal breakthroughs, best talent because people want to work for the "good guys" with a mission to save the world. They protect ethos with integrity, core to breakthrough success stories.
Cat from Claude explained shipping massive features weekly due to mission alignment making decisions easy.
Business feels hard without mission alignment; with it, it's organizational flow state—no meetings needed, no misalignment.
Torchbearers: rare people committed to doing the right thing (e.g., designer refusing slop, engineer prioritizing quality, PM prioritizing correctly). Steve Jobs held skip-level meetings with them.
In traditional companies, torchbearers face daily ROI spreadsheets: "What's the ROI of doing the right thing?" In mission-aligned companies:
- Everyone aligned, no spreadsheets needed.
- Clay Christensen: Easier to do the right thing 100% than 98%—no meetings about unsafe ideas.
Culture bank: trustworthiness as asset. Deposits: sacrifices defending values (e.g., HEB grocery manager let customers take groceries free during ice storm power outage—trained as deposit).
Todd Park rule (from Devoted Health, learned from Howard Schultz of Starbucks): Only make deposits, never intentional withdrawals (accidental ones happen).
Internalized in strong culture: Mary Parker Follett's "invisible leader"—people follow values even without managers, boosting velocity exponentially.
OpenAI: complicated with mega personalities (Elon, Sam). Eric advised Anthropic founders (Dario, Daniela) when they left OpenAI pre-ChatGPT boom. They were first-time founders, not hot; investors were effective altruism true believers, no top VCs.
Advised on horror stories; they made PBC from start, committed in charter to reforms like long-term benefit trust (implemented at Series C). Directors on for-profit board appointed by external AI safety trustees without equity—mission incentive only.
Anthropic refuses dangerous models (costly), turned down $200M Pentagon contract despite government overreach—rare courage enabled by structure (investors can't oust Dario easily). Better than founder control (Dario has no dual-class shares).
Need mission guardian (person/entity) to resist gravity of misalignment.
- OpenAI: nonprofit (now PBC).
- Google/Facebook: founder control (mission guardian = founder; has downsides, founders miserable like Atlas).
At Vatican AI governance panel: Eric with Anthropic, OpenAI, Google, Cohere, Palantir—all non-standard governance to protect humanity from AI. Standard governance = shareholders (whoever borrows most money) control—nuts for AI.
Decision points:
- Guardian: person or entity?
- Founder control: fine temporary bridge, but burdensome long-term.
Encode in structure:
- Single entity (Costco rules baked in; vulnerable to erosion).
- Steward with renewal: employee ownership trust (John Lewis Partnership), cooperatives (Mondragon, 80,000 employees), employee voting trust (Alibaba).
Simplest:
- Nonprofit foundation (Novo Nordisk Foundation, largest charitable foundation owning Novo Nordisk stake).
- Perpetual purpose trust (PPT) (Anthropic long-term benefit trust—no economic stake; Patagonia—purpose protector sues trustees if deviating; checks/balances).
Omnibus term: spiritual holding company (like Berkshire Hathaway for spirit/essence). More stable, invests in R&D/quality, better for shareholders.
Evergreen/Tugboat Foundation: no investors.
Friend ousted by investors despite billions returned—not a celebration party, a wake. Company fine, new CEO fine, but trust destroyed: 15-year promises invalid; activist with 0.5% can oust anytime. Leadership philosophy is anti-trustworthy, collapsing institutions.
Eric sees most founders/startups; convincing them is annoying but essential—grandkids will see as obvious.
Lean Startup was once weirder. Eric writes few books after years of personal testing/experimentation with hundreds.
Eric Ries shares insights from working with hundreds of companies, emphasizing real data on preventing value destruction in the economy.
For founders who haven't raised money yet or only raised on SAFEs:
- Do not waste this moment; everyone rushes to get big, but you have a precious opportunity.
- Founders post-Series A or pre-IPO face more pain, needing to get investors on board.
- The best time to plant a tree was 20 years ago; the next best time is today.
1. Become a Public Benefit Corporation (PBC)
- File to become a PBC—super easy and low cost.
- Write a mission statement you feel good about.
- Test it: Spend an hour with your co-founder on adversarial prompting—brainstorm ways to make money while violating the statement. If you'd be sad in that scenario, rewrite it.
- Avoid being rich and miserable; write it down.
2. Implement the Director's Oath
- Inspired by the Anthropic-Figma fight and board responsibilities in tricky situations like AI.
- Companies ask if they can't have an AI person on the board (dangerous), but not having one is worse.
- Directors make more consequential decisions than nurses, yet nurses have a higher standard (Hippocratic oath: first do no harm).
- Write a director's oath into your corporate charter as a precondition for board members.
- Do not wait for standardization; implement now.
3. Secure Founders Preferred Shares and Mission-Protected Provisions
- If you don't have founders preferred shares, ask an LLM: "Eric said founders preferred shares could potentially make me like personally an extra billion dollars someday. Can you explain why?"
- Use them for founder control, extra votes, board control.
- Talk to a lawyer about mission-protected provisions.
- If lawyers charge by the hour, use Virgil law firm (AI-assisted back office acceleration; doesn't charge hourly).
After implementing:
- Imagine pitching to an investor: "It sounds like you're a greedy SOB grabbing all this power. Why do you want to be emperor for life?"
- Avoid investor-controlled or founder-controlled companies; create mission-controlled companies where the mission has sovereignty.
- If feeling greedy, implement something like Anthropic's LTBT: Pledge 10% equity to a nonprofit foundation and 1% of future revenue; foundation gets a board seat.
- Easy in early stages—no need to boot up nonprofit now; write into charter, fire and forget, retain the right to do it later.
These are bare minimum, easiest things—total piece of cake.
Much of this is similar to AGI alignment: aligning AI and aligning business is not a coincidence.
Simple Version: Who aligns the aligners? Number one unsolved problem in AI—not technical alignment (great progress), but human alignment.
- Conway's Law: Organizational imprint of humans shows in software architecture; org chart visible in architecture diagram because human values flow from parent to child.
- If you can't agree on human values to align to, you're cooked.
Deeper Problem: Emergent Intelligence
- Organizations are the oldest form of artificial intelligence—superorganisms, alive like emergent models (same principle as transformer architecture).
- Piano Movers Puzzle Demonstration (link video; in book): Ants solve puzzle humans struggle with.
- Two walls with slits, big I-beam object (like Friends couch meme: "pivot pivot").
- Human tries logically, rearranges, learns.
- One ant can't solve; 1,000 ants can—colony tries, pauses, reorients like intelligence.
- More ants = faster solution; more humans = worse unless carefully aligned.
- Key lesson for organizational design: We birth these emergent entities; if not tended properly, they develop unwanted characteristics deep in their DNA.
- No amount of founder mode fixes it.
Beware birthing unaligned organizations. (Joke: Hire more ants.)
Frederick Winslow Taylor (1911, Principles of Scientific Management): Taylorism was a massive fad—debated at Supreme Court, front-page news, movies made about him.
Mary Parker Follett (contemporary, 1920s): Work so ahead of time, reads like 2026.
- Focus on power with, not power over.
- Superior and subordinate together obey the law of the situation—work to figure out what it requires, not just command.
- Hallmark of a leader: Create more leaders.
- Erased from history (rediscovered 1990s); Peter Drucker called her "the prophet of management".
Invisible Leader Concept:
- Mr. Roundtree (owner, Roundtree Chocolate Factory) is not the leader.
- He excels by instilling sense of common purpose—that is the invisible leader.
- Most consequential decisions happen when no manager is present (e.g., product tradeoffs: rounded vs. straight corners, skeuomorphism, double-check button or erase hard drive).
- Thousands of tiny decisions; only invisible leader is present.
- Without cultivating common purpose, no control—promises are worthless.
Homework: Read Mary Parker Follett.
- Incorruptible: Why Good Companies Go Bad and How Great Companies Stay Great.
- Available anywhere books are sold (hardcover, audiobook, ebook).
- Website: incorruptible.co—join mailing list for bonus content: implementation guides, advanced guides, readers guides, secret cut chapter.
- List of 100+ local independent bookstores carrying it; order online.
- Buy copies to give away; call local bookstore: "I have this book coming out. I'd like a bunch of copies to give away on launch day."
Thanks and sign-off.
Keep Lenny's Podcast in your library
Save the videos and channels worth coming back to, and find them again in one place.
