Most important take away

Most founders lose control of their companies because they sign standard “best practice” governance documents that legally obligate the board to maximize shareholder value, no matter the cost to mission, customers, or employees — only 20% of venture-backed founders are still CEO three years after IPO. The single highest-leverage protective move any early-stage founder can make this week is to file as a Delaware Public Benefit Corporation (a two-page filing) and write a real, specific mission into the charter. Pair internal “ethos” (a clear purpose plus mission drive, so no one can profit by betraying the principles) with structural “integrity” (governance like PBCs, perpetual purpose trusts, or non-profit foundations such as Anthropic’s Long-Term Benefit Trust) before you raise priced rounds — because it is always too early until it is too late.

Summary

Eric Ries argues that the Lean Startup taught founders how to build successful companies, and his new book Incorruptible (out May 26) teaches them how to protect what they built from “the force no one controls but everyone obeys” — financial gravity that drags great companies into mediocrity, extraction, and founder ouster. He calls this corruption, not in the legal-bribery sense but in the structural sense of bridges with corroded bolts: the failure is predictable, has known mechanisms, and has known countermeasures.

Key themes:

• Standard governance is a trap. Almost every YC/VC-template charter contains a fiduciary duty that would legally compel a board to sell the company to even Philip Morris for $1 more per share. The Vectura case (UK inhaler company actually bought by Philip Morris and destroyed within three years) is the literal embodiment.
• It is always too early until it is too late. Lawyers, VCs, bankers, and CFOs will sequentially tell founders to defer mission-protective provisions through every stage — and then, at IPO, tell them it is too late. Founders only have leverage before they need it.
• Ethos + Integrity is the formula. Ethos = internal alignment to purpose; Integrity = structural ability to keep the promise even under pressure.
• Harder is easier. Committing to quality, safety, design, ethics costs short-term margin but produces trust, which is the single most undervalued asset in business (Cloudflare giving away SSL is the canonical example).
• Mission drive, not mission hope. A company is only truly mission-driven if it cannot profit by betraying its mission. Audit every OKR, bonus, and incentive: can anyone profit by sacrificing safety, quality, performance, design, or innovation? Don’t Be Evil failed at Google because there was no apparatus for it the way there is for the quarterly report.
• The culture bank (Todd Park / Howard Schultz rule). Only make deposits (sacrifices for values), never make intentional withdrawals (greedy shortcuts).
• Spiritual holding companies. Various structures — Public Benefit Corporation (PBC), perpetual purpose trusts (Patagonia), industrial foundations (Novo Nordisk, Zeiss), employee ownership trusts (John Lewis, Alibaba), Anthropic’s Long-Term Benefit Trust — all serve as a “mission guardian.” Companies with these structures are 6x more likely to live to year 50 and have superior return on invested capital.
• Mission-controlled, not founder-controlled or investor-controlled. Founder control burns out founders (Atlas who can’t shrug). Investor control destroys missions. The third path encodes sovereignty in the mission itself.

Actionable career and product advice:

1. Founders pre-priced-round: File as a Delaware Public Benefit Corporation (two-page filing). Write a specific mission into the charter. Adversarially test it: “can we make money while violating this?”
2. Founders at any stage: Read your own corporate charter — most never have. Understand and secure Founders Preferred Shares. Talk to your lawyer about mission-protective provisions (or use Virgil, the law firm Ries co-founded).
3. Adopt a Director’s Oath as a precondition for board service — like the Hippocratic oath, but for directors.
4. Consider an Anthropic-style Long-Term Benefit Trust clause in your charter early, even if you don’t activate it until later: “10% of equity pledged to a non-profit foundation, 1% of future revenue, foundation gets a board seat” — fire and forget, boot it up later.
5. Employees and job candidates: Just ask in interviews, “Is this a mission-driven company? Is the mission in the legal charter?” — this forces the question up the chain and creates internal pressure for change.
6. Operators: Look for “torchbearers” — the rare designers, engineers, PMs who simply won’t ship slop. Steve Jobs did skip-level meetings to find them.
7. Cultivate the “invisible leader” (Mary Parker Follett, 1920): the common purpose that decides what the team does when no manager is in the room.

The AI throughline: every frontier AI company on the Vatican panel (Anthropic, OpenAI, Google, Cohere, Palantir) has non-standard governance because standard governance is too dangerous for a technology this powerful. Anthropic’s growth proves a PBC + LTBT structure does not handicap fundraising or velocity — it produces it, because mission alignment is the organizational equivalent of flow state and is why top talent goes there.

Chapter Summaries

Lean Startup vindicated by AI labs. Lenny opens by noting modern AI companies (Claude Code, ChatGPT, etc.) are textbook Lean Startup — ship a research preview, treat product bets as hypotheses, iterate. Ries notes labs themselves did not predict which of their experiments would explode.

What is “corruption”? Ries reframes corruption not as bribery but as the predictable structural decay of successful organizations — the restaurant taken over by PE that you can taste, the natural-foods brand whose namesake founder gets ousted, Vital Farms ownership concerns. “Their very success became a liability.”

The data on founder ouster. Only 20% of venture-backed founders are still CEO three years after IPO (Harvard Law). A “hot company” warned by Ries pre-IPO ignored him; the founder was ousted five months post-IPO when a competitor got acquired and the stock dropped.

It is always too early until it is too late. A sequential story of how lawyers, VCs, growth investors, and bankers each tell the founder to defer mission-protective provisions until the next stage — until the CFO at IPO road show says “sorry, you should have said something earlier.”

The blueprint: Ethos + Integrity, and the Novo Nordisk story. Marie and August Krogh in 1920 Denmark structured the predecessor to Novo Nordisk as a for-profit owned by a non-profit foundation (an “industrial foundation”). The structure has protected scientific integrity for 100+ years; the trustees once intervened and created $500B+ in shareholder value. Zeiss had this structure in 1885. Companies with this structure are 6x more likely to live to year 50.

Why this is not optional — the Vectura / Philip Morris case. Ries asks audiences to imagine Philip Morris offering $1 over current share price to sell cigarettes to children — most refuse, then learn their own charter legally obligates them to accept. UK inhaler company Vectura actually had this happen: the board said “our hands are tied,” sold for 165p/share; Philip Morris wrote it down $900M within three years and disposed of it.

Harder is easier (the Cloudflare SSL story). Cloudflare’s number-one revenue driver was paid SSL encryption. A junior engineer asked Matthew Prince: “wouldn’t a better internet be an encrypted internet?” Prince said “let’s figure it out” — they re-engineered to give SSL free, took a short-term conversion hit, top-of-funnel went up 10x, today $70B company. Contrast: Groupon’s “one email a day” was ground down by ROI arguments to eight emails a day, destroying the company.

Ethos: write down your purpose (“who would you rather die than betray?”). Avoid consulting speak; use “fiduciaries” and “purpose” instead of “stakeholders” and “mission statement.” The mission statement is not the mission — mission is emergent. Examples: Cloudflare’s “make a better internet,” Devoted Health’s “treat every customer the way you would your own parents.”

Mission drive vs. mission hope. Audit incentives: can anyone profit by betraying safety/quality/performance/design/innovation? Google’s quarterly report ships with 100% reliability because there’s an apparatus; Don’t Be Evil was just a slogan. AI and automated tests make this easier today.

Integrity: structural ability to keep promises. Most founders have never read their own charter (Silicon Valley made a joke of this). “Any lawful act or activity” today means “maximize shareholder returns” because of shareholder primacy doctrine (only ~40 years old). For most of history, companies needed a beneficial purpose; changing it was a crime that could trigger corporate death penalty.

The easiest fix: become a Public Benefit Corporation. Two-page filing in Delaware. Zero meaningful downsides. All major AI labs are PBCs. The only “downside” is investors trying to talk you out of it — and if they do, ask yourself if they’re really the right partner. The book has detailed scripts for talking to lawyers and investors.

The Anthropic story (Ries played a bit part). When Dario Amodei left OpenAI to start Anthropic, an investor sent him to Ries. They wrote PBC status into the charter from day one and reserved the right to enact the Long-Term Benefit Trust, which they implemented at Series C. The LTBT appoints directors to the for-profit board who are AI safety experts without equity — so when Anthropic refuses to ship a dangerous model or turns down a $200M Pentagon contract, the structure makes that possible. Anthropic’s competitive advantages (lower inference costs, faster velocity, focus, top talent) all trace back to ethos + integrity attracting people who want to work for “the good guys.”

Mission guardians and “spiritual holding companies.” Options for the mission guardian role: founder control (temporary, burns founders out), single-entity rules (Costco — fortress but can’t regenerate), non-profit foundations (Novo Nordisk), perpetual purpose trusts (Patagonia — includes a “purpose protector” who can sue trustees), employee ownership trusts (John Lewis), employee voting trusts (Alibaba), cooperatives (Mondragon, 80k employees), or Anthropic-style LTBTs. Ries uses “spiritual holding company” as the umbrella term.

Why this isn’t a drag: the wake party. Ries describes a “celebration” for a founder friend ousted by investors — 1,000 people, including laid-off employees, flew in. He had to explain to a wide-eyed young founder that it was actually a wake. Trust collapses in institutions because we teach a leadership philosophy that is anti-trustworthy.

Three actions for early-stage founders this week: (1) File as PBC with a real mission; (2) Adopt a Director’s Oath in the charter; (3) Understand Founders Preferred Shares, talk to a lawyer about mission-protective provisions, and write Anthropic-LTBT-style language into the charter even if you don’t activate it yet.

AGI alignment parallel. “Who aligns the aligners?” Conway’s Law means human org values flow into product architecture. Corporations are themselves emergent intelligences (the ants-solving-the-piano-puzzle demo): more ants = faster solution; more humans = worse unless carefully aligned. No amount of founder mode fixes deep DNA misalignment.

Mary Parker Follett and the invisible leader. Closing nugget: Follett (1920, erased from history then rediscovered in the 1990s) taught “power with, not power over,” that a leader’s hallmark is creating more leaders, and that the real leader of any organization is the common purpose itself — because the most consequential decisions are made when no manager is in the room.