What this covers

This is a conversation between Keith Rabois and an interviewer. It moves between how Rabois thinks about company-building as an operator and how he thinks about it as an investor. He treats strategy as an equation of weighted variables that a good operator understands well enough to adjust. His examples come from Apple, early PayPal, Opendoor, and Guardant Health, whose liquid biopsy test is replacing invasive tissue biopsies.

The middle of the conversation takes on startup orthodoxy. Rabois calls the lean startup idea harmful, arguing that the most defensible ideas often need capital before there is evidence of product-market fit. He says experiments should be judged by whether a success would deliver a 10x outcome. He also claims he has never found a high-growth company that reached escape velocity and was then beaten by a large competitor in thirty years. On vertical integration, he says that being a component in someone else's stack means long sales cycles and capturing only 10% to 30% of the value. He cites Apple's critics going quiet once it outgrew the open-platform companies. On secrets, he uses Opendoor's bet that homes behave more like commodities than intuition suggests, and says secrets decay as competitors and former employees reverse-engineer them.

The remainder is more personal and more contested. He distinguishes the roughly 70% conviction an executive needs from the 10% to 50% he needs as a seed investor, and he says seed investors win by being directionally right rather than by getting the price right. He questions whether venture operating partners are sustainable. He also dismisses the fake news panic, comparing social platforms to the printing press, and recommends Kelly McGonigal's The Upside of Stress.

Sharpest takeaway

Rabois argues that durable startup success comes from building strategic advantage that compounds over time, through vertical integration, anomalies that signal paradigm shifts, and secrets, rather than from incremental lean experimentation.

  • Owning the full value chain captures more value than being a component in someone else's stack
  • Only 10x paradigm-shift ideas, not 10% incremental improvements, justify building a company from scratch
  • Companies should be easier to run over time because accumulating advantage compounds, and their founding secrets decay as competitors reverse-engineer them

The argument · threads12 threads · 70 claims
0.85

Durable 10x growth comes from anomalies and reinventing the company's secret, not incremental fixes or copyable improvements.

10 pointscentrality 5/5
  • Starting a company from scratch to displace incumbents is an irrational, heroic act, so unless a company has something anomalous in its team, technology, or metrics, it has no chance of becoming one of the top 100 companies of all time.

    “unless you can see something that's special, and very unusual, there's no chance that this is one of the top 100 companies of all time.”

  • Incremental improvements such as better frameworks, UI, or word choice can optimize a startup but cannot produce 10x growth; only paradigm shifts found by asking a series of why questions about anomalous data create 10x growth.

    “They're basically an opportunity to find a paradigm shift and it's the paradigm shifts that allow you to get 10x growth.”

  • Anomalies in people, such as early-career individuals with raw tenacity or resourcefulness, are very hard to identify and require art, while technological anomalies require sophisticated technical understanding, making both equally difficult to spot.

    “The anomalies on people are still really challenging to be able to find those incredible people early in their career”

  • All fundamentally interesting companies are predicated on at least one big secret, and as they succeed, competitors and former employees reverse-engineer that secret, so successful startups must reinvent their secrets every few years to sustain growth.

    “So that you have to reinvent yourself every X years as a successful startup”

  • The criterion for a successful experiment is not whether it is done but whether, if it worked, it would produce a 10x upside scenario, so experiments should be judged by their potential payoff.

    “In my view, the experiment has to have a 10x upside.”

  • Amazon does not offer employees free perks such as free lunches, and startups should be more like Amazon than like failed competitors that spend on perks.

    “I think people don't emulate Amazon enough either. Amazon, for example, doesn't offer employees free perks, they don't serve free lunch there.”

  • The next decade will be driven mostly by new products built on breakthrough hard-science capabilities such as robotics, bioinformatics, lifespan extension, and autonomous driving, rather than by new platforms that others build applications on.

    “I think the next decade is mostly about that, versus a new platform that people build applications on top of.”

  • Since 2010, design and technology have increasingly manipulated the real world, with the smartphone functioning as a remote control for the physical world, so investors want technologies used in the real world rather than escapist digital ones.

    “the iPhone really was the remote control for the real world.”

  • A company needs breakthrough ideas that add orders of magnitude rather than 10% increments, and once a company has scale and accumulating advantage it will eventually run out of 10% ideas, so leaders must ruthlessly say no to incremental ideas to find 10x ones.

    “I want zeros on the dashboard. I don't want 10%.”

  • Even when a business needs magic in the beginning, an anomalous spark should persist, and the strategic leverage is a conceptual vision of where advantage may come from that is more theoretical than empirical early on.

    “I do believe that the strategic advantage comes later. I think it's usually a conceptual answer in the beginning”

0.56

Investors must judge anomalies, mispricing, risk, and capital before evidence of product-market fit.

5 pointscentrality 4/5
  • For investors, anomalies are almost always positive, and empirical anomalies are easier to spot than for operators because founders share their decks and metrics, so an investor can see that a company has, for example, unusually high cohort retention two years out.

    “It's easier to know you found an anomaly, at least an empirical one, as an investor because people give you their decks with all their metrics.”

  • Executives and investors have different risk profiles: investors can accept a chance of zero for a 10x upside, while within a company the risk calculus shifts with accumulated value and the power law operates differently.

    “It's perfectly fine to have a chance of a zero with a 10x upside. In a company that may or may not be the strategy.”

  • The lean startup idea was stupid and harmful to Silicon Valley, because capital often needs to come before evidence of product-market fit for the most differentiated, defensible ideas, though lean approaches are prudent for teams that cannot raise money.

    “I think the lean startup was a stupid idea and followed by stupid people. I think it was poison for Silicon Valley.”

  • Series B and later investments require being right about mispricing and the existence of an opportunity, since price matters more at late stages where valuations are stickier.

    “As you get into series Bs and later you absolutely need to be right about the mispricing and that there's an opportunity there.”

  • Price matters to investors mainly through capital exposure, meaning how much of the portfolio is at risk across stages, industries, and risk-reward equations, rather than through the price of any single company.

    “How much of our portfolio are we risking in various stages, industries, risk-reward equations.”

0.55

Durable advantage comes from accumulating leverage and differentiated services competitors cannot easily copy.

9 pointscentrality 3/5
  • Strategic leverage comes from a business becoming incrementally easier to run over the years rather than requiring the same energy, skill, and talent, because the ability to scale talent at a constant level degrades as companies grow.

    “So you want a business that's actually easier to run over the years than harder to run.”

  • Founders, executives, and investors should be specific about which strategy they follow and where strategic leverage comes from, and the book 7 Powers isolates the only seven ways to create strategic leverage.

    “7 Powers, which really isolates the only seven ways to create strategic leverage.”

  • Early on, strategic leverage is usually a conceptual, theoretical answer about where advantage might come from, and only later does it become empirical as the business shows why it gets easier over time.

    “I do believe that the strategic advantage comes later. I think it's usually a conceptual answer in the beginning”

  • Accumulating advantage is the broader concept, and network effects are a narrow species of it; the goal is to identify why a business gets better every year and why the next year will be easier.

    “One of the reasons why we use the term accumulating advantage versus like network effect is a network effect is a narrow species of an accumulating advantage.”

  • Khosla Ventures' recruiting team with a portfolio of roughly 350 companies can matchmake candidates to 5 or 10 relevant opportunities quickly, which benefits companies, candidates, and the firm simultaneously.

    “And we have a portfolio of roughly 350 companies. So we can match make very quickly for you 5 or 10 opportunities that fit what you're looking for.”

  • To answer a value-proposition question when lacking an answer, ask the three or four people who like you most what they like about you, write it down, and look for common words, since only positive attributes matter for this exercise.

    “Go find the three or four people that like you the most, and ask them to describe what they like about you and just write it down.”

  • Selling a product directly to customers lets a founder control their own fate, while indirect channels leave them waiting on sales that depend on other parties' priorities, though the direct strategy is more capital intensive and riskier.

    “Where you sell a product directly to a customer, you control your own fate.”

  • Companies fail to emulate Apple because its design obsession is easy to state at the jargon level but hard to translate into practical decisions, since Apple makes decisions through creative selection rather than empirically driven processes, and its vertical integration and closed-versus-open system choices are missed.

    “Well, obsession around design. And that's easy to say at jargon level. I think translating into practical decision making is very different”

  • Value-added services like those offered by venture firms are only a real advantage when competitors do not copy them, since if every firm offers identical support services it just raises costs for all firms.

    “it only really is an advantage when other people copy what you're doing, it doesn't work as well.”

0.50

Asymmetric capabilities should be owned end-to-end rather than sold as components inside another stack.

3 pointscentrality 4/5
  • If an entity has an asymmetric ability to perform a function, it should own the entire stack and deliver the end product directly, because being a component in someone else's stack leads to long sales cycles, adoption risk, and capturing only 10% to 30% of the value created.

    “If you're selling to someone else, maybe you capture 10% to 30% of the value create, and that's if you're pretty good at sales and pricing.”

  • Selling as a component inside another company's stack carries adoption risk and long sales cycles that are outside the seller's control, so founders lose control of their fate.

    “You have the adoption risk issue, that includes long sales cycles, et cetera.”

  • Apple has always been vertically integrated, and critics of its closed platform fell silent once Apple became more valuable than the open-platform companies, showing that controlling each component produces the ideal user experience and credit for differentiated advantages.

    “the quintessential example is Apple, which has always been vertically integrated.”

0.48

Concrete examples show secrets, healthcare gaps, and poorly served industries enable beating incumbents.

7 pointscentrality 3/5
  • Opendoor's core secret was that homes are much more like a commodity than people's intuition suggests; pricing a house in the United States is not very difficult given the right data and skills.

    “homes are much more on the scale of commodity than they are works of art.”

  • Early PayPal created trust before it had a brand by offering a $100,000 guarantee insured by Travelers, displaying the Travelers logo on the homepage, and creating a novel way to qualify accounts for FDIC insurance up to $100,000.

    “we had a $100,000 guarantee that if the money didn't get from point A to point B as is supposed to, Traveler's Insurance”

  • Guardant Health uses a liquid biopsy blood test to detect cancer type instead of an invasive, expensive, and risky tissue biopsy, and it is increasingly replacing biopsies while becoming cheaper, less risky, and sometimes more accurate.

    “Guardant is now a successful replacement for a lot of biopsies, it does it more cheaply all the time, less risk all of the time.”

  • Healthcare is 22% of the US economy and delivers poor experiences that people procrastinate about and despise, so there is substantial room for technology to reduce cost, increase convenience, and improve outcomes simultaneously.

    “healthcare is 22% of the US economy. And it's not particularly a high-quality experience for anybody.”

  • Technology is the only lever in history that can improve experience, reduce cost, and improve quality at the same time, which is why both financial services and healthcare are attractive investment areas.

    “Whereas through no other lever in the history of the world can you do all three at the same time.”

  • In healthcare and financial services, trust is an accumulating advantage because a company that has operated for years becomes incrementally easier to believe, as customers have references from neighbors who transacted with it.

    “So for example, trust can be one of those.”

  • Rabois's investment formula identifies large industries with poor experience and procrastination as targets, a pattern that drives investments across healthcare and financial services.

    “That's the formula that drives our investment decisions.”

0.43

Venture investing demands contrarian conviction and accepting that most bets will fail.

5 pointscentrality 3/5
  • Investing resembles baseball because each investor is individually accountable and works alone on decisions, whereas operating resembles football because everyone must execute the same play in sync.

    “it's a lot more like playing baseball, than football, which is every batter in baseball has their own accountability and metrics.”

  • A secret is a belief system about the world that the rest of the world does not appreciate and that can be proven or disproven over time, though some secrets can be independent of evidence until time and effort reveal whether they were true.

    “A secret is a belief system about the world that can be proven or not, that the rest of the world doesn't appreciate.”

  • An investor gets paid to learn new things, meet new people, and try new things, whereas executives must become parochial and go very deep for years or decades, so intellectual curiosity can make investing a better career fit.

    “If you're intellectually curious, for example, being an investor is a much better job.”

  • Early-stage valuations have been roughly flat since 2015, around 25% to 33% below 2013-2015 levels, and at seed and Series A, returns depend on being directionally right about the company rather than on price.

    “You're mostly making money at the very early stages and I mean, specifically seed and series A by being right, directionally right about the company, not the price.”

  • Most things do not work and fail to meet ambitious hopes, and like baseball hitters who succeed roughly 30% of the time, founders and investors must accept that most of their efforts will not succeed.

    “great baseball players roughly get hit set 30% of the time. So you can't succeed in baseball without knowing that roughly 70% of the time you're not going to be happy.”

0.43

Evaluating founders and firms hinges on talent recruitment and differentiated, non-conventional approaches.

5 pointscentrality 3/5
  • Assessing whether a founder can recruit the requisite talent is probably the most primary objective in evaluating a founder, and this ability should show up in prior competitive environments where former colleagues would eagerly join.

    “I think that is maybe a primary, maybe the most primary objective in assessing a founder is can they recruit the requisite talent?”

  • Venture support services such as operating partners are justified only when they are high-leverage for the firm, add value for founders, and benefit potential employees, because services that merely raise costs for all firms are not sustainable.

    “That's my criteria, personally for things I subscribe to is, is this high leverage for KV? Does it work for the founders and add value? And does it work for the potential employees?”

  • Traditional venture capital has seen surprisingly little innovation in the last five to ten years, with only vertical specialists like Khosla Ventures and Lux Capital showing differentiated approaches.

    “I don't know that anybody in the last like 5, 10 years has really innovated that much.”

  • Y Combinator has differentiated itself over 13 years with a different approach, mentality, and economics, and it appears to have a rare network effect in venture capital.

    “I think Y Combinator, for example, has done a phenomenal job over the last 13 years or so, differentiating.”

  • Some people would be very successful founders but chase bad ideas, and redirecting them could let them solve problems as a second to twentieth employee instead of as a founder.

    “And maybe they just have to be the 2nd to 20th employee instead of founder.”

0.32

Career opportunity flows from associating with exceptional people and great leaders who recruit or recognize you.

6 pointscentrality 2/5
  • Attaching yourself to an amazing person is the best way to learn a craft like Silicon Valley's, because talented people who start new things call their former collaborators first, so proximity creates opportunities.

    “if you find somebody who's amazing, they will want you to join them, they're going to do interesting things.”

  • Simply working harder than others, including attention to small details such as remembering customers' names and waking early, is a contrarian advantage that lets someone from an ordinary background open opportunities for decades.

    “If you want to get ahead in life, just put in more effort than other people.”

  • The best investment is in people who show unusual insight or unusual tenacity before others notice, and a person can propel themselves for 10 to 30 years on just a couple of key differentiating dimensions.

    “I'm always looking for people who have some differentiation before other people notice that about them.”

  • Over roughly 20 years, Rabois worked for five notable bosses, Peter Thiel, Reid Hoffman, Max Levchin, Jack Dorsey, and Vinod Khosla, and developed the ability to work with very ambitious, opinionated founders.

    “five I guess, bosses Peter Thiel, Reid Hoffman, Max Levchin, and Jack Dorsey, and Vinod Khosla.”

  • Talented people who join great leaders will call those leaders when they start something new, so a young professional who attaches themselves to an impressive person can learn through osmosis and be on that person's shortlist for future recruitment without needing to do anything except wait.

    “if they go start something or they go join something, what's the first thing they're going to do? Is they're going to call you.”

  • The kindest acts are when people remember and articulate the impact a leader had on their career or life, often unexpectedly and privately, because leaders with large teams and boards have cascading impact they cannot fully track.

    “The things that stick with me are people who remember and articulate in a really sync way, the impact I have in their career or life.”

0.29

Success is learned by copying proven companies, with small focused teams outexecuting incumbents.

5 pointscentrality 2/5
  • Startups are like movies: start with an inspired narrative, cast the right team, raise money to produce, then market with a trailer value proposition to sell tickets, iterating on marketing, which is a better metaphor than bottom-up approaches.

    “I tend to think of startups are more like a movie and the idea behind the movie is you have this inspired narrative”

  • Apple's core advantage is its obsession with design, which translates into decision-making that is not empirically driven, a dynamic described in the book Creative Selection.

    “Well, obsession around design. And that's easy to say at jargon level.”

  • Opendoor decided real estate was broken, raised $10 million to fix it, and then cast the right people to validate the theory, which is a classic fat startup approach rather than a lean one.

    “And so we raised $10 million to fix it.”

  • Success is learned by decomposing successful companies and emulating them, as one learns sports by copying great players, and people should copy successful companies more often than they focus on avoiding failure.

    “I think more people should copy, replicate, aim to reproduce what the successful companies do.”

  • Rabois has never found a satisfactory example of a high-growth startup that reached escape velocity and was then beaten by a large competitor in the last 30 years, so small focused teams with ownership mentality can outexecute large incumbents.

    “is there any high growth startup and I can quantify what that means, but any startup that hit escape velocity that a large competitor ever basically beat?”

0.26

Personal capability grows through stress, serious reading, tenacity, and working with ambitious founders.

4 pointscentrality 2/5
  • Kelly McGonigal's book The Upside of Stress argues that the more stress people experience and the more they adapt their mindset toward it, the healthier, wealthier, and happier they are, based on Boston-area findings that the most stressed people lived longest.

    “she wound up finding evidence in the Boston world that people who are the most stressed were living the longest”

  • Reading serious printed books exercises the brain like a muscle, developing capacities that can be leveraged indirectly in unforeseen ways later, and high-quality reading gives the raw ingredients for seizing opportunities.

    “your brain is basically a muscle you want to exercise your muscle and that develops capacities, capabilities that you can leverage later, indirectly”

  • Recognizing a valuable problem can be separated from the ability to solve it, though successful founders have typically bundled both, and founders who did not originate the solution may lack the tenacity to persist.

    “So recognition of a problem that's worth solving, and that there's a theoretical approach to solve it can possibly be divorced from the ability to solve it.”

  • Rabois built a rare skill by fusing business thinking with design, working with first-rate designers in a way they valued while focusing his own drive on making money and turning growth into cash flow.

    “I became more able and known to be able to fuse business thinking with design, which is again a rare, relatively rare skill”

0.26

Operators and investors require different conviction levels because their decision roles differ fundamentally.

4 pointscentrality 2/5
  • Strategy must be treated as an equation of weighted variables that a strategic operator understands well enough to manipulate, since tweaking one variable has predictable effects elsewhere and can open alternative approaches.

    “every business when it works is like an equation. It's X times Y times Z with some weighting.”

  • Most startup decisions must be made with about 70% conviction, and an executive thrives by getting proficient at making smart decisions at that level, whereas as a seed and Series A investor he needs only 10% to 50% conviction because he does not make the market or hiring decisions.

    “Most decisions in a company, in the startup particularly need to be made with 70% conviction, or 70% confidence.”

  • Investing differs from operating because an investor almost never makes decisions on the company's behalf and instead acts as a psychologist giving founders feedback and trade-offs, while an executive is responsible for shipping output and making decisions with incomplete information.

    “All I am is playing psychologist, which is basically giving the founder feedback on potential avenues”

  • At seed and Series A, investors make money by being directionally right about the company rather than by getting the price right, while Series B and later require being right about mispricing, and capital at risk and portfolio exposure matter more than price itself.

    “As you get into series Bs and later you absolutely need to be right about the mispricing and that there's an opportunity there.”

0.12

Information access leaves society better informed despite fake news and foreign manipulation concerns.

4 pointscentrality 1/5
  • Foreign interference and manipulation of information is not surprising because every government in history has done this, and the United States and other countries engage in it, so it is not a new phenomenon.

    “Every government in the history of the world has been doing stuff like this for centuries.”

  • The fake news debate is ridiculous because the average American is better informed than at any point in history, and the relevant question is whether the net effect of information access makes people smarter, not whether isolated false content exists.

    “It's like the net effect is what matters not any isolated variable.”

  • Social platforms are more democratic versions of the printing press, which similarly enabled revolutions and wars by letting anyone proselytize and undermine existing powers.

    “That's what these platforms are, is they're like more democratic versions of printing presses.”

  • Unlike historical periods when books were a luxury and public libraries existed because people could not afford books, today Americans carry a library in their pocket and can access nearly any information for free.

    “People are walking around with a library in their pocket, literally.”