What this covers

Eric Schmidt reflects on management, technology platforms, and institutional design across business, politics, and civic life. The conversation ranges from how Google scaled its early operations to how incentives shape behavior in companies, cities, and democracies. Schmidt argues that what distinguishes successful organizations is not visionary founders alone but the disciplined systematization of hiring, strategy, and product development—paired with careful incentive alignment and the deliberate use of recognition over coercion.

The discussion spans concrete cases—Google's shift to auction-based ad pricing via Project Drano, the optimal number of job interviews, why Bell Labs and Xerox PARC succeeded in funding pure research—alongside broader institutional critiques. Schmidt contends that tech platforms obey network economics laws that amplify both value creation and human vulnerabilities: social media exploits cognitive biases and enables disinformation at scale, while selection bias leads technologists to mistakenly assume their own values—socially liberal, fiscally conservative—are universal. He makes normative claims about American governance (the Senate's rural bias, the need to reform the quarterly cycle to lengthen corporate time horizons), housing policy (the Bay Area's exclusionary stance), and philanthropic accountability (insisting on measurable impact over emotional appeals). Several positions invite pushback: his skepticism that corporate employees will reliably resist internal wrongdoing, his claim that innovation cannot thrive in separate research labs, and his diagnosis of tech's role in social fragmentation.

Sharpest takeaway

Schmidt argues that successful organizations are systematized human systems—where the quality of people, alignment of incentives, and disciplined mechanisms (not lone-genius founders) drive outcomes—and that technology platforms follow economic laws while creating new societal risks around amplification and information control.

  • Innovation, hiring, and strategy can all be systematized in fast-growing companies
  • Incentive alignment and praise drive human performance more than charisma or coercion
  • Platforms follow network economics laws and amplify both value and human weaknesses

The argument · threads10 threads · 39 claims
0.76

Leadership succeeds by organizing people around shared mission and aligned incentives rather than coercion or status symbols.

5 pointscentrality 5/5
  • Charismatic leaders can motivate people to do things they would not otherwise rationally do—such as working through the night—because team membership and belief in a shared mission overrides individual rational calculation.

    “what I realized is that charismatic leaders can get people to do stuff that they would otherwise not rationally do, but because we were a team and because we believed in our mission”

  • Praise is underused as a management tool: since you cannot coerce employees, adding a preamble of thanks, appreciation, or recognition makes people's 'hearts sing' and yields far more work, yet most great leaders fail to use enough of it.

    “if you take a moment, and you add the preamble of "Thank you" or "I appreciate it," or "I recognize it," people's hearts sing, and you get a lot more work out of them.”

  • Systems run on alignment of incentives; if you reward promotion with status symbols like a rug, you create a market for the rug and distract people, so the goal is to align people on the right incentives—changing the world, impact, scale, great products, serving customers—rather than the wrong ones.

    “If you create a situation where, when you get promoted, you get a rug in your office, and if you're not promoted, you don't have a rug, then you've created a market for a rug. So my goal was always to not have a market for the rug.”

  • The best way to get people to do things is to make your idea also their idea; the ideal manager never has to do anything because the people they work with are self-organizing and so love what they do that they solve their own problems—so leadership is fundamentally about organizing how to get people to act.

    “My father taught me that the best way to get people to do stuff is to have it be their idea. So if you can find a way that your idea is also their idea, then you can go back to working on the people whose idea you don't agree with.”

  • Despite media fascination with brilliant founders, corporate success is a team sport: the people who work every day—putting in 16-hour days because they love the founders and the vision—are who actually change the world, so the quality of the people you work with, hire, and put on your board determines much of your outcome.

    “In the media, there's this fascination with founders and how brilliant they are, but corporate success is a team sport.”

0.76

Nearly all tech and business phenomena follow platform economics and network effects that can be systematized and scaled.

5 pointscentrality 5/5
  • You can systematize nearly anything in a fast-growing company—innovation, strategy, hiring, product development—as long as you have a scale model of it, because thinking in terms of scale and mechanisms (a computer scientist's habit) lets you improve each function at successive stages.

    “You can systematize innovation, and you can systematize pretty much anything in a fast growing company. All you have to do is have a scale model of it.”

  • Nearly everything built in tech is essentially a platform, and platforms have unique structures and scales that follow economic laws, which is why understanding network economics has driven business success.

    “Pretty much everything you all build on is essentially a platform. Platforms have unique structures and scales, and they, in fact, follow economics laws.”

  • Google did its own auction-based IPO not to save money (the fees would likely have been similar) but because it was an auction business that ran more auctions than anyone, so a Vickrey-style double-blind auction fit its identity; there is no evidence it materially changed the outcome, and other companies don't do it because the gain hasn't been worth the cost of being different.

    “we were in the auction business, and we did more auctions than anyone else in the world by far. It made no sense to us to do the traditional IPO, which is not really an auction.”

  • Blockchain is already a good platform for cryptocurrencies and trustless private banking transactions, but the most interesting development is execution on top of blockchain such as Ethereum, and if Ethereum can achieve global synchronization of that activity it becomes a genuinely new and powerful platform.

    “if Ethereum can manage to figure out a way to do global synchronization of that activity, that's a pretty powerful platform. That's a really new invention.”

  • Bell Labs and Xerox PARC produced enormous scientific impact because monopoly structures (legal for AT&T, structural for Xerox) gave them enough money to fund pure research that universities couldn't afford, but separate research labs no longer work and must be integrated with product development.

    “Bell Labs, of course, came out of amonopoly agreement around AT&T, and as part of the way the monopoly was set up, Bell Labs was funded to do interesting things.”

0.58

Hiring strategy for fast-growing companies should prioritize learning potential and cultural fit over domain experience and credentials.

4 pointscentrality 4/5
  • In a fast-growing company in a new field, hiring inexperienced people based on school and GPA rather than experience is a reasonable business analysis, because such people can pivot to new challenges where experienced people's experience would not transfer.

    “in a fast growing company, we need people who can pivot around whatever the new challenge is, and we weren't convinced that the people who had experience — since this was a new field — that that experience would allow them to pivot into the new problems, so it was a reasonable business analysis.”

  • 'Glue people'—very nice individuals who sit between two functions, carrying bags and writing memos but who you don't actually need—make organizational systems less efficient, which is why Google reviewed all offers to ensure it was hiring for genuinely needed specialized skills.

    “Glue people is my term for people who are very, very nice, who sit between two functions. You don't need them, but everyone likes them because they carry the person's bag, and they write memos, and so forth, but you don'tactuallyneed them. They make the system less efficient”

  • A typical American company has about 10 percent annual turnover, so a 20,000-person company hires roughly 2,000 people a year—enormous leverage over a few years—yet CEOs are unwilling to do the surgery needed to bring in the talent required to lead the company into its next generation.

    “A typical American company has a turnover of about 10 percent... if you have a 20,000-person company, you're hiring 2,000 people, plus or minus, okay? That's a lot of leverage over a few years in terms of the hiring pool, and yet the CEOs are unwilling to go in and do the surgery necessary”

  • Google interviewed one candidate sixteen times without deciding; Schmidt capped interviews at eight, and subsequent statistical analysis showed the optimal number is four to five interviews.

    “We've since done a statistical analysis, and the answer today is four to five interviews.”

0.58

Social media and Internet structure amplify human cognitive weaknesses and create vectors for manipulation by state and commercial actors.

4 pointscentrality 4/5
  • The sum of how social media works is organized around human weaknesses rather than strengths: the Internet's structure amplifies individual crazy people to harmful levels, exploits cognitive flaws (e.g., you can't erase the memory of a fake video even when told it's fake), and lets special interests and foreign powers blur the line between legitimate speech and marketing/disinformation—a problem AI will worsen by learning to manipulate individual preferences.

    “I'm worried that the sum of the way social media works is organized around some human weaknesses, not human strengths.”

  • One should read multiple sources and carefully check the source of information, because it is now so easy for people to mislead you and appear legitimate when they are not.

    “I'm careful now to look at the source of information because it's so easy for people to mislead you and look like they're legit, but in fact they're not.”

  • A tech company cannot covertly repurpose user information (e.g., for a bundled hedge fund) because doing so requires people's permission, which you won't get if you'd have to disclose it; large-company employees act as 'independent citizens of humanity' who expose value-inconsistent behavior, so you cannot do such things in secret and wouldn't want to.

    “the answer is always the same, which is that you need people's permission to do that, and you can be sure you won't get that permission, if you follow that reasoning.”

  • Tech culture is not mainstream culture: the average tech person is socially liberal and fiscally conservative, which is the smallest quadrant of the US demographic combinations, and selection bias leads tech people to wrongly assume everyone else shares their values, producing a values uniformity that isn't necessarily bad because teams need common belief to work together.

    “The average tech person, as a gross generalization, is socially liberal and fiscally conservative. If you look at the quadrants of combinations, that percentage is the smallest percentage of the demographic of the United States.”

0.58

Corporate governance and societal institutions suffer from misaligned incentives and short-termism that obstruct effective investment and innovation.

4 pointscentrality 4/5
  • The single most needed reform to American corporate governance is longer time frames, because the quarterly cycle makes it difficult for companies to invest, and misalignment between short-term shareholders, hedge funds, and long-term company interests produces sheepish rather than aggressive leadership that won't take warranted risks.

    “Pretty much everybody agrees that thequarterly cyclemakes it very difficult for companies to invest”

  • Google's failure in social networks—Schmidt's biggest CEO miss—happened because his generation was more comfortable with telephones and email and didn't collectively use social media, so they didn't fully understand how to build it, while a slightly younger generation drove the rise of Facebook.

    “because we didn't collectively use it, I suspect we didn't fully understand how to do it.”

  • Most non-business sectors—health care, education, governments, philanthropies—are poorly run compared to businesses, so people should give money and time only to groups that can demonstrate measurable impact by stating the outcome, probability, and path, rather than responding to emotional marketing pitches.

    “most of the stuff that's not business is not very well managed. So as a general rule, broadly speaking, health care, education, governments, philanthropies — they're not well run compared to businesses.”

  • Despite Schmidt's deep skepticism that Google's ads were worth anything—leading him to impose a cash restriction period—the Project Drano ads conversion to a unified auction produced prices three times higher than the previous fixed pricing.

    “The auction produced a price that was three times higher than the previous prices.Veryinteresting.”

0.57

Urban policy, demographic shifts, and sector-specific productivity gaps reflect structural incentive problems and entrenched interests requiring systemic reform.

7 pointscentrality 3/5
  • US politics has a long-standing structural rural bias, especially in the Senate, because the country was founded as rural, and as more of the population moves to cities there is a growing disconnect between where the people are and where leadership comes from; the fix is either citizens moving to rural areas or allotting more power to where the people are.

    “people who have looked at this all see the brilliance of our founding fathers, but also a rural bias in the structure.”

  • City building is extremely rare and it is almost impossible to change a city's planning assumptions—even in bankrupt or troubled cities—because human incentives and entrenched special-interest groups have locked themselves into the status quo, so real innovation requires exempting areas from rules or building entirely new cities (as China and Saudi Arabia are doing).

    “it's almost impossible to change the planning assumptions about a city because of the way in which the human incentives and all of the very special interest groups that have locked themselves into a state.”

  • The Bay Area's historic no-growth/slow-growth stance benefits incumbent landowners (whose undeveloped land rises in value with demand) while hurting immigrants into the area; the solution is allowing skyscrapers or dense low-rise housing in selected areas, because demand will not slow down and the current situation of blocking all proposals is discriminatory and harmful.

    “theBay Area is subject to a historic no-growth, slow-growth view, which benefits the incumbents and hurts immigrants into the area. I am a landowner, and if I keep my land undeveloped, then the price goes up because there's greater demand.”

  • Because the postwar baby-boom demographic that drove high mid-century growth has inverted, a shrinking working population must fund a growing retired population, which is precisely why AI systems will make workers more productive rather than cause them to lose their jobs.

    “It's why, for example, AI systems will make youmoreproductive and not cause you to lose your jobs. That's a way to understand that.”

  • Cuba is a very poor country needing help, and after 50 years of US blockade the economy will do well only if America works with Cuba; opening it up is the right move, but current US policy is going the wrong way.

    “I think the economy will do wellifAmerica works with Cuba. At the moment, we're going the wrong way.”

  • Sensors of every conceivable kind will explode over the next decade—enabled by smartphones, cheap cameras, and incredible connectivity—and the businesses built on them can solve world problems, including making philanthropy accountable by tracking whether donated goods reach the right place, as RFID tagging already reduced misplaced and stolen activity.

    “What's the thing that's going to explode in the next 10 years in a good way, explosion? Sensors. Sensors, every conceivable kind of sensors.”

  • Middle-class wage stagnation reflects missed opportunities to apply technology to non-tech industries; a small number of software people applied to each industry—examining operations and redesigning business processes—would make a huge difference in productivity, which is what Google's set of cross-industry businesses aims to do.

    “I'mconvincedthat a small number of software people applied to each of these industries — looking at how they operate, redesigning the way their business processes work — will make a huge difference.”

0.45

The Internet will likely bifurcate into Chinese and non-Chinese spheres with divergent governance models affecting freedom and control.

4 pointscentrality 3/5
  • Rather than splintering into many pieces, the Internet is most likely to bifurcate into a Chinese-led Internet and a non-Chinese (American-led) Internet, with China's phenomenal product and service leadership exported via the Belt and Road Initiative to ~60 countries, carrying with it a governance regime of censorship and control with some loss of freedom.

    “I think the most likely scenario now is not splintering but rather a bifurcation into a Chinese-led Internet and a non-Chinese-led Internet, presumably led by America.”

  • Chinese tech company managers feel like Google did 10 years ago—facing crazy growth and problem after problem—but with the added task of 'managing up' to regulators because the government can say yes or no to many of their initiatives.

    “they also manageupbecause the government can say yes or no to many of the things that they were doing.”

  • The technology is finally good enough that intelligent verbal assistants will work at scale, with much of the computing interface becoming verbal rather than text—a shift enabled by image and voice recognition gains that Schmidt thought would be unachievable in his lifetime.

    “Many people believe that we can get to the point where you can haveincrediblyintelligent assistants that are on these devices, and that much of the interface will be verbal, not text.”

  • The effort to scan all the world's books stalled because, while copyright law had to be followed, in many cases a rights holder for an older book (e.g., 50 years old) could not be found, and Google could not reach agreement with the book publisher industry through the legal process.

    “we had a lot of issues with the rights holders because in many cases you'd have a 50-year-old book, but we couldn't find the right holder, and we had to follow the copyright law.”

0.40

Great founders need intellectual disagreement without interpersonal rudeness; social skills matter as much as iconoclasm.

3 pointscentrality 3/5
  • Great founders are 'disagreeable' in that they disagree with ideas, but this does not require being disagreeable with humans; clear-thinking iconoclasm and rough interpersonal behavior are separable traits, and young male founders especially need more social skills.

    “Most of the great, great founders are disagreeable in the sense that they disagree, but that doesn't mean that they need to be disagreeable with humans.”

  • Whatever you think you're going to study or do, you may end up doing something very different, as Schmidt started in architecture, became an engineer, and graduated in electrical engineering without taking EE classes.

    “So whatever you think you're going to do, you may do something very different.”

  • Almost anyone who is successful must start by acknowledging they were lucky—lucky of birth, of family home life, of intellectual upbringing, of being early in their industry—and luck can be partly created by doing things that are interesting.

    “I would say I'm defined by luck, and I think almost anyone who's successful has to start by saying they were lucky”

0.32

Meeting dynamics and organizational communication require deliberate facilitation to prevent groupthink and ensure diverse voices are heard.

1 pointcentrality 3/5
  • In meetings, half the people talk and half don't, and women are often dominated by men in interruptions; deliberately calling on someone who hasn't spoken—who will give a thoughtful answer—and getting others to respond to her forces a leveling of discussion and builds the common knowledge that prevents organizations from going off the rails.

    “you call on the woman who hasn't had said anything. She will inevitably give an extraordinarily interesting and thoughtful answer. Then get people to respond to her, and let her speak. That kind of mechanism forces a leveling of the discussion”

0.22

Authoritarian systems achieve behavioral control through stratified incentives and information control, creating stability but stifling economic dynamism.

2 pointscentrality 2/5
  • The scariest aspect of North Korea is its ability to systematically control everyone's behavior through a hierarchy and a quasi-religious stratified structure with incentives and disincentives, such that the normal debate, complaining, and bitching of free societies does not visibly occur, while the economy remains very poor as a vestige of communism.

    “The scariest thing about North Korea is the ability to control everyone's behavior in a systematic way.”

  • Generally accepted figures put upper-middle-class life span around 90 years, rising about one year every few years, so a child born in 2030-2040 could reasonably expect a life span around 100, driven by antibiotics, better water, and now an explosion in later-life cancer and heart care—though many believe there is a natural asymptote we are bumping against.

    “The generally accepted number is about 90 years old, and the generally accepted number is it goes up on the order of one year every few years. So it's reasonable to expect that a child born in 2030 or 2040 could have a life span of about 100.”