What this covers
This conversation between Reid Hoffman and Tyler Cowen ranges across philosophy, business, technology, and culture, centered on how progress actually happens. Hoffman argues that technology is only one of several "Archimedean levers"—alongside ideas, incentives, culture, and institutions—and that as tech companies accumulate society-level power, they cannot retreat into the claim that they should be left alone. Instead, they must engage deliberately in public discourse, accept accountability for their design choices, and remain responsive to societal feedback. The discussion weaves together concrete business examples with broader philosophical commitments, including Hoffman's operating principle of Intelligent Self-Deception, captured in his "Quixotic" handle: acting as if the future is genuinely open even if determinism holds, because that stance may itself create real possibility.
The conversation covers specific critiques of how technology companies amplify societal agitation through velocity-driven design, how credential systems gate talent unnecessarily, and why persistent institutions—not just startups—produce innovations that require true scale. On AI risk, Hoffman is attentive but skeptical of existential doom, seeing pseudo-AI systems tied to infrastructure as the more concrete danger. He examines talent identification through Network Validation rather than individual picking, defends capitalism as a technology while refusing it as philosophy, and argues that Permanent Beta learning matters more than credentials for labor-market recovery. Throughout, Hoffman draws on Aristotle over Plato, favors aphoristic thinking that catalyzes reader reflection, and maintains that genuine idealism—not naivety—is worth pursuing even while acknowledging tragedy as real.
Hoffman argues that progress comes from pulling many 'Archimedean levers' at once—technology, ideas, incentives, culture, and institutions—and that as tech companies gain society-level power, they must enter discourse with society rather than insist on being left alone.
- Technology companies are an important lever but not the only one; ideas and institutions matter too
- Once a company becomes social infrastructure it must be accountable and responsive to society
- Network-based pattern recognition and non-zero-sum games are superior tools for talent, business, and morality
LinkedIn and business institutions should shift from single-transaction to multi-transaction game design.
- Hoffman hopes LinkedIn can move the business world from single-transaction games to treating their entire business life as multi-transaction games, because acting on a multi-transaction basis increases productivity and evolves society and morality toward preferring non-zero-sum games.
“have all of the business world move in the number of single-transaction games to thinking about their entire business life and relationship and work as multi-transaction games... we should prefer non–zero-sum games.”
- Board games teach that you should have a preference for non-zero-sum games, but because much of life is competitive you need an active theory of how to win when you must compete—and, importantly, of choosing when to compete, which is frequently where people err.
“you want to have a preference for non–zero-sum games, but there’s a lot of life that is competitive sailing... I think it’s also important to choose when you compete, which is frequently an error.”
- Hoffman's comparative strength on the PayPal team was the ability to tolerate very different cultures and mindsets while interfacing with massive platforms and institutions—eBay, Visa, Mastercard, regulators—and winning PayPal a seat at the table without triggering 'organ rejection.'
“my comparative strength was that I was willing to interface with these various massive platforms and institutions... how to not make PayPal be organ rejection to those entities.”
- Later Wittgenstein's insight that we play language games—forming identity and community through how we discourse, the channels we use, and the environments we are in—informed the patterns of communication built into LinkedIn.
“a central part of later Wittgenstein is to think that we play language games, that the way that we form identity and community... is the way that we discourse and the way that we see each other and the way that we elaborate language.”
Systems change requires multiple Archimedean levers—ideas, incentives, and institutional leadership—not single solutions.
- Systems should be changed through a toolset of different Archimedean levers—ideas or memes, incentives or culture, and leaders building or reforming organizations and institutions—rather than relying on any single lever.
“I tend to think in terms of systems — Dewey was a system thinker — but then I tend to think in terms of Archimedean levers, by which you change the systems. And I tend to think that there’s a number of different types of Archimedean levers. It can range from ideas or memes. It can range from incentives or culture, and it can range from leaders building organizations or institutions”
- Hoffman disagrees with the Silicon Valley meme that technology is the only Archimedean lever that matters and that companies should tell society to leave them alone; as tech reaches society-level impact, companies must deliberately engage in public discourse, accept accountability on design principles, and create channels for society to give feedback.
“once you begin to have society-level impact... we need to add in interaction with society, accountability on the design principles and the philosophy by which we are affecting society.”
- Persistent institutions at scale—not startups—tend to produce the inventions and innovations that only scale can deliver, such as building massive new platforms like cloud computing, which is why the renovation story of a great company is as important as the entrepreneurial story.
“persistent institutions at scale that do the inventions, innovations that only scale can do, like building massive new platforms, like cloud computing, tends to come from the big companies, not tends to come from start-ups.”
- Newspaper people hate Facebook more than radio people because of competitive rivalry over lost economics and attention, and Hoffman is unsympathetic to their wish to return to a golden age of newspapers—the right move is to design a better future that preserves the civil functions newspapers performed.
“the thing that I’m always unsympathetic to people on this, is they say, 'Well, I’d like to return to the past. I’d like to go back to the golden age of newspapers.'... 'No, no. Actually, how do you design a better future?'”
Talent assessment must match individual strengths to specific team contexts rather than ranking fixed traits.
- To evaluate a talent-spotter, look for three things: basic competencies (table stakes), awareness of where conventional signals mislead so they find startling talent others miss, and a differential, actively-improving theory of why they are good at spotting—since people wrongly proxy talent to prior success.
“what are the ways that you realize that the table stakes can mislead you and be wrong? Where you can find really startlingly good talent, where the traditional mechanisms don’t work?”
- Rather than ranking people on fixed personality traits, talent should be assessed as which combinations of strengths and weaknesses fit a particular team and task—evaluating whether someone is a team player or individual contributor, driven by teamwork or pure excellence, high-IQ or reliable, and strong in adversity.
“what I tend to be thinking about is which combinations of strengths and weaknesses are the right fit on this team. And I tend to be looking for that set.”
- Aristotle's theory of the metals—ranking people as gold, silver, or bronze within a society—is too simplistic and underdeploys the possibilities of talent; a better lens recognizes that different people excel at different tasks and teams, requiring an active theory matching task to the game being played.
“one of the things I disagree with Aristotle on is his theory of the metals... I think those kinds of hierarchies tend to be simplistic and tend to underdeploy the possibilities of talent.”
- Hoffman's unique insight in talent search is to favor network validation over being a 'picker'—almost never meeting an entrepreneur without a trusted introduction, and trusting different referrers on different attributes—because patterning the network to bubble up talent is far better than individual selection.
“I almost never meet with an entrepreneur that doesn’t come from an introduction from someone I trust. And part of that introduction from someone I trust is, I trust different people on different things.”
- Businesses are bad at certifying talent because no dynamic set of cheaply-recognized certifications exists, so employers default to coarse credentials like 'must have a BA'; richer modern certificates could unlock paths and talent currently gated by expensive reference-checking and the proxy of a degree.
“we haven’t established a dynamic set of certifications that we can use as signifiers that can then be deployed in a lot of circumstances. That’s one of the reasons why most jobs are described as 'must have a BA or a BS'”
- Effective reference checks ask for the person's greatest weakness (threatening to assume the worst if no believable answer is given) and whether the person is a top-five choice for a new venture (then asking who the other four are), which calibrates how analytically credible the referrer is.
“if you don’t give me one that’s believable, I’ll believe that the thing is so bad you can’t tell me about it... 'Is this person the top-five person that you would choose to do a new venture with?'... 'Okay, who are the other four?'”
Human nature is complicated and multidimensional, requiring nuanced theories in strategy, science fiction, and product design.
- Hoffman's view of human beings as complicated entities with many kinds of expression and motives came from childhood and especially boarding school at Putney, while his network and incentive-based systems thinking did not.
“the diversity of human experience, of thinking of human beings as complicated entities that have lots of different kinds of expression and lots of different motives, probably did come from childhood and most especially boarding school at Putney.”
- Science fiction usually fails because its theories of human nature are too simplistic and its characters too one-dimensional; in reality, motives are complicated, ego is a spectrum shaped by triggers, desires, and self-narratives rather than a single binary.
“usually science fiction is too simplistic in its theories of human nature, that the characters are, to some degree, too uni-dimensional.”
- People misunderstand friendship as strict loyalty to the other person, but real loyalty is to the person's aspirational better self—who they are in the world and what they are doing—which is why you would not help a friend who turned out to be a serial murderer.
“what you are is loyal to the person’s aspirational better self... It’s who they are in the world and what they’re doing that’s actually, in fact, where the loyalty should lie.”
- Hoffman is a non-dualist in the Buddhist sense: there may be a spread of substances, possibly including a spiritual nature beyond the material plane, but they are not disconnected, and one should remain open to whether all of it resolves to materialism.
“I am what Buddhists refer to as a non-dualist... there may actually be some kind of spread of substances, but they’re not actually, in fact, disconnected.”
Entrepreneurs should use philosophy to articulate investment theses grounding strategy in explicit theory of world and human nature.
- Entrepreneurs should use philosophy's discipline of clarity to write out their clearest set of investment theses: what they believe the world is and is becoming, their strategy to make it happen, why they can hold a rare or unique position, and an explicit theory of human nature underlying the product.
“to write out your clearest set of investment theses about what it is you believe the world is and is becoming, what your strategy is at making it happen, and why you can have a rare or unique position in so doing.”
- Teaching the skills of entrepreneurship is important for societal progress and for solving the future of middle-class jobs, which is why Hoffman would design a political-economy board game about creating companies.
“I actually think teaching the skills of entrepreneurship is important for how do we make progress in society, how do we solve the future of middle-class jobs”
- Aristotle's intersection of studying the world and thinking is a better form of philosophy than Plato's, because philosophy is not something done sitting in a cave but something you participate in within society.
“philosophy is not the thing that you sit in the cave doing, but something you do as you participate in society.”
- Hoffman reads The Tempest as emphasizing that individual potential matters but the macro point is rejoining society with what you've learned in the magical place—contrasting with Thiel's reading that elevates the primacy of the self-creating individual.
“the reading of The Tempest is the importance of starting in this magical place... but then rejoining society with the things you’ve learned. And your contribution to the society as an individual is the macro reading of The Tempest.”
Regulators and innovators need dialogue, not dismissal, to design better systems around money, speech, and emerging technologies.
- Freedom of speech is something society enables to help it evolve toward truth, but it does not mean any lie will do; specific regulation of internet speech (truth in advertising, liability laws) is compatible with free speech.
“Freedom of speech is something that we enable as a society to help us evolve towards truth, but it doesn’t mean that any lie will do.”
- PayPal persuaded regulators it was not a bank by arguing it was a 'store of value' like gift cards or phone cards rather than a lender, and by demonstrating consumer protection through pass-through FDIC insurance and presenting itself as accretive rather than destabilizing to the system, leading regulators to adopt a wait-and-see stance.
“there’s a lot of things that store value... PayPal is another store of value. But it’s not a store of value in the way that a bank is a store of value, e.g., making loans... we figured out how to actually, in fact, have pass-through FDIC insurance”
- The world will need one or more cryptocurrencies; crypto builders wrongly dismiss banks and sovereigns who actually need an interface point on how monetary systems work, while bankers wrongly dismiss crypto builders as mere hustlers—so a debate format can let both sides learn.
“I argued in a Wired article that I wrote in 2015 that the world will need one or more cryptocurrencies... the crypto folks don’t realize that the banks and the sovereigns also need some interface point on how monetary systems work.”
- The claim that money doesn't influence elections is like saying marketing doesn't work; money does have influence, so while Hoffman would prefer democracies where money has a more muted role, zero money in politics is not the right answer because it would reward only celebrities like Kardashians.
“I actually disagree with the view that money doesn’t have an influence. That view is like saying marketing doesn’t work, and it’s clear that marketing can work.”
Contingency and uncertainty—not determinism or perfect strategy—shape real outcomes more than talent or product-market fit alone.
- The underestimated cause of PayPal's success is that it never targeted eBay—launching a general payment service that took off on eBay almost by accident; contrary to deterministic narratives about talent and product-market fit, if eBay hadn't existed PayPal would never have succeeded.
“when PayPal got started, it actually wasn’t paying any attention to eBay... And if eBay hadn’t existed, PayPal would have never succeeded.”
- Avalon Hill war games are better models of real strategy than chess because they include a gap between epistemology and metaphysics—uncertainty, fog of war, and randomness—forcing players to keep fallback plans and model how opponents think.
“Those games always struck me as inaccurate to real world because they didn’t have a gap between epistemology and metaphysics. They didn’t have the uncertainty, the fog of war. They didn’t have randomness elements to them”
- The strategic choice in blitzscaling is knowing when to move fast and when not to; the rule is not to always move fast but to be capable of moving super fast when you need to.
“the choices — and I’m sure we’ll get to this blitzscaling — is when to move fast and when not to move fast. It’s not always move fast, but be capable of moving super fast when you need to.”
- Even if the universe is fully deterministic, one should act as if it is not—captured in Hoffman's 'Quixotic' handle—because acting that way may create a possibility that the system is not all deterministic, raising the open question of whether this is intelligent self-deception or genuine agency.
“if it’s all determinism, maybe we should act as if it’s not, right? Even if it is.”
Education and policy must address entrepreneurship and continuous learning to support middle-class labor-market adaptation.
- The slow labor-market recovery from the financial crisis stems partly from declining entrepreneurship and new-firm formation across the US (outside Silicon Valley), and partly from educational systems failing to teach the 'permanent beta' of continuous learning needed to adapt to new jobs.
“while entrepreneurship is doing quite well in Silicon Valley and a few other places, it’s actually been on the decrease across the US. We need to get back to the formulation of new firms”
- The nature-nurture question is a false dichotomy; both always combine, so the real question is how they interact rather than which one dominates.
“The nature-nurture question, which is always some blend of both because we always have these false dichotomies — nature or nurture — and it’s always, actually, how do they combine?”
- Top colleges and universities are operating at roughly 30 percent of their potential capacity as talent certifiers, and Silicon Valley is only beginning to compete with them through network-based talent recognition programs and coding camps.
“This will be an unpopular answer because I have a bunch of friends who are great leaders in these institutions, but I would generally say probably about 30 percent.”
- One can articulate very good conservative values that do not try to lock in the past against the future but instead articulate principles—around concepts like liberty—that guide your own evolution of thinking, which is what Matthew Arnold's best values represented.
“one can articulate very good conservative values that are not trying to lock in the past against the future, but trying to articulate principles that help guide you in your own evolution of thinking about concepts like liberty”
- Hoffman attended the Putney School in Vermont, a John Dewey-model school emphasizing diversity of human experience including art, blacksmithing, and farm work such as driving oxen.
“I went to this great high school in Vermont called the Putney School... some shoveling of cow manure, and some driving of oxen.”
- A good aphoristic thinker strings aphorisms together so that the reader catalyzes their own thinking; unlike prose with thesis-argument-conclusion structure, aphorisms aim to shift the reader's Weltanschauung and move them out of their local maxima or minima rather than deterministically from point A to point B.
“An aphorism is to say, we’re trying to change your mindset, the German Weltanschauung... It’s not to a deterministic, move you from point A to point B, but to catalyze you from moving out of your local maxima or minima.”
Science fiction matters as a lens onto society and needs richer human theories and more utopian narratives about technological futures.
- Science fiction matters because by constructing artificial future societies it functions as a lens onto our own, surfacing the interplay of politics, religion, and the evolution of society.
“science fiction is, by making these artificial future societies, a lens onto our own”
- Current science fiction has two failures—too little utopian (especially video) science fiction despite our hope for the future lying in technology, and a failure to address known coming trends like pervasive AI—so new science fiction narratives may be the right form to provoke needed discussion about human progress.
“there’s too little utopic science fiction, and yet if we have a hope for the future, it’s through technology, it’s through change... we know that what’s coming is... artificial intelligence will affect all systems in the world”
- Hoffman is attentive but not worried about AI doom; the more probable danger is not a malicious evolving superintelligence but pseudo-AI systems tied to something like a nuclear defense grid producing unintended-consequence triggers, and a nonzero probability doesn't justify obsession—just as we no longer obsess over nuclear war despite armed, sometimes crazy nations.
“we take pseudo–artificial intelligence systems, we tie them to a nuclear defense grid, and something that’s an unintended consequence... That would be the category of thing that I would have more worry than you have a malicious evolution of a superintelligence.”
Institutional capacity and scale, not only startup disruption, matter for large-platform innovation and economic renewal.
- The topple rate of the S&P 500 is increasing—more previously-listed companies drop out within a decade—reflecting that the networked age brings more revolution in which firms are on top, which is good because it signals competition and new platforms while enduring institutions remain valuable.
“the so-called topple rate of the S&P 500 is increasing, by which the number of companies that were previously in the S&P 500 in the last decade are now dropped out of it altogether.”
- Capitalism is one of the world's most interesting technologies, but it is not a particularly good philosophy—an unconventional view for an entrepreneur or investor to hold.
“the fact that I actually think that the notion of capitalism is one of the world’s leading interesting technologies, but it’s not a particularly good philosophy, and you’d think that’s odd for an entrepreneur or an investor”
- Bitcoin will persist, and a Chinese e-currency is very possible because China's large market could enable a network to start there, with at least one additional significant cryptocurrency likely beyond Bitcoin.
“I think there will be Bitcoin. Don’t yet know on the e-currency, but I think it’s very possible, especially because there’s such a large market within China... And my guess is it’ll be at least one more additional.”
Idealism and optimism about human potential are worthy life goals, compatible with acknowledging tragedy and embracing mystery.
- Being 10-out-of-10 idealistic is not naivety; it is the actual goal of our lives and what we should work toward, and giving up on optimism and idealism is itself a tragedy—though idealism must not deny that tragedy is real.
“part of the reason to be 10-out-of-10 idealistic is not to be naïve... but that is actually, in fact, the goal of our lives and what we should work towards. So, ever giving up on optimism and idealism is, I think, a tragedy in and of itself.”
- Keeping mysticism in the picture without God means not being overly certain about your current theory of the universe—since some present beliefs will look stupid in 50-100 years—staying open to a possible mystical force creating the universe while treating specific revealed deities as less likely.
“the question about keeping mysticism in the picture is to say, 'Don’t be overly certain about your current theory of the universe.'... there are certain things we believe today that 50 or 100 years from now will be thought to be stupid and crazy.”
Velocity-optimized systems amplify negative content and agitation, while GDP metrics miss value creation from free digital goods.
- Productivity and GDP growth appear slow despite more technology because GDP measures dollars flowing through the system rather than value; free tech goods like Wikipedia and entertainment substantially increase value while removing dollars from the system, so the same dollar flow yields a flat GDP even as real value rises.
“one of the problems of the GDP measure, as you know, is that it measures dollars flowing in the system, not in value flowing in the system.”
- A social system designed for velocity as its measure will amplify the content people negatively react to—anger, outrage, fear—because people share links before reading them, producing clickbait cycles and an overall increase in societal agitation even though each individual is just following their own reflex.
“if you have a system that’s designed for velocity as its measure, velocity is more likely to be things that you negatively react to. It’s more likely to be anger. It’s more likely to be outrage. It’s more likely to be fear.”