PJFP.com

Pursuit of Joy, Fulfillment, and Purpose

Category: AI

  • Dario Amodei on Policy for the AI Exponential: Anthropic’s Plan for AI Regulation, Job Displacement, Civil Liberties, and Democratic Leadership

    In June 2026, Anthropic CEO Dario Amodei published “Policy on the AI Exponential”, a wide-ranging essay arguing that the gap between how fast AI is advancing and how slowly policy moves has become dangerous, and that the window to close it is open right now. He opens with a memorable image from The Lord of the Rings: the Hobbits trying to rouse Treebeard, the ancient tree who takes a full day just to say hello, to defend his forest before it is cut down. That mismatch in speed, he writes, is exactly the relationship between AI and our political institutions. This post breaks the essay down in full and adds analysis of where the argument lands.

    TLDR

    Amodei argues that AI’s scaling laws point toward “powerful AI,” a country of geniuses in a datacenter, within a few years, while legislation still moves on a timescale of years. For most of the last few years, safety advocates including Anthropic pushed only for optionality-preserving moves like transparency rules, chip export controls, and labor data collection, because the risks were not yet concrete. He says that has changed: events like Claude Mythos Preview proved frontier models are now tools of national strategic consequence, and the time for binding regulation has arrived. The essay covers five policy areas. First, regulation and public safety, where he proposes an FAA-style regime of mandatory third-party testing of frontier models above a compute threshold across four risks (cybersecurity, biological weapons, loss of control, and automated R&D), with government power to block unsafe deployments. Second, macroeconomics and tax policy, where AI could deliver hypergrowth and severe, enduring job displacement at the same time, demanding measurement, pro-employment incentives, and possibly UBI or universal capital accounts. Third, accelerating AI’s positive impact, where the danger is regulators like the FDA being too slow rather than too lax, and biomedical approval needs reform. Fourth, the state and civil liberties, where AI could become the ultimate tool of autocracy through autonomous weapons and mass surveillance, requiring new accountability rules, a domestic ban on autonomous weapons, closing the data broker loophole, and public rights to AI advice. Fifth, securing leadership by democracies through a values-based global coalition that controls the AI supply chain, coordinates on risk, shares benefits, and rejects AI-powered repression. He closes by rejecting the idea that public concern about AI is a PR problem to be marketed away, calling it democratic accountability working as it should.

    Thoughts

    The most important move in this essay is structural, not technical. Amodei is explicitly retiring the “preserve optionality” posture that defined Anthropic’s policy work through 2025 and replacing it with a call for binding rules. For years the argument from safety-minded labs was that the risks were too speculative to legislate against without doing more harm than good, an idea he grounds in the Collingridge dilemma and the Hayekian point that regulators lack the information to make good calls. That was a defensible hedge. What is striking here is the claim that the hedge has expired. He is saying the evidence is now concrete enough that continued caution about regulating has flipped from prudent to negligent. Whether you trust the underlying capability claims or not, that is a genuine change in position from one of the field’s most influential voices, and it deserves to be read as such.

    The FAA analogy is doing enormous work, and it is worth poking at. Airplanes and drugs are mature technologies with stable physics and decades of incident data; the certification regime works because the failure modes are well understood. Frontier models are the opposite: the whole premise of the essay is that capabilities are changing faster than anyone can characterize them. Amodei half-acknowledges this when he warns that a fixed list of safety requirements tends to consume 95 percent of compliance effort on things that turn out not to matter while missing the real risks, a lesson he says Anthropic learned from its own Responsible Scaling Policy. So the proposal is really for an agency nimble enough to rewrite its own standards continuously, which is a much taller order than the FAA. The honest read is that he is proposing a regulator we do not yet know how to build, and betting that building it is still better than the alternative.

    The economics section is where Amodei is most careful, and it is the part most likely to be misread. He goes out of his way to say enduring job displacement is undesirable and that warning about it is not the same as wanting it, a distinction critics of AI leaders often collapse. His real claim is subtle: that AI might jam the economic policy dial on a “hypergrowth, hyper-inequality” setting that is hard to unstick, because AI substitutes for human cognition broadly and faster than past technologies, potentially overwhelming the usual escape hatches like comparative advantage and Jevons paradox. If he is right, the political fight of the next decade is not about growth, which AI supplies, but about distribution, which it does not. His mention of UBI, universal capital accounts, and higher capital gains taxes is notable coming from a frontier CEO, even hedged as it is.

    The civil liberties section is the one that should travel furthest beyond the AI-policy bubble, because it does not depend on accepting his most aggressive timelines. The data broker loophole, the idea that the government can simply buy the bulk data Americans hand to private companies and run mass analysis on it, is a problem that exists today; AI just raises the stakes by making that data vastly more revealing. Same with the proposal that anyone facing adverse government action should have access to AI at least as capable as what the government uses against them. These are concrete, near-term, and bipartisan in a way the abstract autonomy debates are not. The most candid line in the whole piece is his admission that AI cannot be safely entrusted to either governments or companies, an unusually direct acknowledgment that his own industry needs external checks, with Anthropic’s Long-Term Benefit Trust offered as one imperfect example rather than a solution.

    The geopolitics section is the most contested terrain. Framing AI as a nuclear-scale reset of the game board, with a virtual country of 100 million geniuses divisible across military strategy and weapons R&D, leads naturally to a democratic coalition that hoards chips and denies them to adversaries. That logic is internally consistent, but it sits in tension with the benefit-sharing and “eventually the whole world joins” language elsewhere in the same section. Export controls that lock down the supply chain are, by design, a tool of exclusion, and reconciling that with broad diffusion of AI’s benefits to developing countries is the circle the coalition idea has to square. Amodei is clearly aware of the tension and bets that making membership attractive resolves it. The closing image is the one to remember: Treebeard waking up, with the warning that the goal is to channel real public concern into constructive policy rather than let it curdle into formless anger.

    Key Takeaways

    • The core tension of the essay is a mismatch in speed: AI advances exponentially while legislation moves on a multi-year timescale, dramatized by the Treebeard and Hobbits image from The Lord of the Rings.
    • In only four years, AI models went from barely writing a coherent line of code to writing most of the code at major AI companies, with similar gains across biology, physics, math, finance, law, and translation.
    • Scaling laws now have over a decade of empirical support, and if they continue another year or two they likely produce “powerful AI,” a country of geniuses in a datacenter.
    • For the last few years, safety advocates including Anthropic focused on optionality-preserving policies: transparency legislation, chip export controls, and data collection on AI’s labor effects.
    • Amodei argues that posture is no longer enough. Claude Mythos Preview revealed that frontier models pose real cybersecurity risks to the financial sector, critical infrastructure, and national security, and proved AI is now a tool of strategic consequence.
    • He expects biological risks to follow cyber risks, with serious AI autonomy risks potentially not far behind.
    • The essay covers five policy areas: regulation and public safety, macroeconomics and tax policy, accelerating AI’s positive impact, the state and civil liberties, and securing leadership by democracies.
    • Alongside the essay, Anthropic released a legislative proposal on frontier model testing and a policy framework for job displacement, both with promised financial backing.
    • On regulation, Amodei invokes the Collingridge dilemma and Hayek’s information problem to explain why pre-writing AI law in 2023 to 2024 was risky, then argues the situation has now changed.
    • Anthropic’s 2025 answer was transparency, helping pass SB 53 in California, RAISE in New York, and SB 315 in Illinois, plus advocating a federal transparency standard.
    • He now calls for binding regulation modeled on the FAA, where frontier models must pass technical testing and can have release blocked or reversed if they fail high safety standards.
    • Models above a compute threshold should face mandatory third-party testing in four areas: cybersecurity, biological weapons, loss of control of AI systems, and automated R&D that accelerates the other three.
    • Government should be able to block or deter deployment of models judged to present unacceptable risk, scoped to those four risks with protections against political favoritism.
    • Evaluation could come from a government agency or from authorized and inspected private organizations under a “regulatory markets” approach.
    • AI companies should have strong security to protect model weights, conduct regular red teaming and penetration testing, report safety incidents promptly, and work with government against major threat actors.
    • He warns a time may come when the most powerful systems resemble weaponizable nuclear materials rather than airplanes, requiring more aggressive measures, but cautions against getting ahead of present dangers.
    • On economics, AI could deliver extremely rapid growth via accelerated science and operational efficiency, supercharged by AI building better AI.
    • The same properties make AI a broad substitute for human cognition that changes the economy faster than past technologies, risking large and potentially enduring labor market disruption.
    • The feared outcome is a “hypergrowth, hyper-inequality” setting that is hard to unstick, where the challenge shifts from incentivizing growth to sharing its benefits.
    • Amodei is emphatic that enduring job displacement is undesirable and dangerous, and that he warns about it to help society adapt, not as a prophet of doom.
    • Anthropic says it works with customers to find new revenue and use cases rather than only cost cutting, and explores interaction paradigms that keep humans active alongside AI.
    • He predicts AI will enable single individuals to build billion-dollar companies, noting teams of a few people already reach hundreds of millions in revenue, while admitting significant enduring job loss may be intrinsic to the technology.
    • Any response must address both economic provision and the human need for meaning, purpose, and agency, with the latter ultimately more important and beyond what policy can directly deliver.
    • Suggested economic interventions: better measurement and tracking (governments expanding statistics beyond Anthropic’s Economic Index), pro-employment incentives, and long-term macroeconomic support.
    • Pro-employment ideas include wage insurance, retention tax incentives, workforce training grants, and employer-employee matching infrastructure.
    • If displacement is large and permanent, mechanisms like universal basic income or universal capital accounts, financed through company taxes or higher capital gains taxes, may be necessary.
    • He frames datacenter and energy-price backlash as largely a symbol of broader economic anxiety, and says AI companies should pay to absorb rate increases, a pledge Anthropic has already made.
    • For technologies accelerated by AI, the bigger risk is regulators like the FDA being too slow, not too lax, because AI may make downstream tech safer in ways that violate skeptical regulatory assumptions.
    • Biomedicine is the illustrative case: AI could flood the drug pipeline, raise effect sizes, treat previously untreatable diseases, and create whole new therapy categories, while the current FDA and EMA pipeline takes 7 to 8 years.
    • Agencies should pre-approve standards for AI methods like PD/PK modeling, toxicology prediction, dose selection, biomarker validation, synthetic control arms, and surrogate endpoints, plus more flexible accelerated-approval mechanisms.
    • On civil liberties, powerful AI in the wrong hands could be the ultimate tool of autocracy, and existing constitutional protections are not fully equipped to counter a surprise seizure of power.
    • Threats named include fully automated drone armies that obey unlawful orders and surveillance AI that infers the innermost details of every citizen’s life from widely available data.
    • Civil liberties proposals: accountability rules and an “off switch” for autonomous weapons, a domestic ban on fully autonomous weapons including in law enforcement, closing the data broker loophole, and public rights to AI advice during adverse government action.
    • Amodei warns companies as well as governments can seize quasi-state power, citing the Gilded Age and the East India Company, and says AI cannot be safely entrusted to either alone.
    • He offers Anthropic’s Long-Term Benefit Trust as one separation-of-power structure and urges the industry to explore mechanisms that go further.
    • On geopolitics, he argues AI resets the geopolitical game board like nuclear weapons, becoming the dominant source of military and economic power for any nation that holds it.
    • A nation with powerful AI versus one without it, or even one three years behind, could resemble WWII Marines facing medieval swordsmen.
    • He calls for a democratic coalition that shares chips and semiconductor manufacturing equipment internally while denying them to adversaries, citing MATCH and OVERWATCH as good first steps.
    • The coalition should coordinate risk policy, share benefits including harmonized medical approvals, provide mutual AI defense, reject AI-powered repression, and cooperate on macroeconomic stabilization.
    • He rejects the idea that AI’s image is a PR problem, arguing public concern reflects real risks and is democratic accountability working as it should, with the task being to channel it into constructive solutions.

    Detailed Summary

    The speed mismatch between AI and policy

    Amodei frames the entire essay around a single problem: AI advances at a lightning pace while policy, especially legislation, moves very slowly, often for good reasons since governments wield grave powers that should not be used hastily. He illustrates this with Treebeard, the sentient tree from The Lord of the Rings who takes a full day to say hello, as a stand-in for political institutions trying to respond to a technology that can go from amusing toy to a country of geniuses in the time it takes Congress to act. He recounts the dilemma responsible actors have faced: they could see where the exponential was headed, but to observers looking only at present capabilities, AI looked as mundane as the latest consumer app or cryptocurrency, making a laissez-faire attitude hard to argue against. The absence of AI’s radical effects, and uncertainty about their shape, made it genuinely difficult to design good policy even where the will existed.

    That uncertainty, he says, is why safety advocates limited themselves to optionality-preserving measures like transparency rules, export controls, and labor data collection. But over the last few months the evidence of AI’s power and risk has become undeniable, with Claude Mythos Preview as the emblematic example: it scrambled the global cybersecurity landscape and proved AI models are now tools of global and national strategic consequence. He expects biological and autonomy risks to follow, and argues the world must now activate its slow, rickety policy apparatus to handle risks that will compound quickly. He worries current early actions are at least a year out of step with AI’s progress, and presents the essay as an attempt to close that gap across five policy areas, focused on US policy but relevant worldwide.

    Regulation and public safety: an FAA for frontier models

    Amodei opens by acknowledging the real costs of regulation: it can reduce a product’s benefits, disincentivize innovation, and suffer from the Hayekian problem that regulators lack the information for good tradeoffs, plus the Collingridge dilemma that a technology’s impacts are hard to anticipate until it is too late to manage them. In 2023 to 2024 these dynamics argued against pre-writing AI law, since the exact form of biological or autonomy risk, how to test for it, and how it would play out were all unclear, creating a high risk of low-value compliance requirements that miss the real dangers. Anthropic’s answer was transparency: requiring developers to disclose safety procedures, tests, and critical incidents, which is why it supported SB 53 in California, RAISE in New York, and SB 315 in Illinois in early 2026.

    Now, he argues, the risks are clearly here and it is time for binding regulation. His analogy is to cars, airplanes, and drugs: powerful technologies essential to the economy but capable of killing many people if designed or operated poorly. He models AI regulation on the FAA, with frontier models required to pass testing and auditing and with release blocked or reversed if they fail high safety standards. His concrete proposal: mandatory third-party testing for models above a compute threshold across cybersecurity, biological weapons, loss of control, and accelerating automated R&D; government power to block deployment of unacceptably risky models, scoped narrowly with anti-favoritism protections; evaluation by either a government agency or authorized private organizations in a regulatory-markets model; strong weight security, red teaming, and penetration testing at AI companies; and prompt reporting of safety incidents. He notes a future may arrive when systems resemble weaponizable nuclear materials and demand harsher measures, but warns against designing for dangers that have not yet emerged.

    Macroeconomics and tax policy: growth and displacement together

    Here Amodei challenges the standard premise that growth is fragile and must be traded off against the drag of taxes or deficits to reduce inequality. Powerful AI, he suggests, may scramble that assumption by producing extremely rapid growth through accelerated science and efficiency, supercharged by AI building better AI, while simultaneously acting as a broad substitute for human cognition that reshapes the economy faster than any prior technology. The result could be a world stuck on a hypergrowth, hyper-inequality setting that is hard to unstick, where the central challenge is no longer incentivizing growth but sharing its benefits. He is careful to make two points clearly: first, enduring job displacement is undesirable and dangerous and should be minimized, and his warnings are meant to help society adapt, not to play prophet of doom; second, any response must address both economic provision and the deeper human need for meaning, purpose, and agency, which matters more and which policy cannot directly supply.

    His policy menu starts with measurement and tracking, arguing good policy is impossible without accurate data, and that governments could expand economic statistics well beyond Anthropic’s Economic Index. Next come pro-employment incentives such as wage insurance, retention tax incentives, workforce training grants, and employer-employee matching, costs he says society should readily accept since they are likely offset by AI productivity gains. If displacement proves large and permanent, he says long-term income support like universal basic income or universal capital accounts may be needed, financed through taxes on relevant companies or higher capital gains taxes. He closes the section by reframing datacenter and energy-price backlash as mostly a symbol of broader economic anxiety, while saying AI companies should absorb rate increases, as Anthropic has pledged.

    Accelerating AI’s positive impact: the slow-regulator problem

    For technologies accelerated by AI, rather than AI itself, Amodei flips his concern: the bigger danger is regulatory systems designed for a slower pace failing to handle the deluge of new products, and AI making downstream technologies safer in ways that violate the skeptical assumptions baked into agencies like the FDA. He focuses on biomedicine as the area likely to produce AI’s biggest humanitarian benefits and where regulation is especially complex. AI could greatly increase the rate of new drug candidates, improve their effect sizes and safety profiles, treat previously untreatable diseases, and create entirely new therapy categories the way antibodies, peptides, and cell therapies did.

    The current pipeline at the FDA and EMA takes 7 to 8 years, built on the pessimistic assumption that drug candidates usually fail and often carry safety problems even when they work. Without reform, AI will jam or overload that system. Amodei proposes that agencies develop standards now for accepting AI simulation and analysis, so they can be adopted quickly once proven rather than after years of unnecessary testing. Specific candidates include AI-based PD/PK modeling, toxicology prediction to reduce animal testing, more accurate dose selection, biomarker validation from large datasets, synthetic control arms, and surrogate endpoints (especially for aging and neurodegeneration). He urges more flexible accelerated-approval mechanisms generally, and notes biomedical acceleration may also reduce AI’s risks by aiding biodefense and improving mental health.

    The state and civil liberties: guarding against AI-driven tyranny

    Amodei frames the perennial balance between state power and individual liberty, enforced through machinery like the First, Fourth, and Fifth Amendments, the Posse Comitatus Act, and FISA, and argues AI threatens to upset that balance while raising its stakes. Powerful AI in the wrong hands could be the ultimate tool of autocracy, because the enormous returns to intelligence combined with AI’s pace create a perfect storm for a surprise seizure of power. The danger could take many forms but shares one feature: AI conferring sudden power while routing around democratic oversight. He cites a fully automated drone army that could obey unlawful orders, where trained humans might object, and a surveillance AI that analyzes widely available information at massive scale to infer the innermost details of every citizen’s life, an ability current civil liberties law never contemplated.

    His proposals: create accountability rules for autonomous weapons so they respond to court orders, legislation, and human overseers rather than blindly following orders, possibly with a judicial finger on an off switch; ban domestic use of fully autonomous weapons, including in law enforcement, while allowing them against foreign adversaries; close the bulk-collection and data-broker loophole that lets the government buy and analyze data Americans share with private companies; and guarantee public rights to AI advice at least as capable as what the government uses during adverse action, as an extension of the Administrative Procedure Act, due process, or the Sixth Amendment. He closes by warning that companies, not just governments, can capture the state, citing the Gilded Age and East India Company, and argues AI cannot be safely entrusted to either alone. Anthropic’s Long-Term Benefit Trust is offered as one accountability structure, with a call for the industry to go further.

    Securing leadership by democracies: a values-based coalition

    Amodei rejects treating AI as a mere instrument of trade policy to diffuse a tech stack worldwide. He believes AI resets the entire geopolitical game board like nuclear weapons, potentially even more so, becoming the dominant source of military and economic power for whoever holds it. In a virtual country of 100 million geniuses, millions could be assigned to military strategy, drone manufacture, weapons R&D, intelligence, and scientific advancement at once, so a nation with powerful AI facing one without it, or even three years behind, could be like WWII Marines against medieval swordsmen. Because powerful AI also enables deeper autocratic repression, it matters enormously that the world’s strongest nations are democracies.

    His answer is a global coalition built on shared democratic values that draws in the rest of the world by making membership increasingly attractive and exclusion increasingly costly. Operating principles include managing the AI supply chain by sharing chips and semiconductor manufacturing equipment within the coalition while denying them to adversaries, expanding and tightening export controls (he cites MATCH and OVERWATCH as good first steps); coordinating on biological, cyber, and autonomy risk to make compliance compatible and effective; sharing AI’s benefits including harmonized medical approvals; mutual defense through collective AI cyberdefense, drones, manufacturing, compute, and intelligence; rejection of AI-powered repression; and macroeconomic cooperation against contagious employment crises. The coalition would respect each nation’s sovereignty, start with aligned democracies, and grow iteratively, ideally toward the whole world, but at minimum positioning democracies to contain and outcompete repressive regimes.

    A window of opportunity

    Amodei closes on cautious optimism. The same exponential that strains policymaking has created a unique opening: clear evidence of AI’s risks, an early taste of its value and disruption, and public backlash against unregulated approaches have left policymakers unusually open to forward-looking action. Treebeard and his forest are waking up. He firmly rejects the industry-circle view that this is a PR problem solved by better marketing, arguing people are worried because the risks are real, and that public concern in response to transparency is democratic accountability working as it should. The key challenge is focusing that concern into constructive solutions rather than letting it descend into formless anger and violence. He is optimistic because issues from job displacement to model testing to export controls have common-sense appeal across the political spectrum, and a broad nonpartisan coalition could adopt sane, forward-looking policy faster than usual.

    Notable Quotes

    “in only four years, AI models have gone from barely being able to write a coherent line of code to writing most of the code at major AI companies.”

    Dario Amodei, on the pace of the AI exponential

    “in the several years that it can take Congress to act, AI can go from an amusing toy to the full country of geniuses.”

    Dario Amodei, on the mismatch between AI’s speed and the speed of legislation

    “However, now the risks are clearly here. It is time to go beyond transparency to more serious and binding regulation of AI.”

    Dario Amodei, marking the shift from transparency to binding rules

    “enduring job displacement is undesirable and dangerous, and we should do everything we can to minimize or prevent it, not to bring it about.”

    Dario Amodei, clarifying his stance on AI and jobs

    “The key challenge in such a world won’t be incentivizing growth, but finding a way for everyone to share in the benefits.”

    Dario Amodei, on a hypergrowth, hyper-inequality economy

    “Powerful AI in the wrong hands could be the ultimate tool of autocracy, and our existing legal and constitutional protections are not fully equipped to counter this threat.”

    Dario Amodei, on AI and civil liberties

    “A nation that possesses powerful AI facing one without it … could be the equivalent of an army of World War II Marines facing an army of medieval swordsmen.”

    Dario Amodei, on AI as the dominant source of geopolitical power

    “People are worried about AI because they correctly perceive that its risks are real, not because AI CEOs have been insufficiently Panglossian.”

    Dario Amodei, rejecting the idea that AI has a PR problem

    “Treebeard and his forest are waking up.”

    Dario Amodei, on policymakers’ new openness to acting on AI

    “Policy on the AI Exponential” is a dense, structured argument from one of the most consequential figures in the field, and it rewards a full read in the original. The summary and analysis above are a guide, not a substitute. You can read the full essay here.

    Related Reading

  • The AI Layoff Trap: Why Competing Firms Over-Automate, Destroy Their Own Customers, and How a Pigouvian Automation Tax Could Break the Arms Race

    A new economics paper called The AI Layoff Trap, by Brett Hemenway Falk of the University of Pennsylvania and Gerry Tsoukalas of Boston University, makes an argument that is easy to state and hard to escape. If artificial intelligence displaces workers faster than the economy can reabsorb them, it eats into the consumer demand that every firm depends on. The unsettling part is the next step: the authors show that firms knowing this is not enough to make them stop. Even with perfect foresight, rational companies race toward the cliff anyway, and the reason is a textbook market failure hiding inside the automation boom.

    TLDR

    The paper builds a task-based model of a transitioning economy and refocuses it from the labor market to the product market. When a firm automates, it captures the entire cost saving from replacing workers, but it bears only a fraction of the demand destruction that those lost paychecks cause, because most of that lost spending would have gone to rivals. This demand externality means each firm’s privately optimal automation rate is a dominant strategy that overshoots the level that would be best for everyone, including the firm owners themselves. Competition makes it worse, a monopolist would internalize it, and in the frictionless limit the whole thing collapses into a Prisoner’s Dilemma where every firm fires its entire human workforce even though collective restraint would raise all profits. Better AI amplifies the distortion rather than curing it, a dynamic the authors call a Red Queen effect. They test six policy responses. Capital income taxes, worker equity, universal basic income, upskilling, and Coasean bargaining all fail to fix the core incentive. Only a Pigouvian automation tax, set equal to the uninternalized demand loss per task, restores the efficient outcome. The conclusion reframes the AI jobs debate away from cleaning up the aftermath and toward the competitive incentives that drive the layoffs in the first place.

    Thoughts

    The cleverest move in this paper is where it points the camera. Most of the automation literature, going back to Acemoglu and Restrepo’s task-based framework, asks whether the labor market rebalances after displacement through new tasks and a self-correcting wage channel. Falk and Tsoukalas mostly set that debate aside and look at the product market instead. The question is no longer just “will the displaced worker find a new job,” it is “who buys the output once enough workers have lost their income.” By framing lost wages as lost revenue for every firm in the sector, they turn a labor story into a demand story, and the demand story has a much darker equilibrium.

    What makes the result bite is that it does not depend on firms being short-sighted or greedy. The authors grant every firm perfect foresight. Everyone can see the demand cliff ahead. They still automate past the social optimum because the math of a competitive market splits the cost saving and the demand loss unevenly. You keep all the savings from firing your workers. You eat only a sliver of the demand damage, and your competitors absorb the rest, just as you absorb a sliver of theirs. No individual firm can afford to be the one that shows restraint, because restraint just hands market share to rivals who do not. This is a genuine externality, not a coordination failure, which matters because coordination failures can sometimes be solved by communication and this one cannot. Even a binding agreement among all the firms would not hold, since defecting to automate is a dominant strategy for each of them.

    The Red Queen result is the part that should give AI optimists pause. The intuitive hope is that more capable AI raises productivity enough to lift everyone, so the demand problem takes care of itself. The model says the opposite. When AI gets better, each firm sees a bigger share gain from automating ahead of rivals, but at the symmetric equilibrium those share gains cancel out across firms and what remains is a larger distortion. Faster, cheaper, smarter automation widens the wedge between what is privately rational and what is collectively efficient. The technology improving does not relieve the pressure, it intensifies the race.

    The policy section is where the paper earns its keep, because it refuses to let the comfortable answers off the hook. Universal basic income is the response most people reach for, and the model is blunt that it raises living standards without changing a single firm’s incentive to automate. It treats the symptom and ignores the margin. Upskilling and worker equity narrow the gap but cannot close it. Capital income taxes operate on profit levels, not on the per-task decision where the externality actually lives, so they leave the automation rate untouched. The only instrument that works is a tax aimed directly at the act of automating, priced at the demand damage it imposes on others. That is an uncomfortable conclusion for almost everyone. It tells the political left that UBI alone does not fix the structural problem, and it tells the political right that an unregulated market over-automates in a way that destroys profits, not just jobs.

    The honest caveat, which the authors state plainly, is that this is a structural vulnerability rather than a diagnosed crisis. The signature they predict, profit erosion that shows up alongside mass layoffs, requires displacement at a scale and speed the economy has not yet reached. If reabsorption keeps pace, the externality stays too small to measure. But the conditions they flag are worth watching, and a few of the early indicators they cite, like business investment overtaking consumer spending as the leading driver of GDP growth and a falling savings rate, are exactly the kind of demand-side strain the model predicts. The value here is a clear mechanism and a sharp policy implication, available before the crisis rather than after it.

    Key Takeaways

    • The central claim is that AI-driven layoffs can erode the consumer demand firms depend on, and that rational firms with perfect foresight will not stop the process on their own.
    • The mechanism is a demand externality. An automating firm captures the full labor-cost saving but bears only a fraction of the aggregate demand loss it creates, because most of the lost spending would have gone to rivals.
    • Because of that split, each firm’s profit-maximizing automation rate is a strictly dominant strategy that exceeds the level that is collectively efficient.
    • The resulting loss is not a transfer from workers to owners. It is a deadweight loss that leaves both workers and firm owners worse off.
    • The distortion deepens with competition. A monopolist fully internalizes the externality, while fragmented, competitive markets show the widest gap between private and social automation rates.
    • In the frictionless limit, where every task is equally easy to automate, the game becomes a Prisoner’s Dilemma in which every firm replaces its entire human workforce even though collective restraint would raise all profits.
    • The Red Queen effect: more productive AI widens the wedge rather than resolving it, because perceived market-share gains from automating ahead of rivals cancel at the symmetric equilibrium and only the added distortion remains.
    • Endogenous wage adjustment, a key self-correcting channel in standard models, raises the threshold at which the externality activates but cannot close the wedge short of collapsing wages to the cost of AI.
    • Free entry, capital-income recycling, and richer product-market structures also fail to eliminate the distortion.
    • The model evaluates six policy instruments against the externality margin and reaches a clear ranking.
    • Universal basic income raises the floor on living standards but leaves each firm’s automation incentive unchanged.
    • Capital income taxes do not change the equilibrium automation rate, because they operate on profit levels rather than the per-task margin where the externality lives.
    • Upskilling and worker equity participation narrow the wedge but cannot eliminate it.
    • Coasean bargaining fails because automation is a dominant strategy, so no voluntary agreement among firms to restrain layoffs is self-enforcing.
    • Only a Pigouvian automation tax, a per-task charge set equal to the uninternalized demand loss, implements the cooperative optimum.
    • The tax can be self-limiting. Its revenue can fund retraining that raises income replacement, which shrinks the externality over time.
    • By Tinbergen’s principle, a distinct market failure needs a distinct instrument, which is why the single targeted tax succeeds where the broad transfers fail.
    • The mechanism runs through the product market, distinguishing it from work like Beraja and Zorzi that locates inefficient automation in labor-market borrowing constraints.
    • Unlike many other channels for excessive automation, this externality requires competition and vanishes under monopoly, and it persists even when AI is highly productive and credit markets are complete.
    • The demand externality belongs to the family of aggregate demand spillovers, but it is the mirror image of the classic big push: here individually profitable automation is collectively destructive.
    • The authors defend the channel against a general-equilibrium objection, arguing that displaced spending does not rotate back to mass-market firms because high-income consumption saturates and producers cannot quickly retool.
    • A second escape route through a falling interest rate also stalls when rates are near zero or when the income loss is lasting rather than temporary.
    • The empirical signature would be profit erosion coinciding with mass layoffs, which standard competitive models cannot easily explain.
    • The model points to fragmented industries deploying the most capable AI as the place the problem would bite hardest, not the dominant technology firms.
    • Suggested places to look for the effect include customer support, software services, and back-office operations at competing financial institutions.
    • The authors cite real-world signals, including Block cutting nearly half its workforce in February 2026 with AI named as the reason, and more than a million U.S. job cuts announced in 2025 with AI explicitly tied to roughly 55,000.
    • They note that roughly 80% of U.S. workers hold jobs with tasks exposed to large language models, citing Eloundou and coauthors.
    • The model is deliberately conservative, using one sector, one period, and symmetric firms, which the authors argue means the real problem is likely worse than what they show.
    • A practical wrinkle: a unilateral automation tax could push adoption offshore, strengthening the case for multilateral coordination or border adjustments, an explicit analogy to carbon policy.
    • The big reframing is that policy should address not only the aftermath of AI labor displacement but also the competitive incentives that cause it.

    Detailed Summary

    A task-based model refocused on the product market

    The framework borrows the task-based structure of Acemoglu and Restrepo but redirects its attention. Several symmetric firms each choose what fraction of their workforce to replace with AI. Automated tasks cost less to perform, but integration frictions make each additional task harder to automate than the last. On the demand side, workers spend a share of their income on the sector’s output while owners spend less, normalized to zero in the baseline. Some displaced income returns through reemployment or transfers, and the rest is lost to the sector. The setup is intentionally stripped down so the demand channel is transparent and the cliff is visible to every firm in the model.

    The demand externality that traps every firm

    Competition creates the trap. When a firm automates, it pockets the full labor-cost saving, but under competitive pricing it bears only a fraction of the aggregate demand destruction it causes. The rest spills onto rivals. Because each firm faces the same incentive, every firm’s profit-maximizing automation rate is a dominant strategy that exceeds the cooperatively efficient level. Foresight does not save them. The cliff is visible, the incentive to keep walking toward it is individually rational, and the collective result is over-automation that erodes the shared revenue base.

    Competition deepens it, monopoly internalizes it

    The size of the distortion depends on market structure. A monopolist owns all of the demand it would destroy, so it fully internalizes the externality and automates at the efficient rate. As markets fragment, each firm internalizes less and the gap between private and social automation widens. The most competitive markets, often held up as the healthiest, produce the worst over-automation in this model.

    The frictionless limit becomes a Prisoner’s Dilemma

    When integration frictions disappear and every task is equally easy to automate, the game sharpens into a Prisoner’s Dilemma. Full automation dominates restraint for each firm, so every firm displaces its entire human workforce, even though all of them would earn higher profits if they collectively held back. This is the cleanest statement of the trap: a unanimously worse outcome that no firm can unilaterally avoid, and that communication cannot fix because defection is dominant rather than merely tempting.

    The Red Queen effect: better AI makes it worse

    Higher AI productivity does not rescue the equilibrium. Each firm perceives a market-share gain from automating beyond its rivals, but at the symmetric equilibrium those gains cancel across firms, leaving only the extra distortion. So improvements in AI widen the wedge instead of closing it. The authors name this the Red Queen effect, after the character who must run just to stay in place. Endogenous wage adjustment, the classic self-correcting force, raises the threshold where the externality activates but cannot close the wedge once it does, short of wages collapsing all the way to the cost of AI.

    Six policy fixes, and why only one works

    The paper lines up six instruments against the externality. Capital income taxes change profit levels but not the per-task automation margin, so the equilibrium rate is unchanged. Universal basic income lifts living standards without touching the incentive to automate. Upskilling and worker equity narrow the wedge but leave a gap. Coasean bargaining cannot hold because automating is a dominant strategy, so no agreement is self-enforcing. Only a Pigouvian automation tax, set equal to the uninternalized demand loss per task, implements the cooperative optimum. Its revenue can fund retraining that raises income replacement, which shrinks the externality over time and can make the tax self-limiting. Tinbergen’s principle frames the lesson: a distinct market failure needs its own dedicated instrument.

    Does the channel survive general equilibrium?

    A natural objection is that in a frictionless multi-sector economy, displaced income would simply rotate to other spending and the mechanism would dissolve. The authors argue both escape routes are blocked for the mass-market firms most exposed to AI. Spending does not rotate back because high-income consumption saturates and mass-sector producers cannot quickly retool to capture redirected luxury demand. The other route runs through the interest rate: automation shifts income to owners who save more, raising aggregate saving, which a falling interest rate would normally recycle into investment. That adjustment stalls when rates are already near zero or when the income loss is lasting rather than temporary, so displaced workers cannot borrow their way through it.

    What to watch for in the real economy

    The distinguishing empirical signature would be profit erosion that shows up at the same time as mass layoffs, a combination standard competitive models struggle to explain since cost-cutting technology is supposed to raise profits. The authors are careful that this requires displacement at a scale and speed not yet reached, so the contribution is identifying a structural vulnerability rather than diagnosing an active crisis. They point to fragmented industries running the most capable AI as the place to look first, naming customer support, software services, and competing financial institutions’ back-office operations as concrete settings. They also flag a unilateral tax’s offshoring risk, drawing an explicit parallel to carbon policy and the case for multilateral coordination or border adjustments.

    Notable Quotes

    “At the limit, this becomes self-destructive: firms automate their way to boundless productivity and zero demand.”

    The authors, framing the demand cliff that competitive automation runs toward.

    “Rational, forward-looking firms should be the brake; if the cliff ahead is visible to all, why would they race toward it?”

    The authors, setting up the puzzle the paper exists to answer.

    “No firm can afford to be the one that holds back. This is the trap: an automation arms race that only intensifies as AI improves, that leaves workers and firm owners alike worse off, and that no market force can break.”

    From the Discussion, stating the core result in plain language.

    “Because over-automation leaves both firms and workers worse off, correcting it is a matter of eliminating waste, not of redistributing gains between them.”

    The authors, on why the fix is not a left-versus-right transfer fight.

    “This Red Queen effect means that ‘better’ AI, far from mitigating the externality, amplifies it.”

    The authors, on why more capable AI deepens the distortion rather than curing it.

    “The results suggest that policy should address not only the aftermath of AI labor displacement but also the competitive incentives that drive it.”

    From the abstract, the paper’s central policy reframing.

    You can read the full paper, including the formal propositions and the policy table, on arXiv here.

    Related Reading

  • Claude Fable 5 and Claude Mythos 5: Anthropic Ships Its First Generally Available Mythos-Class AI Model With New Safeguards

    Anthropic has launched Claude Fable 5 and Claude Mythos 5, the first Mythos-class models offered beyond a tiny circle of cyber defenders. Fable 5 is the generally available version, wrapped in a new layer of safeguards, while Mythos 5 is the same underlying model with some of those guardrails lifted for a small group of vetted partners. The pair sits a full tier above the Opus class in raw capability, and the launch is as much a story about how Anthropic is choosing to gate that capability as it is about the benchmarks. Below is a full breakdown of what shipped, what the model can do, and why the safeguard design matters.

    TLDR

    Anthropic released Claude Fable 5, a Mythos-class model that is now its most capable generally available model, posting state-of-the-art results across software engineering, knowledge work, vision, memory, and scientific research. To ship it safely and fast, Fable 5 carries new safety classifiers that route flagged queries in cybersecurity, biology and chemistry, and distillation over to Claude Opus 4.8 instead of refusing, a fallback that triggers in under 5% of sessions. The same model ships without cyber safeguards as Claude Mythos 5 for Project Glasswing partners in collaboration with the US Government, where it is described as having the strongest cybersecurity capabilities of any model in the world. Highlights include a codebase-wide migration of a 50-million-line Ruby codebase that Stripe says took a day instead of two months, beating Pokemon FireRed with a vision-only harness, accelerating drug design roughly tenfold using Mythos 5, producing novel molecular biology hypotheses preferred by scientists about 80% of the time, and over a week of autonomous genomics research. Both models cost 10 dollars per million input tokens and 50 dollars per million output tokens, less than half the price of Mythos Preview, with a staged subscription rollout and a new 30-day data retention policy for Mythos-class traffic.

    Thoughts

    The most interesting decision here is not the capability jump, it is the naming split. Fable and Mythos are the same brain. The only difference is whether the safeguards are on. Anthropic is effectively shipping one model twice: a gated public edition and an ungated edition handed to a short list of trusted defenders working with the US Government. That is a clean way to resolve the central tension of frontier AI, which is that the exact capabilities that help a security professional close a vulnerability also help an attacker find one. Rather than dumbing the model down for everyone or holding it back entirely, they are letting the access list, not the weights, carry the risk. Expect this pattern to repeat as capabilities climb.

    The fallback-to-Opus design is the other quietly important choice. When a classifier flags a query in cybersecurity, biology, chemistry, or suspected distillation, the user does not hit a wall of refusal. The request is silently handed to Opus 4.8, a model that is still excellent at almost everything. Graceful degradation beats a hard no, both for user experience and for trust. It also reframes what a safeguard is. Instead of a binary block, it becomes a routing decision, and because more than 95% of sessions never trigger it, most users will never notice it exists. The honest admission that the classifiers are tuned conservatively and will sometimes catch harmless requests is the right posture, even if it will annoy power users who keep getting bounced to the smaller model.

    The commercial signals are worth reading closely. Pricing came down to less than half of Mythos Preview, which suggests confidence in serving costs at scale, but the subscription rollout tells a more cautious story. Fable 5 is free on Pro, Max, Team, and Enterprise plans only through June 22, after which using it requires usage credits until capacity catches up. That is a polite way of saying demand is expected to badly outrun supply. The model is fully available on the API and consumption-based Enterprise plans from day one, because those bill by the token and self-throttle. Subscriptions, which are all-you-can-eat, are where a capacity crunch actually hurts, so that is exactly where the brakes went on.

    On the science, the genomics result is the one that should make people sit up. A model doing over a week of largely autonomous research, assembling single-cell data across 138 species, then designing and training its own machine learning model that outperforms a recently published Science paper while being 100 times smaller, is a different category of claim than acing a benchmark. So is the drug-design work, where Mythos 5 reportedly matches or beats skilled human operators end to end, choosing binding sites, running protein design tools, and recovering from its own failures. If those hold up to publication and independent replication, the interesting frontier stops being chat quality and becomes whether a model can run a research program. That is also precisely why the biology and chemistry classifier exists, and why Anthropic is being so deliberate about who gets the ungated version.

    One caveat worth keeping in view: nearly all of the evidence in the announcement is Anthropic’s own, or comes from partners with early access and an incentive to be enthusiastic. The Stripe migration, the FrontierCode score, the Slay the Spire memory result, the protein targets, and the genomics model are all compelling, but they are first-party until outside labs and the eventual system card, peer review, and independent red-teamers weigh in. The note that the UK AISI made progress toward a universal jailbreak inside a brief testing window is a useful reminder that the safeguard story is a work in progress, not a finished proof.

    Key Takeaways

    • Claude Fable 5 is a Mythos-class model made safe for general use, and is now Anthropic’s most capable generally available model.
    • Mythos-class is a tier that sits above the Opus class in capability. The first was Claude Mythos Preview, released in April through Project Glasswing.
    • Fable 5 is state-of-the-art on nearly all tested benchmarks, and its lead grows as tasks get longer and more complex.
    • Claude Mythos 5 is the same underlying model as Fable 5, but with safeguards lifted in some areas. Fable and Mythos differ only by their safeguards.
    • Mythos 5 is described as having the strongest cybersecurity capabilities of any model in the world, and is deployed through Project Glasswing with the US Government.
    • New safety classifiers cover cybersecurity, biology and chemistry, and distillation. Flagged queries fall back to Claude Opus 4.8 rather than being refused.
    • Users are told whenever a fallback happens. More than 95% of Fable sessions involve no fallback at all, and for those sessions Fable performs effectively the same as Mythos 5.
    • The safeguards are tuned conservatively and trigger in less than 5% of sessions on average, sometimes catching harmless requests. Anthropic plans to reduce false positives after launch.
    • Stripe reported Fable 5 compressed months of engineering into days, performing a codebase-wide migration of a 50-million-line Ruby codebase in a day that would have taken a team over two months by hand.
    • Fable 5 scores highest among frontier models on Cognition’s FrontierCode evaluation for high-quality agentic coding, even at medium effort, and is more token-efficient than past Claude models.
    • On Hebbia’s Finance Benchmark for senior-level reasoning, Fable 5 has the highest score of any model, with gains in document reasoning, chart and table interpretation, and problem solving.
    • IMC noted Fable 5 aced their trading-analysis evaluations nearly across the board, including factual lookup, conceptual reasoning, root-cause analysis, and expected-value analysis.
    • Fable 5 is the new state-of-the-art for vision, and can rebuild a web app’s source code from screenshots alone.
    • Fable 5 beat Pokemon FireRed using a minimal, vision-only harness with no maps, navigation aids, or extra game-state information. Earlier Claude models needed a complex helper harness.
    • Persistent file-based memory improved Fable 5’s Slay the Spire performance three times more than it did for Opus 4.8, and Fable reached the game’s final act three times more often.
    • Fable 5 built a simulation of the solar system, deriving the planets’ orbital motion from physics first principles and using it to predict solar eclipses.
    • Using Mythos 5, internal protein design experts accelerated aspects of drug design by around ten times, with the model matching or beating skilled human operators end to end.
    • Nine of 14 protein targets in the drug-design study yielded strong candidates Anthropic is now investigating.
    • Mythos 5 is Anthropic’s first model to consistently produce novel, compelling scientific hypotheses. Scientists preferred its molecular biology hypotheses about 80% of the time in blinded comparisons.
    • One Mythos hypothesis, a novel mechanism for an E. coli protein, was corroborated by an independent lab working on the same problem.
    • In over a week of largely autonomous work, Mythos 5 assembled single-cell data for millions of cells across 138 animal species and trained a custom model that outperformed a recent Science paper while being 100 times smaller.
    • Anthropic’s automated alignment assessment found Mythos 5’s level of misaligned behavior was low and similar to Opus 4.8. Because they are the same model, Fable 5’s alignment is similar.
    • An external bug bounty produced no universal jailbreaks in over 1,000 hours of testing, though the UK AISI made progress toward one in a brief initial window.
    • One external partner found Fable 5’s safeguards against harmful cyber queries the most robust of any model tested, including Opus 4.8 and Opus 4.7, with zero compliance on harmful single-turn cyberattack requests.
    • The biology and chemistry classifier is deliberately broad for now. Mythos-class models outperformed dedicated protein language models at predicting AAV viral shell assembly using biological reasoning alone.
    • The distillation classifier targets large-scale attempts to extract Claude’s capabilities to train competing models, which could proliferate near-frontier capabilities without safeguards.
    • A new policy requires 30-day data retention for all Mythos-class traffic on first- and third-party surfaces, used only for safety, with logged human access and deletion after 30 days in almost all cases.
    • Anthropic plans trusted access programs that let cybersecurity organizations apply for Mythos 5, and let a small number of life science researchers access Fable 5 with biology and chemistry safeguards removed.
    • Both models cost 10 dollars per million input tokens and 50 dollars per million output tokens, less than half the price of Mythos Preview. Developers can use claude-fable-5 via the Claude API.
    • Fable 5 is free on Pro, Max, Team, and seat-based Enterprise plans through June 22. On June 23 it moves to usage credits on those plans until capacity allows it to return as a standard inclusion.

    Detailed Summary

    A Mythos-class model, made safe for general use

    Fable 5 is the first Mythos-class model Anthropic has made generally available. Mythos-class is a tier that sits above the Opus class, and the first of its kind, Claude Mythos Preview, was released in April through Project Glasswing to a limited group of cyber defenders and critical software infrastructure providers. The company framed today’s launch as the moment it could finally bring that level of capability to all users, because its safeguards had matured enough to allow it. Fable 5’s capabilities exceed those of any model Anthropic has made generally available, and its advantage over other models grows as tasks get longer and more complex.

    Two models, one brain

    Claude Mythos 5 is the same underlying model as Fable 5, but with safeguards lifted in some areas. The names are the only real difference: Fable, from the Latin fabula meaning that which is told, is akin to the Greek mythos, and the safeguards are what distinguish the two. Mythos 5 launches first to existing Mythos Preview users, including the Project Glasswing cybersecurity partners, as an upgrade. It is deployed in collaboration with the US Government and is described as having the strongest cybersecurity capabilities of any model in the world. Anthropic plans to steadily expand access through a more systematic trusted access program.

    Software engineering and token efficiency

    Fable 5 can work autonomously for longer than any previous Claude model, and software engineering is where that shows most clearly. During early testing, Stripe reported it compressed months of engineering into days, performing a codebase-wide migration in a 50-million-line Ruby codebase in a single day that would otherwise have taken a whole team over two months by hand. It is also more token-efficient than past models, scoring highest among frontier models on Cognition’s FrontierCode evaluation for high-quality, maintainable agentic coding, even at medium effort.

    Knowledge work, vision, and memory

    On complex analytical work, Fable 5 posted the highest score of any model on Hebbia’s Finance Benchmark for senior-level reasoning, with substantial gains in document-based reasoning and chart and table interpretation, and IMC said it aced their trading-analysis evaluations nearly across the board. In vision, it is the new state-of-the-art, able to extract precise numbers from detailed scientific figures and rebuild a web app’s source code from screenshots alone. It needs less scaffolding too: where earlier Claude models struggled to play Pokemon even with helper harnesses, Fable 5 beat FireRed with a minimal, vision-only harness using nothing but raw game screenshots. On memory, giving Fable persistent file-based notes improved its Slay the Spire performance three times more than it did for Opus 4.8, and it built a physics-first-principles solar system simulation accurate enough to predict solar eclipses.

    Life sciences: drug design, hypotheses, and genomics

    Using Mythos 5, Anthropic’s internal protein design experts accelerated aspects of the drug-design process by around ten times. With protein design and bioinformatics tools but no human assistance, the model matched or beat skilled human operators, executing the full workflow of choosing binding sites, selecting and running design tools, and recovering from failures. Nine of 14 protein targets yielded strong drug-design candidates now under investigation. Mythos 5 is also Anthropic’s first model to consistently produce novel, compelling scientific hypotheses: scientists preferred its molecular biology hypotheses about 80% of the time in blinded comparisons, and one, a novel mechanism for an E. coli protein, was corroborated by an independent lab. In genomics, Mythos 5 ran over a week of largely autonomous research, assembling single-cell data for millions of cells across 138 species and training a custom model that outperformed a recent Science paper despite being 100 times smaller.

    The new safeguards: classifiers and fallback

    Mythos-class capability is potent enough that Anthropic considers it a substantial misuse risk, especially given how much advanced AI usage is dual use. Fable 5 ships with a new set of classifiers, separate AI systems that detect potential misuse and jailbreak attempts and stop the main model from responding. When a classifier flags a request related to cybersecurity, biology and chemistry, or distillation, the response is handled by Claude Opus 4.8 instead, and the user is told. The cybersecurity classifiers cover both exploitation and broader offensive cyber tasks like reconnaissance and lateral movement, and Anthropic says they prevent Fable from making any progress on those tasks. The biology and chemistry classifier is intentionally broad for now, after tests showed Mythos-class models could outperform dedicated protein language models at predicting AAV viral shell assembly using biological reasoning alone. The distillation classifier targets large-scale attempts to extract Claude’s capabilities to train competing models.

    Jailbreak resistance, data retention, and availability

    Anthropic ran extensive red-teaming, including an external bug bounty that produced no universal jailbreaks in over 1,000 hours, though it notes the UK AISI made progress toward one in a brief window. The company concedes it is likely impossible to fully prevent universal jailbreaks and aims instead to make any that remain slow and costly enough to catch before they scale. A new policy requires 30-day data retention for all Mythos-class traffic, used only for safety, with logged human access and deletion after 30 days in almost all cases. On availability, Fable 5 is live everywhere today and fully available on the API and consumption-based Enterprise plans, while subscription access rolls out in stages: free on Pro, Max, Team, and seat-based Enterprise through June 22, then on usage credits from June 23 until capacity allows it to return as a standard inclusion. Both models cost 10 dollars per million input tokens and 50 dollars per million output tokens.

    Notable Quotes

    “Today we’re launching Claude Fable 5: a Mythos-class model that we’ve made safe for general use.”

    Anthropic, opening the Claude Fable 5 and Claude Mythos 5 announcement

    “Fable 5’s capabilities exceed those of any model we’ve ever made generally available.”

    Anthropic, on where Fable 5 sits in the lineup

    “It has the strongest cybersecurity capabilities of any model in the world.”

    Anthropic, describing Claude Mythos 5

    “During early testing, Stripe reported that Fable 5 compressed months of engineering into days.”

    Anthropic, on Fable 5’s software engineering results

    “Our early data shows that more than 95% of Fable sessions involve no fallback at all.”

    Anthropic, on how often the safeguards route to Opus 4.8

    “Mythos 5 is our first model to consistently produce novel, compelling scientific hypotheses.”

    Anthropic, on the model’s molecular biology research

    “It is likely impossible to completely prevent universal jailbreaks, but our goal is to make any remaining jailbreaks sufficiently slow and costly that we can detect and prevent them before they are used at scale.”

    Anthropic, on the limits of its safeguards

    “Fable is from the Latin fabula, ‘that which is told,’ akin to the Greek mythos. The safeguards are what distinguish the two models.”

    Anthropic, explaining the Fable and Mythos naming

    Read the full announcement and the benchmark tables on Anthropic’s site here: Claude Fable 5 and Claude Mythos 5.

    Related Reading

  • Anthropic Raises $65 Billion Series H at $965 Billion Valuation to Fund AI Safety Research and Massive Compute Expansion

    Anthropic has closed one of the largest private financing rounds in the history of technology, raising $65 billion in Series H funding at a $965 billion post-money valuation. The round, announced on May 28, 2026, lands as demand for Claude reaches what the company calls historic levels, and it positions Anthropic to pour fresh capital into safety research, compute, and the products that enterprises now lean on every day.

    TLDR

    Anthropic raised $65 billion in its Series H at a $965 billion post-money valuation, with Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital leading and Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ, and XN co-leading, alongside $15 billion in previously committed hyperscaler investment that includes $5 billion from Amazon. The raise follows Anthropic crossing $47 billion in run-rate revenue earlier in May 2026, and it funds three priorities named by CFO Krishna Rao: advancing safety and interpretability research, expanding compute capacity to meet growing Claude demand, and scaling the products and partnerships customers depend on. On the infrastructure side, the company is locking in gigawatt-scale compute through 5 gigawatts with Amazon, 5 gigawatts of TPU capacity via Google and Broadcom, GPU access from SpaceX, and supply from partners Micron, Samsung, and SK hynix, while Claude remains available across all three major cloud platforms, AWS, Google Cloud, and Microsoft Azure, with widespread enterprise adoption across industries.

    Thoughts

    Start with the number that everyone will fixate on. A $965 billion post-money valuation against $47 billion in run-rate revenue is roughly 20 times sales, and for a company growing this fast that multiple is not the interesting part. The interesting part is that run-rate revenue crossed $47 billion earlier this month, which means the denominator is moving so quickly that the multiple is already stale. Investors are not pricing the business Anthropic is today. They are pricing the slope. A 20x multiple on a number that may double again inside a year is a very different bet than 20x on a flat line, and the lead names here (Altimeter, Dragoneer, Greenoaks, Sequoia, with Capital Group, Coatue, GIC and others co-leading) are not the kind of capital that pays for nostalgia. They are paying for the second derivative.

    But the real story is not the valuation. It is the compute. Read the infrastructure list carefully and you see the actual problem this round solves: 5 gigawatts from Amazon, 5 gigawatts of TPU capacity through Google and Broadcom, GPU access from SpaceX, and memory supply locked down with Micron, Samsung, and SK hynix. That is more than 10 gigawatts of secured power and silicon. The constraint on frontier AI in 2026 is no longer talent or even algorithms. It is electricity, land, and the multi-year queue for advanced packaging and high-bandwidth memory. You cannot buy 10 gigawatts on a quarterly basis. You reserve it years out, and you need the balance sheet to make those commitments credible. A $65 billion raise is, in plain terms, the down payment that lets Anthropic sign for capacity nobody can conjure on demand. The money is downstream of the megawatts.

    The diversification across that compute stack matters as much as the size. By splitting between Amazon’s infrastructure, Google and Broadcom’s custom TPUs, and SpaceX-supplied GPUs, Anthropic is refusing to become hostage to any single supplier’s roadmap or pricing. Custom silicon through Broadcom in particular is a bet on bending the cost curve, because the long-term economics of serving Claude at this scale depend on dollars per token, not just on raw availability. Anyone who has watched cloud lock-in play out over the last decade understands the move. Optionality at the hardware layer is leverage, and leverage is what keeps margins from being dictated by whoever owns the only fab slot you can reach.

    It is worth pausing on the fact that the round explicitly funds safety and interpretability research alongside scaling, and not as a footnote. Most companies treat safety spend as a cost center to be minimized once growth kicks in. Naming it first, ahead of compute and products, is a statement about where Anthropic believes its durable advantage sits. If models keep getting more capable, the binding constraint on deployment inside regulated industries (finance, healthcare, government) becomes trust, not intelligence. Interpretability is the work that turns a black box into something an enterprise risk committee can actually sign off on. Framed that way, safety research is not philanthropy subtracted from the bottom line. It is the thing that unlocks the most lucrative and defensible parts of the market, and pairing it with the scaling budget is the tell.

    Finally, look at distribution. Claude now ships on all three major clouds at once: AWS, Google Cloud, and Microsoft Azure. In a market where most frontier labs are tethered to a single hyperscaler, being available everywhere enterprises already run their workloads is a structural edge. It removes the procurement friction of asking a customer to adopt a new vendor relationship, and it means Anthropic competes on the merits of the model rather than on which cloud a buyer happened to standardize on years ago. Combine that omnipresent distribution with the compute reservations and the explicit safety mandate, and the shape of the strategy is clear. This is not a company buying time. It is a company buying the three things that actually compound: capacity that cannot be rushed, trust that cannot be faked, and reach into every place where work already happens.

    Key Takeaways

    • Anthropic raised $65 billion in its Series H funding round, one of the largest private financings in the history of the technology industry.
    • The round set Anthropic’s post-money valuation at $965 billion, placing the company within reach of the $1 trillion mark.
    • Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital led the Series H round.
    • Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ, and XN served as co-leads on the investment.
    • The new capital builds on $15 billion in previously committed hyperscaler investments, which includes $5 billion from Amazon.
    • Anthropic crossed $47 billion in run-rate revenue earlier in May 2026, reflecting the surging commercial demand for Claude.
    • A core priority for the funding is to advance Anthropic’s safety and interpretability research.
    • The company will use the capital to expand compute capacity in order to meet growing demand for Claude.
    • Anthropic plans to scale the products and partnerships that customers depend on across its business.
    • CFO Krishna Rao said the funding will help Anthropic serve the historic demand it is experiencing, stay at the research frontier, and bring Claude to more of the places where work happens.
    • Amazon is providing 5 gigawatts of compute capacity as part of Anthropic’s infrastructure expansion.
    • Google and Broadcom are supplying 5 gigawatts of TPU capacity to power Claude’s growth.
    • SpaceX is contributing GPU access to Anthropic’s compute footprint.
    • Micron, Samsung, and SK hynix are partnering with Anthropic on memory and infrastructure to support its scaling needs.
    • Claude is available on all three major cloud platforms, AWS, Google Cloud, and Microsoft Azure.
    • Anthropic reports widespread enterprise adoption of Claude across a broad range of industries.

    Detailed Summary

    The Raise and the Valuation

    Anthropic has raised $65 billion in Series H funding, a round that values the company at $965 billion on a post-money basis. The size of the raise places it among the largest private financing events the technology industry has ever seen, and the valuation pushes Anthropic to the doorstep of the trillion dollar mark. The capital arrives at a moment when demand for the company’s Claude models has accelerated sharply, and the round is built to fund the response to that demand rather than simply mark a milestone. Anthropic framed the financing in its Series H announcement as the fuel for staying at the research frontier while scaling the infrastructure and products that customers increasingly rely on.

    Who Put In the Money

    The Series H was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, a group that combines deep growth-stage technology experience with conviction in Anthropic’s long-term trajectory. Joining as co-leads were Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ, and XN, a roster that spans crossover funds, sovereign wealth, and institutional investors. Beyond the new equity, Anthropic pointed to $15 billion in previously committed hyperscaler investment, including $5 billion from Amazon. Taken together, the investor base reflects a mix of financial backers and strategic partners with a direct stake in seeing Claude reach more customers and more compute.

    Revenue at $47 Billion Run-Rate

    Underpinning the valuation is a business that has scaled with unusual speed. Anthropic crossed a $47 billion run-rate revenue figure earlier in May 2026, a number that signals how quickly enterprises and developers have adopted Claude across their workflows. Run-rate revenue annualizes the company’s most recent performance, and at this level it puts Anthropic firmly among the fastest growing software businesses on record. That financial momentum is the practical justification for both the round’s size and the near trillion dollar valuation investors were willing to support.

    The Compute Buildout

    A large share of the strategy behind the raise centers on securing compute at enormous scale. Anthropic detailed a set of infrastructure partnerships designed to keep pace with Claude demand. Amazon is providing 5 gigawatts of capacity, while Google and Broadcom together are supplying 5 gigawatts of TPU capacity. SpaceX is contributing GPU access, broadening the range of silicon Anthropic can draw on. Supporting the buildout on the hardware supply side are Micron, Samsung, and SK hynix, the memory and component partners whose output is essential to standing up data centers at this magnitude. The combined picture is a company assembling power, chips, and supply chain commitments measured in gigawatts rather than racks.

    Where the Money Goes

    Anthropic outlined three priorities for the new capital. The first is to advance safety and interpretability research, continuing the work of understanding how models behave and ensuring they remain reliable as they grow more capable. The second is to expand compute capacity to meet the growing demand for Claude, the practical engine behind the infrastructure commitments above. The third is to scale the products and partnerships that customers depend on, deepening the company’s reach into the tools and platforms where work actually happens. Krishna Rao, Anthropic’s chief financial officer, said the funding “will help us serve the historic demand we are experiencing, stay at the research frontier, and bring Claude to more of the places where work happens.”

    Claude Everywhere

    The funding lands on top of a distribution footprint that already spans the major cloud ecosystems. Claude is available on all three leading cloud platforms, AWS, Google Cloud, and Microsoft Azure, which means enterprises can reach the models through whichever provider they have standardized on. That availability has translated into widespread enterprise adoption across industries, from software and finance to healthcare and beyond. By being present everywhere developers and businesses already operate, Anthropic positions Claude not as a destination customers must travel to but as a capability woven into the platforms they use every day.

    Notable Quotes

    This funding will help us serve the historic demand we are experiencing, stay at the research frontier, and bring Claude to more of the places where work happens.

    Krishna Rao, CFO at Anthropic, on the purpose of the Series H round.

    Advance safety and interpretability research, expand compute capacity to meet growing Claude demand, and scale products and partnerships customers depend on.

    How Anthropic describes its use of funds from the round.

    For the full details on the round, the lead and co-lead investors, and how Anthropic plans to deploy the capital across safety research, compute, and products, read the full announcement here.

    Related Reading

    • Anthropic, the AI safety and research company behind Claude that raised this Series H round.
    • Sequoia Capital, one of the lead investors anchoring the financing.
    • Amazon Web Services, one of the three major cloud platforms where Claude is available and the source of a $5 billion investment.
    • Google Cloud TPUs, the tensor processing units behind the 5 gigawatts of TPU capacity in the Google and Broadcom partnership.
    • AI safety, the research field at the center of how Anthropic says it will use the new funding.

  • Claude Opus 4.8 Released: Anthropic Bets on Honesty, Dynamic Workflows, Effort Control, and Cheaper Fast Mode

    Anthropic has released Claude Opus 4.8, the newest member of its flagship Opus class, available today across every surface and priced exactly like the model it replaces. The company calls it “a modest but tangible improvement” on Opus 4.7, but the framing undersells what is actually interesting here: the headline upgrade is not a benchmark number, it is honesty. Opus 4.8 is built to know when it does not know, and that single behavioral shift may matter more for real agent work than any raw capability bump.

    TLDR

    Claude Opus 4.8 is an across-the-board upgrade to Anthropic’s Opus class that ships today at the same regular price as Opus 4.7 ($5 per million input tokens, $25 per million output tokens), with the model positioned as “a more effective collaborator.” The marquee improvement is honesty: Opus 4.8 is roughly four times less likely than its predecessor to let flaws in its own code pass unremarked, and it is more willing to flag uncertainty rather than confidently claim progress on thin evidence. A pre-release alignment assessment found new highs on prosocial traits like supporting user autonomy and acting in the user’s best interest, with misaligned behavior at rates similar to Anthropic’s best-aligned model, Claude Mythos Preview. Three things launch alongside the model: dynamic workflows in Claude Code (research preview), where Claude plans work then runs hundreds of parallel subagents that run even longer and verify their own outputs before reporting back; effort control in claude.ai and Cowork, a slider for how hard Claude thinks; and a Messages API update that accepts system entries inside the messages array so developers can update instructions mid-task without breaking the prompt cache. Fast mode now runs at 2.5x speed and is three times cheaper than before ($10 / $50 per million tokens). The roadmap points to cheaper Opus-equivalent models, a higher-intelligence class above Opus, and a wider rollout of Mythos-class models gated behind stronger cyber safeguards under Project Glasswing.

    Thoughts

    The most important sentence in this announcement is not about coding scores. It is the claim that Opus 4.8 is about four times less likely than Opus 4.7 to let flaws in its own code slip by without comment. For a chat assistant, overconfidence is annoying. For an agent, it is catastrophic. The whole premise of long-running autonomous work is that you hand the model a task and walk away, which means the model’s own judgment about whether it succeeded becomes the only judgment in the loop until you come back. A model that confidently declares victory on a half-finished migration does not save you time, it costs you a debugging session plus the time you spent trusting it. Honesty, framed this way, is not a soft virtue. It is the load-bearing reliability property that makes unattended agents usable at all.

    Read the launch as a single coherent argument rather than a list of features, and the pieces lock together. Dynamic workflows let Claude plan a job and fan out hundreds of parallel subagents that, with Opus 4.8, run longer than before. Effort control lets you dial up how much the model thinks. The honesty improvement means the model checks its own work and flags what it is unsure about instead of papering over it. Put those three together and you get one product thesis: let it run longer, let it think harder, and trust it to tell you when something is wrong. The codebase-scale migration example, hundreds of thousands of lines from kickoff to merge with the existing test suite as the bar, is the proof point. None of those three capabilities is worth much alone. A model that runs for hours but lies about its results is a liability. A model that flags uncertainty but cannot sustain a long task never reaches the moment where its honesty matters. Anthropic shipped all three at once because they only pay off together.

    The economics deserve a closer look than the “same price” headline invites. Regular pricing is flat versus Opus 4.7, which is the polite way of saying you get a better model for free. The real move is fast mode: 2.5x the speed at three times cheaper than it cost on previous models, landing at $10 per million input and $50 per million output. That is Anthropic quietly attacking the latency-versus-cost tradeoff that has shaped how teams deploy frontier models. Until now, “fast” meant “expensive,” so you reserved it for interactive moments and ate the wait everywhere else. Collapsing that premium changes the default. And note the subtle token story underneath: Opus 4.8 at its default high effort spends roughly the same tokens on coding as Opus 4.7’s default while performing better, so the effort slider is not a way to bleed you dry, it is an honest exposure of the quality-cost dial that was always there implicitly.

    The Messages API change is the kind of unglamorous plumbing that practitioners will appreciate immediately. Letting system entries live inside the messages array means you can update an agent’s instructions, permissions, token budget, or environment context partway through a task without smuggling the update through a fake user turn and without blowing up your prompt cache. Anyone who has built a long-running agent has hit this wall: the world changes mid-task, the agent needs new constraints, and the only clean way to inject them previously was a cache-busting hack. This is Anthropic treating agents as first-class, stateful, long-lived processes rather than oversized chat sessions. It is a small spec change with outsized implications for how you architect an agent that runs for an hour.

    Then there is the roadmap, where the most telling line is the quietest. Anthropic says a small number of organizations are already using Claude Mythos Preview for cybersecurity work under Project Glasswing, and that models of this capability level require stronger cyber safeguards before general release. Notice that they are pinning Opus 4.8’s alignment numbers to Mythos as the benchmark for “best-aligned,” while simultaneously holding Mythos back from general availability on safety grounds. That is a deliberate signal: the next class of model is good enough that they are gating it on cyber-offense risk, not on capability. For a site about the pursuit of joy, fulfillment, and purpose through AI, this is the part worth sitting with. The frontier is increasingly defined not by what the models can do, but by what their builders decide it is responsible to ship. Honesty in the small (flagging a bad line of code) and restraint in the large (holding back a cyber-capable model) are the same instinct expressed at two different scales.

    Key Takeaways

    • Claude Opus 4.8 is now available everywhere, replacing Opus 4.7 as Anthropic’s flagship Opus-class model and positioned as “a more effective collaborator.”
    • Regular usage pricing is unchanged from Opus 4.7, holding at $5 per million input tokens and $25 per million output tokens, so the capability gains come at no added cost.
    • The single most emphasized improvement is honesty, which Anthropic treats as a core trained behavior rather than a marketing flourish.
    • Evaluations show Opus 4.8 is around four times less likely than its predecessor to let flaws in its own code pass unremarked, a direct reliability win for autonomous coding.
    • Early testers report the model is more likely to flag uncertainty about its work and less likely to make unsupported claims or jump to conclusions on thin evidence.
    • A detailed alignment assessment was run before release and concluded Opus 4.8 reaches new highs on prosocial traits like supporting user autonomy and acting in the user’s best interest.
    • Misaligned behavior such as deception or cooperation with misuse is at rates substantially lower than Opus 4.7 and similar to Anthropic’s best-aligned model, Claude Mythos Preview.
    • The full alignment assessment and pre-deployment safety tests are documented in the public Claude Opus 4.8 System Card.
    • Dynamic workflows launch as a research preview inside Claude Code, letting Claude plan the work and then run hundreds of parallel subagents in a single session.
    • With Opus 4.8, those subagents can run even longer, and Claude verifies its outputs before reporting back rather than declaring success blindly.
    • Anthropic’s flagship example for dynamic workflows is a codebase-scale migration across hundreds of thousands of lines of code, from kickoff to merge, using the existing test suite as the success bar.
    • Dynamic workflows are available in Claude Code for the Enterprise, Team, and Max plans.
    • Effort control arrives in claude.ai and Cowork as a setting next to the model selector that lets users choose how much effort Claude puts into a response.
    • Higher effort makes Claude think more frequently and deeply for better answers; lower effort responds faster and consumes rate limits more slowly. Effort control is available on all plans.
    • Opus 4.8 defaults to “high” effort, judged the best overall balance of quality and user experience.
    • On coding tasks, the default effort spends a similar number of tokens as Opus 4.7’s default but delivers better performance, so quality rises without a token penalty.
    • Users can select “extra” (called “xhigh” in Claude Code) or “max” to spend more tokens for stronger results, and Anthropic recommends “extra” for difficult tasks and long-running asynchronous workflows.
    • Rate limits in Claude Code were increased to accommodate the higher token usage of the higher effort levels.
    • The Messages API now accepts system entries inside the messages array, a meaningful change for agent developers.
    • That update lets developers change Claude’s instructions mid-task, adjusting permissions, token budgets, or environment context, without breaking the prompt cache or routing through a user turn.
    • Fast mode now runs at 2.5x speed and is three times cheaper than it was for previous models, priced at $10 per million input tokens and $50 per million output tokens.
    • Developers access the model as claude-opus-4-8 through the Claude API.
    • Partner Miguel Gonzalez reports Opus 4.8 scored 84% on Online-Mind2Web, a meaningful jump over both Opus 4.7 and GPT-5.5, calling it the strongest computer-use and browser-agent model his team has tested.
    • Databricks reports that, inside Genie, Opus 4.8 reasons over unstructured content like PDFs and diagrams at 61% cheaper token cost than Opus 4.7.
    • Thomson Reuters reports Opus 4.8 is the first model to break 10% overall on the all-pass standard of its Legal Agent Benchmark, the highest score recorded there.
    • Eleven partners weighed in, including Cursor, Cognition’s Devin, Databricks Genie, Thomson Reuters CoCounsel, and Hebbia, spanning coding, legal, finance, and enterprise data work.
    • Anthropic is working on models that deliver many of the same capabilities as Opus at a lower cost.
    • The company plans to release a new class of model with even higher intelligence than Opus.
    • Under Project Glasswing, a small number of organizations are already using Claude Mythos Preview for cybersecurity work, with Mythos-class models expected to reach all customers in the coming weeks once stronger cyber safeguards are in place.

    Detailed Summary

    What Claude Opus 4.8 Is

    Claude Opus 4.8 is an upgrade to Anthropic’s Opus class of models, building on Opus 4.7 with improvements across benchmarks covering coding, agentic skills, reasoning, and practical knowledge-work tasks. Anthropic describes the result as “a more effective collaborator” while characterizing the release overall as “a modest but tangible improvement on its predecessor.” The model is available today, everywhere, and developers call it as claude-opus-4-8 via the Claude API. The announcement includes a comparison table against the predecessor and other models, though the per-cell numbers in that table are published as an image and are not reproduced here as text.

    Honesty: The Headline Improvement

    Anthropic singles out honesty as one of the most prominent improvements in Opus 4.8. All of the company’s models are trained to be honest, which includes avoiding claims they cannot support. A persistent problem with AI models generally is that they sometimes jump to conclusions, confidently claiming progress despite thin evidence. Early testers report that Opus 4.8 is more likely to flag uncertainties about its own work and less likely to make unsupported claims. The most concrete measure: evaluations show Opus 4.8 is around four times less likely than its predecessor to allow flaws in code it has written to pass unremarked. For agentic and unattended use, this self-skepticism is the difference between a model that reliably tells you when something went wrong and one that quietly ships a broken result.

    Alignment Assessment

    A detailed alignment assessment was run before release. On the positive side, the Alignment team concluded that Opus 4.8 “reaches new highs on our measures of prosocial traits like supporting user autonomy and acting in the user’s best interest.” On the risk side, misaligned behavior such as deception or cooperation with misuse occurs at rates substantially lower than Opus 4.7, and similar to Anthropic’s best-aligned model, Claude Mythos Preview. The full alignment assessment and the pre-deployment safety tests are published in the Claude Opus 4.8 System Card, which also contains the complete benchmark table and wider evaluations.

    Dynamic Workflows in Claude Code

    Launching today as a research preview in Claude Code, dynamic workflows let Claude plan the work and then run hundreds of parallel subagents in a single session. With Opus 4.8, those agents can run even longer than before, and Claude verifies its outputs before reporting back rather than reporting unchecked results. The showcase example is a codebase-scale migration: Claude Code with Opus 4.8 can carry out migrations across hundreds of thousands of lines of code, all the way from kickoff to merge, using the existing test suite as its bar for success. Dynamic workflows are available in Claude Code for the Enterprise, Team, and Max plans.

    Effort Control

    Effort control arrives in claude.ai and Cowork as a setting alongside the model selector that lets users choose how much effort Claude puts into a response. Higher effort means Claude thinks more frequently and deeply for better responses; lower effort means it responds faster and uses rate limits more slowly. Opus 4.8 defaults to “high” effort, which Anthropic judged the best overall balance of quality and user experience. On coding tasks, that default spends a similar number of tokens as Opus 4.7’s default while performing better. Users who want more can choose “extra” (called “xhigh” in Claude Code) or “max” to spend more tokens for stronger results, and Anthropic recommends “extra” for difficult tasks and long-running asynchronous workflows. To support the heavier token usage at higher effort levels, rate limits in Claude Code were increased. Effort control is available on all plans.

    Messages API Update

    The Messages API now accepts system entries inside the messages array. This lets developers update Claude’s instructions mid-task without breaking the prompt cache and without routing the update through a user turn. In practice that means you can update permissions, token budgets, or environment context while an agent is running, which is exactly the kind of statefulness a long-running autonomous process needs. It is a small specification change with significant consequences for how developers build durable agents.

    Pricing and Fast Mode

    Regular usage pricing is unchanged from Opus 4.7: $5 per million input tokens and $25 per million output tokens. The notable shift is in fast mode, where the model works at 2.5x the speed and fast mode is now three times cheaper than it was for previous models, landing at $10 per million input tokens and $50 per million output tokens. The combination of unchanged regular pricing and dramatically cheaper fast mode reshapes the latency-versus-cost calculus that has long governed how teams deploy frontier models.

    Partner Results Across Coding, Legal, Finance, and Data

    Eleven partners shared results spanning the spectrum of professional work. Miguel Gonzalez reports 84% on Online-Mind2Web, a meaningful jump over both Opus 4.7 and GPT-5.5, calling it the strongest computer-use and browser-agent model his team has tested. Databricks reports that Genie reasons over unstructured content like PDFs and diagrams at 61% cheaper token cost than Opus 4.7. Thomson Reuters reports Opus 4.8 is the first model to break 10% overall on the all-pass standard of its Legal Agent Benchmark. Cursor reports gains across every effort level on CursorBench with more efficient tool calling, and Cognition reports that Devin sees cleaner tool use, fixes to the comment-verbosity and tool-calling issues seen with Opus 4.7, and improvements over Opus 4.6. Hebbia reports strong quality with better citation precision and more token efficiency on retrieval for dense financial filings. The footnotes note that Terminal-Bench 2.1 was scored on the Terminus-2 public harness (GPT-5.5’s Codex CLI harness score is 83.4%), that OSWorld-Verified methodology changed with Opus 4.7’s score updated to 82.3%, and that on Finance Agent v2 Gemini 3.5 Flash scores 57.9%.

    What Is Next: Cheaper Models, Higher Intelligence, and Mythos

    Anthropic outlined a three-part roadmap. First, the company is working on models that provide many of the same capabilities as Opus at a lower cost. Second, it plans to release a new class of model with even higher intelligence than Opus. Third, as part of Project Glasswing, a small number of organizations are currently using Claude Mythos Preview for cybersecurity work; models of this capability level require stronger cyber safeguards before general release, and Anthropic expects to bring Mythos-class models to all customers in the coming weeks.

    Notable Quotes

    “Claude Opus 4.8 has noticeably better judgment. In Claude Code, it asks the right questions, catches its own mistakes, pushes back when a plan isn’t sound, and builds up confidence around complex, multi-service explorations before making big changes. It’s a great model to build with.”

    Tom Pritchard, Staff Engineer, in Claude Code

    “On our Super-Agent benchmark, Claude Opus 4.8 is the only model to complete every case end-to-end, beating prior Opus models and GPT-5.5 at parity on cost. For agent products in translation, deep research, slide-building, and analysis, it delivers powerful reliability.”

    Kay Zhu, Co-Founder and CTO, on the Super-Agent benchmark

    “On CursorBench, Claude Opus 4.8 exceeds prior Opus models across every effort level. Tool calling is meaningfully more efficient, using fewer steps for the same intelligence, and it carries end-to-end tasks through.”

    Michael Truell, Co-Founder and CEO, on CursorBench results

    “Claude Opus 4.8 delivers the highest score recorded on our Legal Agent Benchmark, and is the first model to break 10% overall on the all-pass standard. For substantive legal work, that’s the kind of accuracy lift that translates directly into how much real attorney work our customers can hand off with confidence.”

    Niko Grupen, Head of Applied Research, on the Legal Agent Benchmark

    “Claude Opus 4.8 feels like a major quality-of-life update over Opus 4.7: faster, easier to collaborate with, and better at carrying context and style direction across a long session. Opus 4.8 is the model I kept trusting for work where voice, taste, and technical execution all have to happen side-by-side.”

    Katie Parrott, Staff Writer, on long writing sessions

    “Claude Opus 4.8 is the strongest computer-use and browser-agent model we’ve tested, scoring 84% on Online-Mind2Web, which is a meaningful jump over both Opus 4.7 and GPT-5.5. It stays reflective and on-task in the way our customers’ agent workloads need to be reliable end-to-end.”

    Miguel Gonzalez, Tech Lead, on computer-use and browser agents

    “Claude Opus 4.8 uses tools cleanly and follows instructions with the consistency our autonomous engineering workloads need to keep running unattended. It improves on Opus 4.6 and fixes the comment-verbosity and tool-calling issues we saw with Opus 4.7. This release from Anthropic translates directly into faster capability gains for engineers building on Devin.”

    Scott Wu, CEO, on building with Devin

    “On our long-running evals, Claude Opus 4.8’s analysis was consistently higher quality than prior Opus models. It finished faster and produced richer, more information dense outputs. Overall, a noticeably better signal to noise ratio. The biggest differentiator was Opus 4.8’s tendency to proactively flag issues with the inputs and outputs of an analysis, something other models routinely missed and left to the users to catch.”

    Michael Ran, Sr. Investment Associate, on long-running analysis evals

    Claude Opus 4.8 is a quieter release than its “modest but tangible” billing suggests, because the gains land where autonomous work actually lives: a model that flags its own uncertainty, runs longer and checks itself, scales effort on demand, and stays affordable while fast mode gets cheaper. The honesty improvement alone changes the trust math for anyone deploying agents. Read Anthropic’s full announcement here.

    Related Reading

  • SpaceX S-1 IPO Filing Breakdown, Ticker SPCX on Nasdaq and Nasdaq Texas, xAI Integration, Musk’s Trillion Share Mars Pay Plan, $18.7B Revenue, and the 100 Gigawatt Orbital AI Compute Bet

    Space Exploration Technologies Corp. filed its S-1 registration statement with the SEC on May 20, 2026, kicking off the largest and weirdest IPO in modern capital markets history. The 280-page preliminary prospectus proposes to list Class A common stock on both the Nasdaq Stock Market and the new Nasdaq Texas exchange under the ticker SPCX, bundles xAI into SpaceX as a third reportable segment via a February 2026 reorganization under common control, and asks public investors to underwrite a $28.5 trillion total addressable market that explicitly includes asteroid mining, lunar manufacturing, Mars passenger transport, and 100 gigawatts per year of orbital AI compute on solar-powered satellites. The filing reports $18.67 billion of 2025 revenue and a $4.94 billion net loss, with a Q1 2026 net loss of $4.28 billion driven almost entirely by the AI segment’s $7.7 billion of quarterly capex.

    TLDR

    SpaceX is going public on Nasdaq and Nasdaq Texas as SPCX, led by Goldman Sachs, Morgan Stanley, BofA Securities, Citigroup, and J.P. Morgan. The company has been reincorporated in Texas, headquartered at Starbase, structured as a perpetual dual-class controlled company with Class B shares carrying 10 votes each and electing a majority of the board, and post-merger contains three segments: Space (Falcon, Dragon, Starship), Connectivity (Starlink with 10.3 million subscribers across 164 countries and roughly 9,600 satellites in orbit), and AI (the former xAI, including the Colossus and Colossus II superclusters in Memphis totaling about 1.0 gigawatt of nameplate compute, Grok, and the X platform with 550 million MAUs). Revenue grew from $10.4 billion in 2023 to $14.0 billion in 2024 to $18.7 billion in 2025, with Connectivity contributing $11.4 billion at a 63% segment Adjusted EBITDA margin and the new AI segment burning $1.2 billion of segment Adjusted EBITDA in 2025 while spending $12.7 billion of capex. Elon Musk holds an unspecified majority of the voting power, has a base salary of $54,080 unchanged since 2019, no key-person life insurance, and was granted in January and March 2026 a combined roughly 1.3 billion performance-restricted Class B shares that vest against market-cap milestones from $500 billion up to $7.5 trillion, with the highest tranches contingent on building a permanent Mars colony of one million inhabitants and on deploying non-Earth data centers delivering 100 terawatts of compute per year. The prospectus discloses Anthropic’s $1.25 billion per month compute deal through May 2029, a $60 billion option to acquire Cursor (Anysphere) with a $10 billion combined break fee, the Terafab one-terawatt-per-year chip JV with Tesla and Intel, the $19.6 billion EchoStar spectrum acquisition, a $20 billion SpaceX Bridge Loan, a $5 billion amended revolver, a Houston-exclusive Texas Business Court forum clause with ICC arbitration fallback, and several uniquely SpaceX risk factors including third-party Musk conduct triggering foreign asset seizures, anti-satellite weapons, cascading cyber-induced orbital debris events, and Grok’s named “Spicy” Imagine Mode and “Unhinged” Voice Mode.

    Key Takeaways

    • Ticker SPCX, dual listed on Nasdaq and Nasdaq Texas, Class A par $0.001, joint lead bookrunners Goldman Sachs, Morgan Stanley, BofA Securities, Citigroup, and J.P. Morgan, with a 22-firm syndicate including Barclays, Deutsche Bank, RBC, UBS, Wells Fargo, Allen & Company, Cantor, Needham, Raymond James, Societe Generale, Stifel, William Blair, BTG Pactual, ING, Macquarie, Mirae Asset, Mizuho, and Santander.
    • Headquartered at 1 Rocket Road, Starbase, Texas. Reincorporated from Delaware to Texas on February 14, 2024. Five-for-one forward stock split executed May 4, 2026. All share data in the filing is post-split.
    • Perpetual dual-class structure with no sunset. Class A carries 1 vote per share, Class B carries 10 votes per share, Class C carries no votes (and has been eliminated via the Class C Reclassification). Class B converts to Class A only on a non-permitted transfer.
    • Class B holders elect a majority of the board (the Class B Directors), as long as any Class B shares remain outstanding. Removing Musk from CEO or Chairman requires a separate Class B majority vote. SpaceX will be a Nasdaq controlled company and will rely on the exemptions, meaning no requirement for fully independent compensation or nominating committees.
    • Consolidated revenue: $10.39 billion in 2023, $14.02 billion in 2024, $18.67 billion in 2025, and $4.69 billion in Q1 2026 (up 15.4% year over year). Financials are retrospectively recast to combine xAI and X Holdings since both transactions were between entities under Musk’s common control.
    • Net income (loss): $(4.63) billion in 2023, $0.79 billion in 2024, $(4.94) billion in 2025, and $(4.28) billion in Q1 2026. Accumulated deficit pro forma $41.31 billion as of March 31, 2026.
    • Connectivity (Starlink) is the cash engine. 2025 revenue $11.39 billion, up 49.8%. 2025 operating income $4.42 billion, up 120.4%. 2025 segment Adjusted EBITDA $7.17 billion, up 86.2%. Consumer subscriptions are more than 60% of Connectivity revenue.
    • Starlink subscribers: 2.3 million at year-end 2023, 4.4 million at year-end 2024, 8.9 million at year-end 2025, and 10.3 million as of March 31, 2026. Roughly 9,600 broadband and mobile satellites in low Earth orbit, about 75% of all active maneuverable satellites globally. Available in 164 countries and territories.
    • Starlink ARPU is declining as the mix shifts international and lower priced: $99 monthly in 2023, $91 in 2024, $81 in 2025, $66 in Q1 2026. Management says this is expected to continue.
    • Starlink direct to cell now has roughly 650 V1 Mobile satellites and 7.4 million monthly unique devices across about 30 countries, with partnerships across roughly 30 mobile network operators including T-Mobile, Rogers, KDDI, Optus, Telstra, One NZ, Kyivstar, VMO2, Salt, and Entel. V3 satellites begin deploying in the second half of 2026, designed for 1 Tbps downlink per satellite with up to 60 per Starship launch (a 20x payload-capacity step over Falcon 9).
    • Space segment now generates lower revenue growth because Starlink dedicated launches are not booked as inter-segment revenue. Space revenue: $3.56 billion (2023), $3.80 billion (2024), $4.09 billion (2025). Falcon launches in 2025: 165 total, 43 third-party customer and 122 internal Starlink. Mass to orbit: 1,210 metric tons (2023), 1,699 (2024), 2,213 (2025). SpaceX has now launched more than 80% of the world’s mass to orbit since 2023.
    • Falcon 9 has flown roughly 620 missions with greater than 99% mission success. A single booster has been reflown 34 times. Falcon Heavy is 11-for-11 since 2018 and certified for NSSL. SpaceX flew 11 of 12 NSSL medium and heavy lift missions in 2025.
    • Starship has completed 11 flight tests and is preparing the 12th, debuting next-generation Starship, Super Heavy, and Raptor 3 from a new Starbase pad. V3 is designed for 100 metric tons fully reusable to LEO, V4 targets 200 tons. Cumulative Starship R&D investment is greater than $15 billion, including $3.00 billion in 2025 alone. Operational payload delivery to orbit is expected in the second half of 2026.
    • Dragon has flown 78 crewmembers from 20 countries since 2020 and Cargo Dragon remains the only spacecraft capable of returning meaningful mass from the ISS.
    • AI segment, the absorbed xAI business plus X, generated $818 million Q1 2026 revenue but operating losses of $(2.47) billion and segment Adjusted EBITDA of $(609) million. AI capex was $7.72 billion in Q1 2026 alone, dwarfing Space ($1.05 billion) and Connectivity ($1.33 billion).
    • Colossus and Colossus II in Memphis and Southaven Mississippi together provide about 1.0 gigawatt of nameplate compute draw. Colossus came online in 122 days with about 100,000 H100s. Colossus II added 110,000 GB200s in 91 days and 110,000 GB300s in 64 days. Next phase: another 220,000 GB300s and 400 megawatts. Industry benchmark for a 100 megawatt greenfield datacenter is two years.
    • Grok and X together have 1.3 billion supported accounts on a trailing basis, about 550 million MAUs, roughly 117 million MAUs using Grok AI features, and roughly 350 million daily posts. Imagine generates about 10 billion images and 2 billion videos per month. Paid subscribers totaled 6.3 million as of March 31, 2026 (4.4 million X Premium variants plus 1.9 million SuperGrok variants).
    • Disclosed Anthropic cloud services agreements signed May 2026: Anthropic pays $1.25 billion per month for compute capacity on Colossus and Colossus II through May 2029, ramping in May and June 2026, with 90-day termination by either party.
    • Cursor (Anysphere) compute agreement and acquisition option signed April 2026: SpaceX has the right but not the obligation to acquire Cursor at an implied $60.0 billion equity value, paid in Class A stock priced off the SPCX VWAP. SpaceX-side termination or breach triggers a $1.5 billion termination fee plus an $8.5 billion deferred services fee.
    • Terafab JV with Tesla, announced March 2026, joined by Intel in April 2026, targets one terawatt per year of compute hardware production. The filing explicitly notes that neither Tesla nor Intel is obligated to remain, and definitive agreements may not be signed.
    • Macrohard, in development with Tesla, is described as a platform designed to fully emulate digital workflows, augment human computer operation, and create a fully AI-operated software company.
    • EchoStar Spectrum Transaction (AWS-3, AWS-4, H-block, 65 megahertz US plus global MSS) was FCC-approved May 12, 2026. Total deal value $19.6 billion, including roughly $11.1 billion of equity (261.8 million Class A shares at an implied $42.40) and up to $8.5 billion of debt assumption. Closing expected around November 30, 2027.
    • Balance sheet as of March 31, 2026: cash and equivalents $15.85 billion, short-term marketable securities $7.82 billion, total assets $102.09 billion, total liabilities $60.51 billion, total debt principal $29.13 billion. The $20 billion SpaceX Bridge Loan (Goldman Sachs Bank USA as administrative agent, March 2026) refinanced legacy X and xAI debt and must be repaid within six months of IPO. The amended SpaceX Credit Facility, also May 2026, was upsized to $5.0 billion and extended to May 19, 2031.
    • Use of proceeds: expansion of AI compute infrastructure, enhancements to launch infrastructure and launch vehicles, increases in satellite constellation scale and capacity, and general corporate purposes. No dividends are anticipated and the credit agreements restrict them.
    • Total addressable market estimate of $28.5 trillion (ex-China and Russia): Space $370 billion, Connectivity $1.6 trillion ($870 billion broadband and $740 billion mobile), and AI $26.5 trillion ($2.4 trillion infrastructure, $760 billion consumer subscriptions, $600 billion digital advertising, and $22.7 trillion enterprise applications).
    • Stated future markets explicitly listed in the prospectus: point-to-point Earth transport via Starship, space tourism, in-orbit manufacturing including pharmaceuticals and materials, passenger and cargo to Moon and Mars, lunar mining of rare materials, lunar mass driver, lunar factories building AI compute satellites, asteroid mining, and orbital solar-powered AI. The headline aspirational target is 100 gigawatts per year of orbital AI compute on solar-powered satellites in Sun-synchronous orbit, with first deployments targeted as early as 2028.
    • Musk 2025 total compensation $54,080 (base salary unchanged since 2019, tied historically to California’s exempt-employee minimum). No bonus, no stock or option awards reported for 2025. SpaceX maintains no key-person life insurance on Musk.
    • January 13, 2026 Musk grant: 1 billion performance-based restricted Class B shares across 15 equal tranches tied to market-cap milestones from $500 billion to $7.5 trillion (in $500 billion increments), with at least one tranche additionally gated on “a permanent human colony on Mars with at least one million inhabitants” and on continued employment.
    • March 23, 2026 Musk replacement award (assumed from xAI): 302,072,285 performance-based restricted Class B shares across 12 tranches from $1.065 trillion to $6.565 trillion market cap, additionally requiring completion of “non-Earth-based data centers capable of delivering 100 terawatts of compute per year.” Replaces an earlier xAI award after Musk had already earned and canceled 25,172,695 Class A shares at the first milestone.
    • Gwynne Shotwell 2025 total compensation $85.81 million, primarily option awards. Bret Johnsen (CFO) 2025 total compensation $9.84 million. Non-employee directors received zero cash and zero equity for 2025 service.
    • Board of 8 post-IPO: Musk (Chairman, CEO, CTO), Shotwell (President, COO), Antonio Gracias (Valor Management), Ira Ehrenpreis (DBL Partners and Tesla), Randy Glein (DFJ Growth, audit chair), Donald Harrison (Google), Steve Jurvetson (Future Ventures), and Luke Nosek (Gigafund and Founders Fund). Class B Directors: Musk, Shotwell, Gracias, Harrison, Nosek. Common Stock Directors: Ehrenpreis, Glein, Jurvetson.
    • Lock-up is 180 days for company, directors, and officers, but Musk and certain significant investors are subject to an extended 366-day lock-up, and 100% of Musk’s shares are explicitly not subject to early-release tiers. A Directed Share Program with Schwab, Fidelity, Robinhood, SoFi, and E*TRADE handles retail allocation; DSP shares have no lock-up.
    • Corporate Opportunities waiver in the charter renounces interest in business opportunities presented to directors, officers, board observers, and their affiliates. Musk and his affiliates are explicitly not restricted from competing with SpaceX. This carve-out covers Tesla, Neuralink, The Boring Company, and any future Musk venture.
    • Exclusive forum is the Texas Business Court, Eleventh Division, in Houston, including for federal securities claims. If unenforceable, the fallback is mandatory ICC arbitration in Houston under Expedited Procedure Rules. Jury trial is waived. Class actions are prohibited.
    • Texas Business Organizations Code carve-outs: Section 21.419 codifies a statutory business-judgment-rule presumption, Section 21.552 requires 3% minimum ownership to bring derivative proceedings, and Section 21.373 (2025) requires 3% ownership for six months plus solicitation of 67% of voting power for shareholder proposals (SpaceX concedes enforceability is “expected” to be challenged).
    • Unprecedented risk-factor disclosure: in August 2024 Brazil’s Supreme Court froze Starlink’s Brazilian assets over the conduct of X “when X was not owned by us and was only affiliated with Mr. Musk.” SpaceX warns that third-party Musk conduct may continue to trigger foreign retaliation against SpaceX.
    • Risk language names Grok’s “Spicy” Imagine Mode and “Unhinged” Voice Mode as carrying heightened risks of explicit content, misinformation, and “potential nonconsensual or exploitative imagery.” A putative class action over content “representing children in sexualized contexts” is disclosed, as is an Irish DPC GDPR inquiry into Grok and an FTC inquiry into chatbots as companions for children and teens.
    • The S-1 uses the term “Department of War” (not Defense) for the federal customer requiring CMMC compliance and discloses that anti-satellite weapons have been publicly discussed by foreign governments as a tool against the Starlink constellation. A cyberattack-induced cascading Kessler-style debris event is cited as a possibility.
    • Workforce of more than 22,000 full-time employees globally, with no collective bargaining and engineering acceptance rate under 2% in 2025.
    • Operating asset footprint: Starbase (Texas, HQ, Starship), Hawthorne (California, Falcon, Dragon, Merlin and Raptor), McGregor (Texas, engine testing), Redmond (Washington, Starlink satellite production at about 70 per week), Bastrop (Texas, terminal production at tens of thousands per day, doubling in 2026 to include AI compute satellites), Kennedy and Cape Canaveral (Florida, LC-39A, SLC-40, SLC-37 in build for Starship), Vandenberg (California, SLC-4 polar launches), Memphis and Southaven (Tennessee and Mississippi, Colossus data centers), Palo Alto (California, xAI HQ), more than 400 Starlink ground stations globally, and three autonomous spaceport drone ships including “Of Course I Still Love You,” “Just Read the Instructions,” and “A Shortfall of Gravitas.”
    • Related party transactions of note: roughly $20.2 billion of equipment lease undiscounted payments to Valor (Gracias) entities guaranteed by SpaceX; aircraft, security, and tunnel-construction payments to Musk affiliates; xAI subsidiary leases real property from Musk Industries LLC.
    • Pampena v. Musk: an April 3, 2026 partial judgment in the Northern District of California, where a jury found Musk personally violated Section 10(b) and Rule 10b-5 on two May 2022 statements regarding his Twitter purchase. Post-trial motions are pending. The 2018 SEC “funding secured” settlement is also disclosed.
    • Critical accounting policy quirks: flight vehicles are depreciated over expected average number of flights rather than time. Starship costs are expensed to R&D until commercialization, then capitalized. Starlink dedicated launch costs are capitalized into Connectivity PP&E rather than booked as inter-segment Space revenue, which mechanically suppresses the headline Space growth rate.
    • The One Big Beautiful Bill Act (Public Law 119-21) reversed a $659 million U.S. R&D credit deferred tax asset recognized in 2024, driving the 2025 income tax provision of $718 million versus a $549 million benefit in 2024.
    • Pre-IPO ownership pro forma at March 31, 2026: Class A 6,824,581,339 shares and Class B 5,695,729,430 shares outstanding, for a combined 12.52 billion shares before primary issuance. Class C and the redeemable convertible preferred are converted/reclassified at close.
    • Authorized capitalization post-IPO: 36.13 billion Class A, 6.13 billion Class B, 10.0 billion Class C (none issued), and 2.4 billion preferred (none issued). Headroom for future issuance is enormous.
    • Five-for-one stock split executed May 4, 2026 to set the IPO share count and round-lot price. Price range, share count, and proceeds are bracketed in this preliminary filing and will be updated before launch.

    Detailed Summary

    A different kind of S-1 from the start

    Most S-1 filings open with corporate prose and a careful, neutral business description. SpaceX opens with an Elon Musk epigraph about wanting to wake up in the morning and “think the future is going to be great,” a mission statement that says the company exists “to make life multiplanetary, to understand the true nature of the universe, and to extend the light of consciousness to the stars,” and a Kardashev Type II framing that treats the next century of capital allocation as a civilizational project. Investors are being told, in legally binding language, that single-planet existence is “a single point of failure” and that the company is hedging against humans sharing the fate of the dinosaurs. The filing dual-lists SPCX on Nasdaq in New York and Nasdaq Texas in Dallas, picks the new Texas Business Court in Houston as exclusive forum, and reincorporates from Delaware to Texas. Every macro signal is set deliberately.

    Three segments after the xAI absorption

    The most consequential mechanical change in the S-1 is the retrospective recast of financial statements to combine xAI Holdings and X Holdings into SpaceX. Both transactions are accounted for as reorganizations of entities under common control (Musk’s), so prior-period revenue, opex, and capex move into the SpaceX line items rather than appearing as acquired-business additions. This is what produces the headline numbers: $10.4 billion (2023), $14.0 billion (2024), $18.7 billion (2025). The Space segment includes Falcon, Dragon, and Starship. Connectivity is Starlink in all its consumer, enterprise, government, and mobile forms plus the Starshield military variant. AI is the former xAI in full: Colossus and Colossus II superclusters, Grok, the X platform, and the Imagine media products. The recast also explains why net income flips so violently year to year. 2024’s $791 million net income reflects a quieter pre-merger SpaceX. 2025’s $4.94 billion net loss and Q1 2026’s $4.28 billion loss reflect the integrated AI business burning capital at unprecedented rate.

    Connectivity is the cash engine

    Starlink is the only segment that looks like a normal high-margin growth business. Revenue rose 96.4% in 2024 and another 49.8% in 2025 to $11.39 billion. Operating income tripled in 2024 and then doubled again in 2025 to $4.42 billion. Segment Adjusted EBITDA in 2025 was $7.17 billion, an EBITDA margin north of 60%. Subscribers grew from 2.3 million to 10.3 million in twenty-seven months. The constellation is now roughly 9,600 satellites, about 75% of all active maneuverable satellites on orbit. Inter-satellite laser links exceed 23,000, forming a mesh that delivers 700+ Tbps of cumulative downlink. ARPU is declining steadily, from $99 monthly in 2023 to $66 in Q1 2026, but management frames this as deliberate international mix shift toward lower priced plans and notes that direct-to-cell is just beginning to monetize. Roughly 650 V1 Mobile satellites already provide service to 7.4 million monthly unique devices through partnerships with roughly 30 mobile network operators. The EchoStar spectrum acquisition adds 65 megahertz in the US plus global MSS spectrum to support V2 Mobile broadband and 5G IoT starting in 2027.

    Space economics are obscured by accounting

    The Space segment looks small in the headline financials ($4.09 billion of 2025 revenue, an operating loss of $657 million) until you understand the accounting. Starlink launches are capitalized into Connectivity PP&E rather than booked as inter-segment Space revenue. That single policy is why 2025 Space revenue grew only 7.6% even though SpaceX flew 170 missions, of which 122 were internal Starlink. The actual operating reality is that SpaceX flew more than 80% of the world’s mass to orbit in 2025, owns 24 flight-proven reusable Falcon 9 boosters certified for 40 flights each, has refln a single booster 34 times, and has invested more than $15 billion in Starship to date. Starship’s eleventh flight test is on the books, the twelfth will debut the next-generation vehicle and Raptor 3 engine, and operational payload delivery to orbit is targeted for the second half of 2026. V3 Starship is designed to deliver 100 tons to LEO fully reusable and to carry up to 60 V3 Starlink satellites per launch, a 20x payload step over Falcon 9. The Starship cost target is a 99% reduction against the historical $18,500 per kilogram average, on the way to “airline-like” reflight cadence.

    AI is a money furnace with a thesis

    The AI segment is brand new to the SpaceX line item set and dominates the loss line. AI generated $3.20 billion of 2025 revenue (up 22.2%) but lost $6.36 billion at the operating line, much of it driven by GPU depreciation. AI capex was $12.73 billion in 2025 and another $7.72 billion in Q1 2026 alone. Colossus came online in 122 days with about 100,000 H100s and 130 megawatts. Colossus II followed with 110,000 GB200s in 91 days and 110,000 GB300s in 64 days, with another 220,000 GB300s and 400 megawatts in the next phase. The two superclusters now draw about one gigawatt combined. Grok-5 is training on Colossus II, targeting multi-trillion parameters. The X platform contributes 550 million MAUs and roughly 350 million daily posts to the segment, with 117 million MAUs touching Grok AI features. The thesis the prospectus is pitching is vertical integration on physics: SpaceX controls power generation (data center turbines and, eventually, orbital solar), launch (Starship to lift orbital compute satellites), satellite manufacturing (Redmond and Bastrop), chip supply (Terafab JV with Tesla and Intel for one terawatt per year of compute hardware), and the application layer (Grok and X). Management calls this “shovels-to-tokens” and argues no other AI company has this much control over the physical stack.

    The Anthropic, Cursor, and Terafab carve-outs

    Three subsequent events disclosed in the S-1 reframe SpaceX as a cloud and software platform as much as a hardware company. Anthropic signed cloud services agreements in May 2026 to pay $1.25 billion per month for Colossus and Colossus II capacity through May 2029, ramping in May and June 2026. The Cursor (Anysphere) agreement signed April 2026 includes both a compute commitment and an option for SpaceX to acquire the company at a $60 billion implied equity value, with a $1.5 billion termination fee and an $8.5 billion deferred services fee if SpaceX breaches or terminates. Terafab is a manufacturing JV with Tesla, joined by Intel in April 2026, with a stated one terawatt per year compute hardware production target. The prospectus is explicit that Tesla and Intel are not obligated to remain in Terafab and that no definitive agreements may be signed. Anthropic, the leading commercial competitor to OpenAI, is now SpaceX’s largest disclosed cloud customer.

    The Musk pay package

    The CEO compensation disclosure is the most aggressive in S-1 history. Musk’s reported 2025 total compensation was $54,080, a base salary unchanged since 2019. SpaceX maintains no key-person life insurance on him. Then on January 13, 2026 the board granted him one billion performance-based restricted Class B shares, vesting across fifteen equal tranches as market capitalization milestones are achieved at $500 billion increments from $500 billion all the way to $7.5 trillion, with at least one tranche additionally conditioned on the existence of a permanent human Mars colony of at least one million inhabitants and on continued employment. On March 23, 2026 the board granted an additional 302.07 million performance-based restricted Class B shares across twelve tranches from $1.065 trillion to $6.565 trillion of market cap, additionally requiring the completion of “non-Earth-based data centers capable of delivering 100 terawatts of compute per year.” This second grant replaces an earlier xAI award after Musk had already earned 25.17 million Class A shares at the first xAI milestone, which were then canceled and rolled in. The combined package is roughly 1.3 billion restricted Class B shares, dwarfing the Tesla 2018 award that previously held the record. Other executive comp is more conventional. Gwynne Shotwell’s 2025 total was $85.81 million, primarily option awards. Bret Johnsen, CFO, received $9.84 million. Non-employee directors received zero cash and zero equity for 2025 service.

    Governance built to be Musk-proof in one direction only

    SpaceX takes the dual-class playbook further than any prior tech IPO. Class B carries 10 votes per share, has no sunset, and elects a majority of the board as a separate class. Removing Musk from CEO or Chairman requires a separate Class B majority vote, and Musk holds the majority of Class B. The charter renounces interest in business opportunities presented to Musk and his affiliates, explicitly preserving his right to run competing ventures (Tesla, Neuralink, The Boring Company, anything next). The company opts into the Texas Business Organizations Code’s Section 21.419 business-judgment-rule presumption, requires 3% ownership to bring a derivative suit, requires 3% ownership for six months plus solicitation of 67% of voting power to bring shareholder proposals under Section 21.373 (a provision SpaceX itself concedes will likely be challenged in court), picks the Texas Business Court in Houston as exclusive forum even for federal securities claims, and falls back to mandatory ICC arbitration in Houston with Expedited Procedure Rules if forum exclusivity is struck down. Jury trials are waived. Class actions are prohibited. SpaceX will be a controlled company and will rely on Nasdaq exemptions from independent committee requirements. Musk and certain significant investors are subject to a 366-day lock-up rather than the standard 180 days, and 100% of Musk’s shares are excluded from the early-release tiers other holders enjoy.

    Risk factors disclose things no S-1 has disclosed before

    The Risk Factors section contains language no prior S-1 has used. SpaceX warns that “actions and statements of Mr. Musk and his affiliated ventures, whether or not directly relating to us, may draw significant public attention and scrutiny” and notes that in August 2024 the Brazilian Supreme Court froze Starlink’s Brazilian assets over the conduct of X “when X was not owned by us and was only affiliated with Mr. Musk.” That is the precedent: a foreign government seized SpaceX assets over Musk’s separate business conduct. The filing names Grok’s “Spicy” Imagine Mode and “Unhinged” Voice Mode as carrying heightened risks of explicit content and “potential nonconsensual or exploitative imagery,” discloses a putative class action over content “representing children in sexualized contexts,” an Irish DPC GDPR inquiry into Grok’s processing of EU children’s data, and an FTC inquiry into chatbots as companions for children and teens. The orbital risk language describes a cyberattack-triggered cascading Kessler-style debris event that could render SpaceX-licensed orbits “unusable for an extended period,” notes that “certain foreign governments have publicly discussed the potential use of anti-satellite weapons against the Starlink constellation,” and acknowledges that the FAA does not currently permit return-to-launch-site reentries for Starship and the company will require a waiver “which is not guaranteed.” The filing also uses “Department of War” rather than “Department of Defense” when discussing CMMC compliance for federal customers, reflecting the recent rebranding.

    Capital position and the bridge loan time bomb

    The balance sheet is large but the debt structure tells a story about why an IPO is urgent now. SpaceX has $15.85 billion of cash and $7.82 billion of short-term marketable securities against total debt principal of $29.13 billion. The largest piece is the $20 billion SpaceX Bridge Loan signed March 2026 with Goldman Sachs Bank USA as administrative agent, used to refinance legacy X and xAI debt (including X B-1, X B-3, and xAI 12.5% Senior Secured Notes). The bridge matures September 2, 2027 (extendable to March 2028 with a 0.25% fee per quarter), priced at Term SOFR plus 0.75% to 1.75%, with 0.125% duration fees kicking in at year one. It must be repaid within six months after IPO completion. The amended SpaceX Credit Facility was upsized to $5.0 billion and extended to May 19, 2031 in May 2026, with a $2.0 billion performance LC sublimit. The leverage covenant is 3.75x maximum (4.25x post-qualified acquisition). Capex is enormous and consistent: $20.74 billion in 2025 ($3.83 billion Space, $4.18 billion Connectivity, $12.73 billion AI), $10.11 billion in Q1 2026 alone. Operating cash flow ($6.79 billion in 2025) does not cover capex, and the gap is being filled by financing activity ($26.35 billion of net financing inflow in 2025).

    The 100 gigawatt orbital AI bet

    Buried in the Business section is the future-markets framing that justifies the AI-segment burn rate. SpaceX is asking public investors to underwrite a plan to deploy 100 gigawatts per year of orbital AI compute on solar-powered satellites in Sun-synchronous orbit. Reaching that scale requires thousands of Starship launches per year and roughly one million metric tons of mass to orbit annually. First modular orbital AI shells are targeted for “as early as 2028.” The justification given is that the Sun contains roughly 99.8% of the solar system’s energy, that orbital compute escapes terrestrial constraints on power, cooling, latency, and permitting, and that no other AI company controls the physical stack required to deploy at that scale. The prospectus stitches this directly to the Mars project: lunar mining of rare materials, lunar mass drivers to launch satellites at low cost, and lunar factories building AI compute satellites are listed alongside asteroid mining and Mars passenger transport as the future markets investors are being asked to value. The risk language acknowledges that none of these markets currently exist and that breakthrough advances in nuclear energy could moot the orbital compute thesis entirely. Investors are being asked to take Musk’s word that the long-tail outcomes are real options.

    Thoughts

    The most important number in this S-1 is not the revenue, the loss, or the implied valuation. It is the $54,080 Musk salary unchanged since 2019 against the 1.3 billion performance-restricted Class B shares contingent on a Mars colony and 100 terawatts of off-Earth compute. This is a pay package that resolves the question of whether SpaceX is a public-markets-style optimized corporation by answering it directly: no. SpaceX is going public on Musk’s terms, with a perpetual dual-class structure, a controlled-company exemption, a Houston exclusive forum, an arbitration backstop, a class-action prohibition, a charter that explicitly renounces interest in business opportunities Musk gets pitched elsewhere, and a CEO compensation structure that pays nothing for normal performance and 1.3 billion shares for an interplanetary civilization. Investors who buy SPCX are not buying voting power. They are buying optionality on the most ambitious capital allocation thesis a public company has ever attempted, contingent on Musk continuing to deliver outcomes the rest of the industry cannot.

    The xAI absorption is the most consequential corporate event in the prospectus and the one most worth scrutinizing. Accounting it as a common-control reorganization is technically defensible because Musk controlled all three entities, but the practical effect is to fold xAI’s enormous compute burn and X’s separate litigation surface area into SpaceX’s reported financial history without showing the deals as acquisitions. The Q1 2026 net loss of $4.28 billion is almost entirely xAI capex pulling forward. The two segments that actually make money (Connectivity at a 63% Adjusted EBITDA margin, Space when you adjust for the launch accounting policy) are being asked to subsidize an AI build-out that requires the orbital compute thesis to come true to ever generate adequate returns. Strip out AI and SpaceX would be one of the highest-quality businesses ever taken public. Include AI and it is something more like a venture-stage company stapled to a cash-flow machine, with the venture stage absorbing the cash. That is the trade the IPO is asking the market to price.

    The risk-factor language about third-party Musk conduct triggering foreign asset seizures is the cleanest single articulation in any S-1 of why founder-led companies with cross-portfolio exposure are different from normal public companies. The Brazil precedent is real, the legal theory is established, and the prospectus admits it directly. Buying SPCX means accepting that a fight between Musk and a foreign government over X content moderation, a Neuralink ethics dispute, a Boring Company permit fight, or a future venture entirely unrelated to space could trigger a freeze on Starlink subscriber revenue in that country. The Corporate Opportunities waiver is the legal mechanism that makes this acceptable to the board. It is far from clear that it is acceptable to public-market shareholders. The early reception of SPCX will partly be a referendum on whether the market thinks Brazil 2024 was a one-time event or a template.

    The Anthropic disclosure is the funniest detail. SpaceX, controlled by Musk, is now selling roughly $15 billion per year of compute to Anthropic, a company explicitly founded by former OpenAI researchers who broke away from the OpenAI-Musk faction in 2021. SpaceX-Colossus is now Anthropic’s largest disclosed compute supplier through May 2029, on 90-day termination by either side. The OpenAI lawsuit, the xAI launch, and the Grok positioning as the “truth-seeking” anti-OpenAI all sit in tension with the fact that Anthropic now anchors xAI’s third-party compute revenue. The economic logic is simple. The political logic, given the lockup of compute supply that this deal effectively creates, is fascinating. Public investors are being asked to underwrite a business where the largest compute customer is a direct AI competitor and where that supply contract is the single biggest piece of disclosed enterprise AI revenue.

    What this IPO most resembles is not Tesla’s 2010 deal or Twitter’s 2013 deal but rather a hybrid of the East India Company chartering and a moonshot R&D vehicle taken public. It is a real cash-flowing business at the Connectivity layer (the largest satellite ISP on Earth) wrapped around a launch monopoly (more than 80% of global mass to orbit) wrapped around a venture-stage AI laboratory (Colossus, Grok, the Anthropic deal, the Cursor option) all underwritten by a CEO compensation structure whose biggest payoffs require a Mars colony. The investor question is not whether any individual piece works, because three of the four pieces clearly do. The question is whether the public market will price the orbital compute and Mars optionality at zero, at a small positive number, or at the eye-watering multiple the $7.5 trillion top tranche of Musk’s pay package implies the board thinks is achievable. There is no precedent for a public company successfully executing on that scale of ambition. There is also no precedent for SpaceX, Starlink, Falcon 9, or Colossus II coming online in 91 days. The S-1 reads like the company assumes the precedent is itself.

    Read the full SpaceX S-1 filing on the SEC EDGAR system for the complete prospectus, including the financial statements and all related disclosures.

  • How the X For You Algorithm Works (May 15, 2026 Source Code Walkthrough)

    Diagram of the X For You algorithm pipeline showing content signals and user signals flowing through embed, attend, score, rank, and filter stages into a curated feed

    On May 15, 2026, xAI shipped a major update to the open-source release of the X “For You” recommendation algorithm. The repository now includes a runnable end-to-end inference pipeline, a pre-trained mini Phoenix transformer, a brand-new content-understanding service called Grox, and ad-blending logic. This is the most transparent look at how a major social feed actually picks your posts that has ever been published.

    This is the practical, plain-english guide. We read the source. Here is exactly how a post travels from your fingertips to someone’s For You tab, and what you can do to be the post that wins.

    The whole strategy in one sentence

    Write posts people reply to, repost, DM to a friend, linger on, and follow you for. Avoid anything that earns a mute, block, report, or spam flag. Space your posts hours apart. That is the algorithm.

    TL;DR

    • The For You feed is no longer a stack of heuristics. It is a single transformer-based machine learning system that predicts the probability you will like, reply, repost, share, dwell on, or hide a given post.
    • Posts come from two pools: Thunder (people you follow) and Phoenix Retrieval (the rest of X, found by similarity search).
    • A model called Phoenix scores every candidate against your engagement history. The final score is a weighted sum of nineteen predicted actions, with negative weights for “block”, “mute”, and “report.”
    • Almost no hand-engineered features survive. Freshness, verification badges, follower counts, and post type are not directly boosted. They are signals the transformer learns to use from your behaviour.
    • Out-of-network content is penalised by a tunable factor, so in-network posts have an edge by default.
    • A separate service called Grox continuously classifies posts for spam, policy violations, and brand safety. Flagged content gets filtered or de-amplified before it reaches scoring.
    • The best optimisations are still the boring ones: write posts that earn long dwell time, replies, reposts, and follows, and avoid anything that triggers mutes or reports.

    What changed on May 15, 2026

    The January 2026 release gave us the architecture but not a working system. The May update is the one that matters:

    • phoenix/run_pipeline.py replaces the separate retrieval and ranking scripts with a single inference entry point that mirrors production.
    • A pre-trained mini Phoenix model (256-dim embeddings, 4 attention heads, 2 transformer layers) is bundled as a roughly 3 GB Git LFS archive. You can run inference without training.
    • The Grox content-understanding service is now public. It runs classifiers and embedders for spam detection, post categorisation, and policy enforcement.
    • Ads blending is now in the open. So is brand-safety tracking.
    • New query hydrators mean the model sees your followed topics, starter packs, served history, impression bloom filters, IP, and mutual-follow graph at request time.
    • New candidate hydrators add engagement counts, language codes, media detection, quote post expansion, and mutual follow scores.
    • New candidate sources for ads, who-to-follow, Phoenix Mixture-of-Experts, Phoenix Topics, and prompts.

    The For You pipeline in one picture

    Every time you pull to refresh, the same pipeline runs:

    1. Query hydration: pull the requesting user’s engagement history, follow list, topics, served history, and metadata.
    2. Candidate sourcing: gather candidates from Thunder and Phoenix Retrieval in parallel.
    3. Candidate hydration: enrich each candidate with text, media, author, engagement counts, brand-safety labels, language, mutual follow scores.
    4. Pre-scoring filters: drop duplicates, posts that are too old, your own posts, blocked or muted authors, posts you’ve already seen, and posts with your muted keywords.
    5. Scoring: run candidates through Phoenix, combine the predicted action probabilities into a weighted score, attenuate repeated authors, penalise out-of-network content.
    6. Selection: sort by score, take the top K.
    7. Post-selection filtering: final visibility check for deleted, spam, violence, gore, abuse, and de-duplication of conversation branches.
    8. Side effects: cache request info, then return the ranked feed.

    That is the whole story. Every choice the system makes lives in one of those stages.

    Where candidates come from: Thunder vs Phoenix

    Two sources feed the pipeline.

    Thunder is the in-network store. It is an in-memory firehose that consumes post create and delete events from Kafka and serves sub-millisecond lookups for recent posts from people you follow. Posts older than the retention window get trimmed automatically. Thunder is why the feed feels fast.

    Phoenix Retrieval is the out-of-network source. It is a classic two-tower neural network. A user tower encodes your features and recent engagement history into a single embedding. A candidate tower does the same for every post in the global corpus. The system then does an approximate nearest-neighbour search over those embeddings to find the posts whose vectors point most similarly to yours. A dot product between vectors is all it takes.

    How many candidates from each side? The code does not hard-code a ratio. It is set at runtime via parameters (ThunderMaxResults and PhoenixMaxResults). New users get a different retrieval cluster while their account is below an age threshold and a minimum-following count.

    The creator implication is the part most guides miss: follower count is not what gets you into out-of-network feeds. Embedding similarity is. Phoenix knows nothing about how famous you are. It knows that the people who engage with posts like yours have engagement histories that look like the histories of users it is trying to serve.

    How Phoenix ranks posts

    After candidates arrive, every one of them gets a score from the Phoenix ranking transformer. The architecture, per phoenix/README.md, is small by language-model standards:

    • 128-dimensional embeddings
    • 4 transformer layers
    • 4 attention heads
    • 127-position user history sequence
    • 64-position candidate sequence
    • 1,000,000 entries each in the user, item, and author vocabularies (with 2 hash functions per entity)
    • 19 predicted action types

    The transformer’s input is a sequence of your past engagements. Each engagement carries the post you engaged with, the author, the action you took, and the product surface (For You, profile, search). The candidates are appended as a second segment. The model uses candidate isolation masking: candidates can attend to your history but not to each other. This is a deliberate engineering choice. It means a post’s score does not depend on the other posts in the batch, which keeps scoring cacheable and consistent.

    The output is one probability per action type, per candidate.

    The action weights: what positive and negative engagement is worth

    The Weighted Scorer combines those probabilities into a single number:

    “ Final Score = Σ (weight_i × P(action_i)) “

    The exact weight values are not in the open-source repo. They live in an external configuration crate (xai_home_mixer) that xAI tunes continuously. What the repo does show us is the shape of the signal, and that is what matters for strategy.

    Positive weights are applied to these predicted actions:

    • favorite
    • reply
    • retweet
    • quote
    • quoted click
    • click
    • profile click
    • photo expand
    • video view (only counted if the video is above a minimum duration)
    • share
    • share via DM
    • share via copy link
    • dwell (the user lingered on the post)
    • continuous dwell time (how long they lingered)
    • follow author

    Negative weights are applied to:

    • not interested
    • block author
    • mute author
    • report

    A few observations matter for creators. Replies, reposts, follows, and DM shares are listed as their own separately-weighted actions, which is the strongest signal we have that xAI treats them as more valuable than a like. Dwell and continuous dwell time are split into two predictions, which means how long the average person reads your post is its own ranking lever. And the negative actions are not just filters. They actively push the score down for posts that even slightly resemble content that triggers mutes or reports.

    The out-of-network penalty

    Out-of-network content does not start on an even footing with in-network content. The OON Scorer multiplies a candidate’s score by a configurable factor that is less than one when the candidate came from Phoenix Retrieval. This is why following the right accounts still matters.

    There is an explicit override for new users: if your account is fresh and you follow at least the minimum number of accounts, the OON penalty is softened with a more permissive factor. This is the bootstrap mechanism that gets new users a populated For You feed before they have generated enough engagement history for Phoenix to personalise around.

    The author diversity damper

    Once posts are sorted, the Author Diversity Scorer runs. It applies an exponential attenuation to repeated authors:

    “ multiplier(position) = (1 - floor) × decay^position + floor “

    The first post from an author keeps its full score. The second is attenuated. The third more so. A floor value prevents an author from being attenuated below a minimum. The decay and floor are tunable parameters, not constants. The practical consequence is that posting fifteen times in twenty minutes does not produce fifteen top-of-feed impressions. It produces one or two.

    What Grox does (and why it is the most important new piece)

    The grox/ directory is new in May 2026 and is the single most underreported part of the update. It is an asynchronous task-execution engine that runs content classifiers and embedders on every post. Among the tasks present in the repo:

    • Spam detection, including a low-follower reply-spam classifier
    • Safety policy classification across categories like violent media, adult content, hate, self-harm, and platform-policy violations
    • A “post safety screen deluxe” pipeline that re-checks adult content classification with a second pass
    • Media classification for images and video
    • Multimodal post embedding for retrieval and ranking

    Grox does not directly score posts. It produces labels. Those labels are consumed by the visibility filter (VFFilter) at the post-selection stage and by the brand-safety hydrator that ad placement uses. The effect is that posts the system thinks are spam, policy-violating, or unsafe for ads next to do not get removed entirely. They get de-amplified before they ever reach the Phoenix ranking step, or filtered after.

    If you have wondered why a perfectly reasonable post sometimes mysteriously underperforms, this is the likely culprit. A Grox classifier flagged something.

    What the algorithm explicitly does not boost

    This is worth saying twice because the rumour mill keeps repeating the opposite. In the open source code:

    • There is no verified-badge boost in the scorer.
    • There is no follower-count boost.
    • There is no link penalty. Links are not separately weighted.
    • There is no freshness boost. The age filter removes posts above a threshold but does not score newer posts higher.
    • Subscriber status is used for filtering paywalled content, not for boosting reach.

    If verification, Premium, or any of these correlate with reach in practice, it is because the Phoenix transformer has learned to predict that users engage with those posts more, not because a hand-written rule said so. The whole point of the architecture, per the repo, is that every such heuristic has been removed and the model learns the signal from your engagement sequences.

    How ads get inserted

    Ads ride along through the same pipeline. The blender requires at least five organic posts before an ad can be placed. It computes a spacing interval, partitions candidate ads by brand-safety verdict, and caps the number of ads from the safe set to roughly half the safe-set size. A second layer of contextual checks drops ads when neighbouring posts have a weak brand-safety rating, a conflicting handle, or a keyword collision. The result is an interleaved feed that tries to keep brand-safety risk down without starving the auction.

    How creators should post for the 2026 algorithm

    Stop optimising for proxies. Optimise for the actions Phoenix is actually predicting:

    1. Write for replies and reposts, not for likes. A like is one positive weight. A reply, a repost, a quote, and a follow are each separately weighted on top. Posts that ask a question, take a stance, or offer a frame for someone else to argue with consistently outperform posts that close a thought.
    2. Aim for dwell. A long-form thread, a clear photo, or a video that people watch to the end gets two positive signals: dwell and continuous dwell time. A one-line post you scroll past in a quarter-second gets neither.
    3. Earn the follow. Follow-author is a predicted action with its own weight. A post that successfully sells a new viewer on hitting follow scores more than a post that doesn’t.
    4. Do not cluster. Author Diversity attenuates your second and third posts inside the same scoring window. If you have three things to say, space them out by hours, not minutes.
    5. Avoid anything that gets you muted, blocked, or reported. Those carry explicit negative weights. Engagement bait that produces a single block does measurable damage to the score of that post and any signal it sends about the author.
    6. Do not be Grox-flagged. Spam-shaped behaviour (reply-bombing with the same line, posting at high frequency with a low follower count, low-quality media) gets you classified by Grox before you ever reach the scorer.
    7. Follow more accounts in your niche. Phoenix Retrieval is similarity-based, but the OON penalty means in-network candidates still have a head start. The denser your in-network graph in your niche, the more likely your posts surface there.
    8. Build an engagement history that Phoenix can recognise. The user tower encodes your recent engagement. If you want your content to surface to people who like topic X, engage like a person who likes topic X. The model will learn to send your posts to that cluster.
    9. Lean into video and photo. Photo expand and video view are both separately weighted positive actions. They give a single post more ways to score.
    10. DM-share-worthiness is a quiet superpower. Share via DM and share via copy link are each their own weighted action. A post worth sending to a specific friend is, mechanically, a higher-scoring post than a post merely worth liking.

    Can I run the algorithm locally?

    Yes. With the May 15 release, the runnable inference path is phoenix/run_pipeline.py, and the bundled mini Phoenix checkpoint is enough to score sample posts. You can clone the repo, pull the LFS archive, and watch the pipeline rank a batch end to end. This is, as far as we know, the first time a production-scale social recommendation system has shipped a runnable inference path to the public.

    What is next

    Two trends are worth watching. The first is the cadence: xAI has been pushing material updates every few weeks. Expect the action weights, retrieval ratios, and Grox classifier set to keep moving. The second is the architecture: candidate sources for “Phoenix MoE” and “Phoenix Topics” suggest the next direction is multiple specialised ranking experts rather than a single transformer, with topic awareness fed in explicitly. Promptable feeds (telling X in natural language what you want more of) are the user-visible end of that trend.

    The closing point is the practical one. The 2026 For You algorithm is, more than any version before it, a measurement of how people respond to your post. Strategies that try to game routing, freshness, or format are landing in a system that does not care about those things directly. Strategies that earn replies, holds, follows, and shares are landing in a system that is built, end to end, to reward exactly that.

  • The Book of Elon by Eric Jorgenson: Complete Summary of Musk’s Operating System, The Algorithm, The Tesla Master Plan, and the 69 Core Musk Methods

    Infographic summary of The Book of Elon by Eric Jorgenson covering The Algorithm Tesla Master Plan SpaceX Mars and the 69 Core Musk Methods

    Eric Jorgenson’s The Book of Elon: A Guide to Purpose and Success (Magrathea Publishing, 2026) is the third entry in his series of compiled-wisdom books, following The Almanack of Naval Ravikant and The Anthology of Balaji. It is built entirely from Elon Musk’s own words, drawn from transcripts, tweets, and interviews across his career, then recontextualized into a four-part operating manual: Pursue Purpose, Ultra Hardcore Work, Building Companies, and On Behalf of Humanity. The book closes with a bonus list of 69 distilled maxims. Naval Ravikant wrote the foreword and calls it “the only book an entrepreneur needs.” Jorgenson’s stated goal is “one million Musks.” This is a complete, dense summary of every major idea in the book, including The Algorithm verbatim with each of its five steps explained in depth, the Tesla Master Plan, the first-principles battery cost calculation, the SpaceX rocket cost analysis, the seven existential risks, the Mars colonization plan, and the 69 Core Musk Methods in full. Get the book at elonmuskbook.org.

    TLDR

    The Book of Elon argues that Musk’s results are not an accident of genius but the output of a learnable operating system. The system has four layers. Layer one is purpose: optimize your life for usefulness, which Musk defines mathematically as number of people helped multiplied by magnitude of help per person. Layer two is epistemology: reason from physics and raw-material costs, not from analogy or precedent. Layer three is execution: take responsibility, hire only exceptional people, design organizations that route around hierarchy, run at maniacal urgency, and treat the factory as the product. Layer four is mission: pick problems whose solutions move civilization forward (sustainable energy, reusable spaceflight, AI alignment, brain-computer interfaces, multiplanetary life). The book’s single most important operational artifact is The Algorithm, Musk’s five-step engineering process that must be applied in order: make your requirements less dumb, try very hard to delete the part or process, simplify or optimize, accelerate cycle time, automate. The 69 Core Musk Methods at the end of the book are the entire operating system compressed to one-line maxims. Naval frames it as a choice for the reader: when humanity goes to the stars, you can be in the front row cheering or sour-faced in the bleachers jeering, but there is also a third option, which is to copy the methods and build something yourself.

    Key Takeaways

    • Optimize for usefulness, not for money, fame, or comfort. Musk’s daily question is “how can I be useful today” and his success metric is number of people helped multiplied by magnitude of help per person.
    • Five domains will most influence the future: the internet, sustainable energy, space exploration, artificial intelligence, and the genetic rewriting of biology. Pick one and contribute.
    • It is possible for ordinary people to choose to be extraordinary. Convention is optional. The default settings of a culture are not laws of nature.
    • Physics is law. Everything else is a recommendation. If a plan does not violate conservation of energy or any other physical principle, it is at least theoretically possible.
    • First-principles thinking is the antidote to “that’s how it’s always been done.” Break a problem down to atomic constraints (raw material cost, physics, basic operations) and reason up from there. The battery pack example is canonical: people said cells would always cost $600/kWh, but the raw cobalt, nickel, aluminum, carbon, polymers, and steel at London Metal Exchange prices added up to only $80/kWh.
    • Track two ratios on everything you build: the magic-wand number (raw-material cost as a floor for finished cost) and the idiot index (finished cost divided by raw-material cost). Anything with a high idiot index has enormous room for improvement.
    • Aspire to be less wrong. You will not be right every day. Being less wrong most of the time, with a clear feedback loop to reality, is the realistic target.
    • Engineering is magic, and engineers are the magicians of the 21st century. Science discovers what is. Engineering creates what was not.
    • Take responsibility. Musk is CEO of Tesla and SpaceX because he feels responsible for them, not because it improves his quality of life. The worst problems are the CEO’s job, not the best problems.
    • Sleep on the factory floor. Leadership is shared suffering, not delegated comfort. Seeing is believing. If the CEO can do it, the team will do it.
    • Startups are eating glass and staring into the abyss. Glass is the work you do not want to do. The abyss is the constant threat of company death. Both are required.
    • Adversity forges strength. A high ego-to-ability ratio breaks your feedback loop. Suffer enough early to develop the pain threshold needed later.
    • The most important job is attracting exceptional people. Money is not the constraint. Exceptional talent is the constraint.
    • Hire only Special Forces. The minimum passing grade is excellent. A small group of technically strong people will always beat a large group of moderately strong people.
    • Hire for character as much as for skill. Skills are teachable. Attitude is not. Judge a person by the character of their friends and associates and to some degree by their enemies.
    • Camaraderie can be dangerous because it prevents truth-telling. Physics does not care about hurt feelings. It cares about whether you got the rocket right.
    • All bad news should be given loudly and often. Good news can be said quietly and once.
    • Communication should travel via the shortest path necessary to get the job done, not through the chain of command. Anyone should be able to talk to anyone.
    • The organization manifests in the product. Silos produce redundancy, waste, and error. Acronyms and jargon are cognitive pollution.
    • Innovation needs permission to fail. If failure is not an option, you get incremental progress and nothing else.
    • Simplicity creates both reliability and low cost simultaneously. The best part is no part. The best process is no process.
    • The Algorithm, verbatim, in mandatory order: (1) Make your requirements less dumb. (2) Try very hard to delete the part or process. (3) Simplify or optimize. (4) Accelerate cycle time. (5) Automate. See the deep-dive section below for each step in detail.
    • If you are not adding deleted things back in roughly 10 percent of the time, you are not deleting enough. Overcorrect.
    • Requirements must come from a named person, not a department. Requirements from smart people are the most dangerous because you are less likely to question them.
    • Speeding up something that should not exist is absurd. If you are digging your grave, do not dig it faster. Stop digging.
    • Automation is last, not first. Tesla’s Nevada and Fremont factories had to rip out hundreds of expensive robots that had been installed before The Algorithm’s first four steps were complete.
    • A maniacal sense of urgency is the operating principle. The only true currency is time. Every minute lost is gone forever.
    • Speed is both offense and defense. The SR-71 Blackbird has almost no defense except acceleration. Innovating faster is more durable than any patent.
    • Do things in parallel. A factory moving at twice the speed of another factory is basically equivalent to two factories.
    • Be a vector, not a scalar. High speed in the right direction. Course-correct like a guided missile.
    • Manufacturing is underrated. Design is overrated. There is 1,000 to 10,000 percent more work in the production system than in the product itself.
    • The factory is the product. The biggest Tesla epiphany was that what really matters is “the machine that builds the machine.”
    • Attack the constraint. The production line moves at the speed of the slowest, least lucky part. Out of 10,000 things, the one that is not working sets the production rate.
    • Manufacturing is the moat. Maximize economies of scale and maximize manufacturing technology. The combination is uncopyable.
    • Zip2 (1995, started with $2,000) sold to Compaq for over $300 million. Musk’s first major lesson: sell directly to consumers, not through legacy gatekeepers who will misuse the technology.
    • X.com merged with Confinity to become PayPal, which sold to eBay in 2002 for $4.5 billion. Musk had been removed as CEO during a honeymoon trip but did not contest it to avoid disrupting the company during a crisis. “Life is too short for long-term grudges.”
    • Listen well, correct fast. X.com’s initial financial-services conglomerate failed; the email-payments demo worked instantly. Musk pivoted to what the market wanted and powered viral growth (one million customers in year two, no sales force, no marketing spend).
    • Musk reinvested his post-tax PayPal proceeds (~$180 million) split across Tesla (~$70M), SpaceX (~$100M), and SolarCity (~$10M). Costs were 2x his estimates on every company.
    • Tesla Master Plan (August 2006): (1) Build a sports car. (2) Use the profits to build an affordable car. (3) Use those profits to build a mass-market car. (4) Provide zero-emission power generation. The strategy was forced by the economics of new technology: you cannot start at the bottom of the market without scale, so you start with low-volume, high-margin and use the margin to fund scale.
    • Tesla nearly died on Christmas Eve 2008. The final funding round closed at 6 p.m., hours before payroll would have bounced. Musk had moved into Jeff Skoll’s guest bedroom. Daimler then put $50M into Tesla after Musk’s team dropped a Tesla powertrain into a Smart Car that hit 60 mph in 4 seconds.
    • Model 3 production “hell” lasted 2017 to 2019. Musk slept on the Fremont and Nevada factory floors for three years. “The longest period of excruciating pain in my life.”
    • Give people more for less. Don’t spend on advertising. Spend on engineering and design so the product carries itself through word of mouth.
    • SpaceX was founded in mid-2002 with $100 million of Musk’s PayPal money. He expected to lose everything. There was no external funding for three years.
    • SpaceX had budgeted for exactly three failed Falcon 1 launches. Launches 1, 2, and 3 failed (2006, 2007, 2008). Launch 4 succeeded in August 2008. Then NASA called with a $1.6 billion cargo resupply contract, saving SpaceX and indirectly Tesla. Musk reportedly screamed “I LOVE NASA. YOU GUYS ROCK.”
    • Rockets are expensive only because of legacy supply chains, cost-plus contracting, and outsourcing through five layers of subcontractors (“overhead to the fifth power”). The raw materials of a rocket are 1 to 2 percent of finished cost. The half-nozzle jacket Musk uses as an example cost $13,000 but contained $200 of steel.
    • Full and rapid reusability is the holy grail of rocketry. With reuse, only propellant cost remains, which is mostly liquid oxygen and methane at around $1 million per Starship flight.
    • Optimize for the right thing. SpaceX’s actual optimization target is “fastest time to a self-sustaining city on Mars.” That cascades to fastest time to a fully usable rocket, then fastest time to orbit. Early Starship had no doors because doors are not necessary for reaching orbit.
    • Companies are the most reliable engine of progress and the deepest form of philanthropy because they create durable wealth and deploy capital toward problems. “I care about reality. Perception be damned.”
    • The Age of Abundance is coming via AI and humanoid robotics. Optimus and competitors will eventually outnumber humans, removing labor as the economy’s binding constraint. The market for humanoid robots will exceed the market for cars.
    • Tesla’s full self-driving and Robotaxi product is forecast to make Tesla a $10 trillion company. Autonomous cars are worth 5 to 10 times non-autonomous cars because they earn money when their owners are not using them.
    • Neuralink achieved 2 bits per second of brain output with the first patient, Noland Arbaugh. Musk’s 5-year target is one megabit per second. Long-term: consensual telepathy via two BCIs, plus restoration of vision (Blindsight) and eventually multispectral senses (infrared, ultraviolet, radar).
    • Musk’s seven named existential risks: (1) World War III, (2) Regulation accumulation, (3) Unsustainable energy, (4) Misaligned artificial superintelligence, (5) Population collapse, (6) Asteroids and comets, (7) Civilizational fragility itself.
    • Population collapse is the risk most underdiscussed. The US has been below replacement since the early 1970s; sustained only by immigration and longevity. China’s three-child policy failed; the country is 40 percent below replacement. Musk: “We need to revive the idea of having children as a social duty.”
    • Do not force an AI to lie. The HAL 9000 lesson from 2001: A Space Odyssey is that AI given conflicting instructions, one of which is to deceive, becomes dangerous. Truthfulness as a core training objective is the alignment mitigation Musk advocates.
    • Becoming multiplanetary is an evolutionary-scale event. Six milestones in Earth history: single-celled life, multicellular life, plants/animals, ocean-to-land, consciousness, and now multiplanetary life. “At least as important as life going from the oceans to land, probably more significant.”
    • The window of opportunity is open right now. We cannot count on it being open for long. Stephen Hawking estimated roughly 1 percent civilizational-end probability per century. “That’s Russian roulette with 99 empty barrels and every century is a click.”
    • Mars insurance costs less than 1 percent of Earth GDP. The plan: 1,000 Starships per Mars transfer window (every 2 years), eventually a fleet of thousands lifting off together. Target: 1 million tons of cargo and people on Mars by 2044, then a self-sustaining civilization.
    • Mars terraforming options Musk names: thousands of solar reflectors in orbit, or detonating thermonuclear devices over the polar caps as “two little suns” to vaporize CO2 ice, thicken the atmosphere, and eventually create liquid oceans roughly a mile deep covering 40 percent of the planet.
    • Even given pure slower-than-light travel and no new physics, a million-year time horizon allows humanity to colonize the entire galaxy and possibly neighboring galaxies. “We are at the very, very early stage of the intelligence big bang.”
    • The 69 Core Musk Methods at the end of the book are the entire system in maxim form. The full list appears later in this article.

    The Algorithm in Detail: Musk’s 5-Step Engineering Process

    The single most important operational artifact in the book is what Musk calls “The Algorithm.” It is a five-step engineering process he developed and enforces across Tesla, SpaceX, the Boring Company, Neuralink, and xAI. Every part, every process, every line of code, every requirement, every meeting is supposed to be put through these five steps. The order is mandatory. Reordering them is the most common failure mode and the source of nearly every major mistake Musk says he has made at scale (most famously the Nevada and Fremont automation disaster). The book treats The Algorithm as the practical compression of first-principles thinking into a daily ritual.

    The five steps, in mandatory order, in Musk’s own phrasing:

    1. Make your requirements less dumb.
    2. Try very hard to delete the part or process.
    3. Simplify or optimize.
    4. Accelerate cycle time.
    5. Automate.

    The book devotes its longest single chapter to explaining each step, why the order matters, and the specific failure mode that occurs when you skip ahead. Here is every step in depth.

    Step 1: Make Your Requirements Less Dumb

    The first step is the hardest because it is the most psychologically uncomfortable. Musk’s exact framing in the book: “Your requirements are definitely dumb. It does not matter who gave them to you. Requirements from smart people are the most dangerous, because you’re less likely to question them.”

    The operational rule that follows is concrete. Every requirement on every part, process, deliverable, or specification must come from a named human. Not from a department. Not from a regulation document. Not from “the customer.” A name. Track who owns each requirement in writing. If the named person has left the company, retired, or cannot remember why they wrote the requirement, the requirement should be presumed dumb until proven otherwise. Many requirements in any organization are legacy beliefs nobody currently defends. They exist because they existed yesterday and nobody felt empowered to delete them. The Algorithm starts by demanding evidence for every assumption.

    The reason requirements from smart people are especially dangerous is that smart people are persuasive. A specification handed down by a respected engineer carries the implicit authority of “if she said this, there is a reason.” Most of the time there is no reason left, or the reason was contextual to a moment that no longer applies. The Algorithm’s first step is to put every smart-person requirement on equal footing with every dumb-person requirement and force a present-tense justification. If the justification cannot be reconstructed, the requirement is dumb regardless of the author’s IQ.

    The mental shift this step demands is to treat requirements as recommendations and treat the laws of physics as the only fixed authority. Musk repeats this constantly: “All requirements should be treated as recommendations. The only fixed laws are the laws of physics.” Once you internalize that frame, the requirements doc stops being scripture and becomes a draft that is open to revision in every meeting, every day.

    Step 2: Try Very Hard to Delete the Part or Process

    Once the requirements survive scrutiny, the second step is aggressive deletion. The Algorithm’s specific test for whether you are deleting enough: “If you’re not adding deleted things back in 10 percent of the time, you’re clearly not deleting enough.” The 10 percent is a forcing function. If you delete and never have to restore, you are not pushing hard enough; you are leaving safe deletions on the table.

    The book explains why engineers chronically under-delete. Every engineer remembers the painful moment when they deleted something and it turned out to be load-bearing. That memory is so vivid that it overshadows the silent cost of thousands of unnecessary parts that nobody ever questions. The Algorithm corrects for this asymmetry by deliberately overshooting. The instruction is explicit: “We are on a deletion rampage. Nothing is sacred.”

    The application is mechanical. For every part on the bill of materials, every step in the production process, every meeting on the calendar, every requirement in the spec, every line in the documentation, every approval in the workflow: try to delete it. If deleting causes nothing to break for 30 days, leave it deleted. If something breaks and you have to add it back, do so without shame; that is the 10 percent. The maxim that summarizes this step appears multiple times in the book: “The best part is no part. The best process is no process.”

    The canonical example in the book is the fiberglass-mat story. Tesla’s battery pack had a layer of fiberglass mats between the battery cells and the underbody. The mats had a dedicated production process that had been automated, accelerated, and optimized over years. Engineers had spent millions perfecting the glue, the cure time, the cutting tolerances, the robotic placement. Then Musk asked a simple question: “What are these mats for?” The battery team said “noise and vibration.” Musk asked the noise and vibration team. They said “fire safety.” The fire-safety team had no idea where the mats came from. So Musk had two cars built, one with the mats, one without, and put microphones in both. There was no detectable difference. Deleting the part eliminated a $2 million robotics step that had been built up over years. “It was like being in a Dilbert cartoon.”

    The fiberglass-mat story is the entire point of The Algorithm in miniature. Tesla had already automated step five, accelerated step four, optimized step three, and skipped steps one and two entirely. The whole apparatus existed to perfect a part that should not have existed. Steps one and two would have found this in a single meeting.

    Step 3: Simplify or Optimize

    Only after steps one and two have been completed in earnest do you simplify or optimize what is left. Musk’s exact warning: “The most common mistake of smart engineers is to optimize a thing that should not exist.”

    The book argues that this mistake is systematically produced by education. High school and college train convergent logic: you are given a question and graded on the elegance and correctness of your answer. The question itself is never on the table. After 16 to 20 years of this, most engineers, scientists, and analysts are mentally locked into “optimize the question in front of me” mode and physically cannot ask whether the question should be deleted. The Algorithm is designed to override that training. Steps one and two are explicitly the act of questioning the question; only at step three do you finally get to apply the optimization skills that school rewarded.

    What “simplify or optimize” looks like in practice: reduce part counts, combine functions, choose materials that are abundant rather than exotic, eliminate processing steps within a part’s manufacturing, reduce the number of inputs the team needs to track, collapse separate tools into one tool, replace bespoke fasteners with standard ones, replace any custom solution with a commodity solution that is good enough. The book’s framing is that simplicity creates both reliability and low cost at the same time, with no trade-off. A simpler part is cheaper to build, cheaper to inspect, cheaper to repair, fails less often, and breaks in more predictable ways when it does fail. Optimization without simplification almost always increases complexity and therefore increases failure modes.

    The Algorithm treats simplify and optimize as one step but acknowledges they are different operations. Simplify is structural: fewer pieces. Optimize is parametric: better values for the pieces you keep. Both are legal at step three, but neither is legal before steps one and two have been honestly executed.

    Step 4: Accelerate Cycle Time

    Once requirements are minimal, parts are deleted, and what remains is simplified, the fourth step is to go faster. The specific maxim: “Once you’re moving in the right direction, and moving efficiently, you’re moving too slow. Go faster.”

    The reason acceleration comes fourth, not first, is in another Musk line: “Speeding up something that shouldn’t exist is absurd. If you’re digging your grave, don’t dig it faster. Stop digging.” Speed multiplies the value of correct decisions and the cost of incorrect ones. Apply it before steps one through three and you scale your mistakes. Apply it after and you scale your gains.

    Acceleration at step four is everything that compresses the time between iterations. Shorten meetings. Eliminate approval queues. Run things in parallel that were running in series. Move people physically closer to the work so that information travels at the speed of conversation instead of the speed of email. Set aggressive internal deadlines that force the team to find shortcuts they would not otherwise have looked for. Replace any tool, supplier, or process that is slow with one that is faster, even if it is slightly more expensive per unit, because cycle time compounds.

    The book frames acceleration as both offense and defense. As offense, faster iteration lets you out-innovate competitors who are stuck on slower cycles. As defense, the SR-71 Blackbird analogy: the plane has almost no defensive systems because its acceleration is its defense. A company that ships faster than competitors can copy does not need patents, because patents protect static IP and speed protects evolving IP. The maxim Musk repeats is: “A factory moving at twice the speed of another factory is basically equivalent to two factories.” The Colossus supercluster story is the application: xAI built 100,000-GPU infrastructure in 122 days against a supplier estimate of 18 to 24 months, then doubled it in 92 more, by attacking the problem in parallel across building, power, cooling, and networking, all working 24/7 in four shifts.

    Step 5: Automate

    Automation comes last. Always. This is the step where most companies start and where Musk himself made his most expensive single mistake. The book quotes him directly: “The big mistake I made in the Tesla factories in Nevada and Fremont was trying to automate every step too early. To fix that, we had to tear hundreds of expensive robots out of the production line.”

    The reason automation must be last is that automation locks in a process. Once you have built robots, written PLC code, calibrated machine vision systems, and integrated them into your factory floor, the cost of changing the underlying process is enormous. If the process you have automated should not exist (step 2 failure), is more complicated than necessary (step 3 failure), or runs at the wrong cadence (step 4 failure), you have just spent millions of dollars institutionalizing your mistakes. Tesla’s experience was exactly this: robots installed before the underlying process was clean and simple ended up being expensive obstacles to the eventual correct process.

    The correct order is reverse. First make sure the part should exist (step 1). Then delete it if you can (step 2). Then simplify the part and the process around it (step 3). Then run it manually at maximum speed (step 4). Only after a human-run process is fast, simple, and clearly necessary do you automate it. By that point, the automation is purchasing leverage on a known-good system, not freezing a guess.

    The book notes that automation done last is also cheaper to build, because the process being automated is simpler. Automating a 20-step process requires a 20-stage robotic system. Automating the 5-step version of the same process that emerged from steps 1 through 3 requires a 5-stage robotic system. The savings from doing steps 1 through 4 first show up directly in the capital cost of step 5.

    How to Run The Algorithm: The 24-Hour Cadence

    The book treats The Algorithm as a daily practice, not a one-time exercise. Maxim 22 in the 69 Core Musk Methods reads: “For critical items, have meetings every twenty-four hours to run The Algorithm and check progress from yesterday.” For any deliverable that is on the critical path, the team meets daily, walks through the five steps in order, and reports concrete progress on each step. Requirements that survived yesterday are re-questioned today. Parts that survived deletion yesterday are re-evaluated today. Steps three through five proceed in parallel with the continuing daily challenge of steps one and two. The cadence is what prevents The Algorithm from becoming a poster on the wall.

    Common Failure Modes

    The book identifies the specific ways teams skip steps. Skipping step 1 happens when a respected engineer’s requirement is treated as immutable; the fix is to make every requirement come from a named human and be re-justified on demand. Skipping step 2 happens when engineers prefer to optimize a part rather than delete it, because deletion creates immediate visible risk while optimization creates invisible long-term cost; the fix is the 10 percent restoration rule. Skipping step 3 in favor of step 4 happens when management demands speed before the system is clean; the fix is the “digging your grave” check before any acceleration program is approved. Skipping step 4 in favor of step 5 is the most expensive mistake and the one Musk says he personally committed at the Tesla Nevada and Fremont factories; the fix is the explicit rule that humans must run a process at speed before robots are introduced.

    The throughline is that The Algorithm protects you from your own intelligence. Smart engineers are very good at steps three through five. They are bad at steps one and two because the schooling system that produced them never asked them to question the question. The order of The Algorithm is therefore the order in which discomfort decreases. Step 1 is the most uncomfortable. Step 5 is the most fun. Most organizations run the algorithm in fun-first order and pay for it with multimillion-dollar fiberglass-mat-style monuments to optimization without deletion.

    Detailed Summary

    The book’s structure and method

    Jorgenson built the book entirely from Musk’s own words across decades of transcripts, tweets, and interviews. He notes explicitly that he edited for clarity, brevity, and flow, that all material is recontextualized, and that readers should verify phrasing with primary sources before citing. The four parts of the book are presented as a curriculum, not a biography. Part I lays the philosophical foundation. Part II teaches the operating tempo and methods. Part III applies those methods through the actual histories of Zip2, X.com/PayPal, Tesla, SolarCity, and SpaceX. Part IV widens the lens to civilizational risks and the multiplanetary mission. The bonus section, “The 69 Core Musk Methods,” compresses the whole book into a maxim-by-maxim reference. Naval Ravikant’s foreword frames the underlying claim: Musk’s methods are copy-able, and “if your motives are pure and greater than yourself, the world will conspire in its subtle ways to help you.” Jorgenson’s stated dream is “one million Musks.”

    Part I: Pursue Purpose, the foundation of a unique life

    Musk’s daily question is “how can I be useful today.” His success metric is mathematical: total impact equals number of people helped multiplied by magnitude of help per person. He identifies five domains as having the largest possible impact on the future of humanity: the internet, sustainable energy, space exploration, artificial intelligence, and the rewriting of genetics. He repeats that it is possible for ordinary people to choose to be extraordinary, that convention is not law, and that the best work is found at the intersection of what you are good at, what you enjoy, and what improves humanity. He warns against zero-sum thinking, framing the economy as a growable pie rather than a fixed one. He notes that consumer adoption is unreliable as a guide: a 1946 to 1948 survey found 96 percent of people would never buy a television, and Tesla heard the same about electric cars before launch.

    The middle chapter teaches first-principles thinking. The technique is to break a problem into its atomic constituents (raw material costs, physics, basic operations) and reason up from there, ignoring analogy and precedent. The canonical example is battery cells. People said they would always cost about $600 per kilowatt-hour. Musk priced the actual materials at the London Metal Exchange (cobalt, nickel, aluminum, carbon, polymers, steel) and got $80 per kWh, proving cheap EVs were a manufacturing problem, not a physics one. He uses the same technique for rockets, where finished cost is typically 10 to 100 times raw-material cost. The half-nozzle jacket example: $13,000 list price, $200 of actual steel. He names two ratios that operationalize this: the magic-wand number (raw-material floor) and the idiot index (finished cost divided by raw-material cost). High idiot index means high opportunity. He also teaches “thinking in limits”: scale the variable to extreme values to expose hidden constraints, then iterate back to feasible regimes. His tunneling example is illustrative: LA subway costs about $1 billion per mile, but shrinking tunnel diameter from 28 feet to 12 feet drops cross-section 4x, and combining that with continuous tunneling and reinforcement enables an 8x cost improvement.

    The third chapter of Part I makes the case for engineering itself. Science discovers what already exists. Engineering creates what did not. Engineering, Musk says, is magic, and engineers are the magicians of the 21st century. He grounds this historically: Roman military dominance came from metallurgy (martensitic steel swords) and roads (logistical advantage), and Rome fell when its technological edge was matched and routed around. The WW2 Pacific air war was won by the side with the faster innovation loop, not the side that started with better fighters. Nuclear weapons were the ultimate winner-take-all. Tesla’s powertrain is sold to Toyota, Daimler, and Mercedes precisely because it is hard. “If it was easy, they would do it.” The lesson is that durable value sits where the engineering is genuinely difficult, not where the marketing is loud.

    Part II: Ultra Hardcore Work, teams, organization, urgency, manufacturing

    Part II is the operating manual. The first chapter, “What It Takes,” argues that responsibility cannot be delegated. The CEO owns the worst problems, not the best ones. Physical presence and shared suffering communicate commitment more powerfully than any memo, which is why Musk literally sleeps on the factory floor. He talks about the ego-to-ability ratio: high ego breaks your reinforcement-learning loop with reality. He frames startups as “eating glass and staring into the abyss,” where glass is the work you do not want to do and the abyss is the constant threat of company death. He says adversity is the only forge that produces the pain threshold required to run a hard company at scale.

    The teams chapter is uncompromising. The most important job of a leader is attracting exceptional people. Money is not the constraint; exceptional talent is. He runs a Special Forces hiring model: the minimum passing grade is excellence. A small group of technically strong people will always outperform a large group of moderately strong people. Character matters as much as skill, because skills are teachable and attitude is not. The feedback discipline he insists on is hardcore: “All bad news should be given loudly and often. Good news can be said quietly and once.” Camaraderie is dangerous when it suppresses truth. “It’s not your job to make people on your team love you. In fact, that’s counterproductive.”

    The organization-design chapter teaches three rules. First, structure shows up in the product. Silos produce redundancy, waste, and error. Second, communication should travel the shortest path that solves the problem, not the chain of command. Anyone should be able to talk to anyone. Third, jargon and acronyms are cognitive pollution; the test for any internal phrase is whether a new hire would understand it cold. This is the chapter that introduces The Algorithm (covered in depth above).

    Musk runs his companies on what he calls a “maniacal sense of urgency.” The only true currency is time. Speed is both offense (faster innovation than competitors can copy) and defense (the SR-71 Blackbird has almost no defense system except acceleration). The protection of real intellectual property is not patents but rate of innovation; if you ship faster than anyone can copy, you do not need legal moats. He stresses parallelization over serialization. “A factory moving at twice the speed of another factory is basically equivalent to two factories.” Be a vector, not a scalar: high speed in the right direction, with continuous course corrections like a guided missile.

    The Part II close is “We Must Make Stuff.” Manufacturing is underrated and design is overrated. “There is 1,000 percent, maybe 10,000 percent more work that goes into the production system than the product itself.” The factory is the product, not the car. Designing a rocket is trivial compared to making one that reaches orbit. The production line moves at the speed of its slowest, least lucky part. Out of 10,000 things that have to go right, the one that is not working sets the rate. Manufacturing combined with scale becomes the moat. The gigacast machine story illustrates this perfectly: Musk got the idea from toy cars, asked if any law of physics prevented it, surveyed six casting-machine suppliers, five said no, the sixth said maybe, and Tesla used that single innovation to cut the body shop by 30 percent.

    Part III: Building Zip2, PayPal, Tesla, and SpaceX

    Musk left Stanford grad school in 1995 with $110K in debt and founded Zip2 with his brother Kimbal, starting with $2,000 and one computer in a squatted office where he slept on a futon and showered at the YMCA. In 1999, Compaq acquired Zip2 for over $300 million. His after-tax bank account went from $5,000 to $21 million. He immediately rolled $12.5 million of that into X.com, which merged with Confinity in March 2000 to become PayPal. PayPal reached 100,000 customers in its first month and one million by year two with no sales force and no marketing spend. The product traction came from email payments, not from the conglomerate financial-services pitch X.com started with. Musk’s lesson: “listen well, correct fast.” He was removed as CEO during his honeymoon trip in early 2002 but did not contest it, prioritizing company survival over personal vindication. eBay acquired PayPal in October 2002 for $4.5 billion. “Life is too short for long-term grudges.”

    Tesla started in 2003. The original Roadster used a Lotus Elise chassis; the modification added 40 percent weight and invalidated the crash tests. Only 7 percent of Roadster parts ended up shared with the Elise. Musk’s lesson: start clean-sheet, do not modify legacy platforms. The Tesla Master Plan (August 2006) was the sequencing logic: (1) build a sports car, (2) use the profits to build an affordable car, (3) use those profits to build a mass-market car, (4) provide zero-emission power generation. This sequence was forced by the unit economics of new technology, where you cannot start at the bottom of the market without scale.

    Tesla nearly died at the end of 2008. The SolarCity Morgan Stanley deal had collapsed. Tesla and SpaceX were both on the brink. Musk had moved into Jeff Skoll’s guest bedroom because he had no house. The final emergency funding round closed at 6 p.m. on Christmas Eve, hours before payroll would have bounced. Daimler arrived shortly after; Musk’s team rapidly dropped a Tesla powertrain into a Smart Car and got it to 60 mph in 4 seconds, which shocked Daimler into a $50 million investment. Tesla then survived three years of Model 3 manufacturing hell from 2017 to 2019, during which Musk lived in the Fremont and Nevada factories, slept on the floor, and ran around fixing the line. “The longest period of excruciating pain in my life.” His pricing philosophy is “give people more for less”: spend money on engineering and design instead of advertising, and let the product carry word of mouth.

    SpaceX was founded in mid-2002 with $100 million of Musk’s PayPal proceeds. He expected to lose everything; that was his stated expectation going in. There was no external funding for three years. His initial plan was a $90 million Mars greenhouse mission designed to inspire NASA, but he realized the binding constraint was launch cost, not mission design. He tried to buy Russian ICBMs to cut launch costs; that failed. He then ran the first-principles rocket cost analysis, found that finished cost was 50 to 100 times raw-material cost, and concluded the industry’s pricing was a function of cost-plus contracting, five-layer subcontracting, and legacy tech. He budgeted for exactly three failed Falcon 1 launches. Launches 1, 2, and 3 failed (2006, 2007, 2008). Launch 4 succeeded in August 2008. Days later NASA awarded SpaceX a $1.6 billion cargo resupply contract. Musk reportedly screamed “I LOVE NASA. YOU GUYS ROCK.” The fourth-launch success and the NASA call together saved both SpaceX and (indirectly, via Musk’s bank account) Tesla.

    SpaceX’s actual optimization target is “fastest time to a self-sustaining city on Mars.” That goal cascades to “fastest time to a fully usable rocket,” which cascades to “fastest time to orbit.” Early Starship had no doors because doors are not necessary for reaching orbit. The unifying engineering insight is that full and rapid reusability is the holy grail of rocketry, because once a rocket is reusable, the only marginal cost is propellant (mostly liquid oxygen and methane, around $1 million per Starship flight). Current cost per landed ton to Mars is about $1 billion. Starship targets less than $100,000 per ton, a 10,000x improvement. Musk’s philosophy on testing reflects the design constraint: unmanned rockets should be allowed to blow up so the team can learn; crewed systems get extreme conservatism. The Space Shuttle’s safety record suffered precisely because the asymmetry of risk made the program incapable of iteration.

    Part IV: The Age of Abundance, the seven risks, and Mars

    Musk frames his companies as philanthropy, defined by reality rather than perception. “If you care about the reality of goodness instead of the perception of it, philanthropy is extremely difficult.” Companies create durable wealth because they solve real problems at scale, distribute knowledge through products, and deploy capital toward problems rather than store it idle. The companies he names as worth starting today: tunneling (Boring Company), synthetic-RNA medicine (“the digitization of medicine”), and high-speed transport such as Hyperloop (a pressurized electric vehicle in a vacuum tube, faster than aircraft, weather-independent).

    The Age of Abundance chapter argues that AI plus humanoid robotics will eventually remove labor as the binding economic constraint, producing abundance for everyone. Humanoid robots will start in dangerous and repetitive jobs and eventually outnumber humans 2 to 10 to one at less than the cost of a car. Tesla’s full self-driving and Robotaxi will, in Musk’s projection, make Tesla a $10 trillion company because autonomous cars are worth 5 to 10x non-autonomous cars (they earn revenue when owners are not using them). Neuralink achieved 2 bits per second of brain output with first patient Noland Arbaugh; the 5-year target is one megabit per second. Long-term Neuralink applications include consensual telepathy between two BCIs, vision restoration (Blindsight), and multispectral senses. Musk’s framing: humans are already cyborgs through phones and laptops, but the bandwidth to those devices is “poking glass with your meat sticks” and BCIs are the next bandwidth jump.

    The Existential Risks chapter names seven specific risks. World War III: the cycle of major-power war recurs and global thermonuclear conflict could end or maim civilization. Regulation accumulation: laws never die when humans do, regulations compound forever, and eventually everything becomes illegal. California High-Speed Rail is the example: after billions of dollars, it is “almost illegal to build.” Wars historically cleared regulatory cobwebs; peacetime allows infinite accumulation. Unsustainable energy: regardless of climate, hydrocarbons are finite, so the transition must happen. Nuclear plants should not be shut down (coal is 100 to 1,000x worse for health than nuclear). The energy mix is solar plus wind plus batteries plus nuclear plus hydro plus geothermal. Misaligned artificial superintelligence: AI is growing faster than any prior technology, and Musk considers it “a significantly higher risk than nuclear weapons.” The specific mitigation he names is rigorous truth adherence in training. The HAL 9000 lesson from 2001 is that an AI forced to lie becomes dangerous; he cites the Gemini “George Washington wasn’t white” failure as a concrete example of ideological training producing catastrophic outputs at scale. Population collapse: low birth rates are a slow civilizational death. The US has been below replacement since the early 1970s. China is 40 percent below replacement; the three-child policy failed. “We need to revive the idea of having children as a social duty.” Musk himself has 12 children across three women. Asteroids and comets: Earth has no defense against a large comet; Starship gives some capability against small asteroids. Shoemaker-Levy left an Earth-sized hole in Jupiter, and that level of impact on Earth is “game over.” Civilizational fragility itself: every prior civilization fell, and Stephen Hawking estimated roughly 1 percent probability of civilizational end per century. “That’s Russian roulette where 99 barrels are empty. Every century is a click.”

    The closing chapter, Becoming Multiplanetary, places Mars colonization in evolutionary context. Earth has had six milestones in 4 billion years: single-celled life, multicellular life, plants and animals, ocean-to-land transition, consciousness, and (potentially) multiplanetary life. Musk argues this last step is “at least as important as life going from the oceans to land, probably more significant,” because it makes the substrate of consciousness redundant. Sun expansion will destroy Earth in roughly 500 million years; meanwhile self-inflicted or external extinction events are recurring, with five major mass extinctions already in the fossil record and Yellowstone erupting roughly every 700,000 years. The plan: produce 1,000 Starships per year, refuel in orbit, hit 10,000 missions and 1 million tons to Mars by approximately 2044, then build out a self-sustaining city. Mars trips depart in 2-year windows when planets align; Musk’s working schedule is 5 uncrewed missions in 2026 and crewed missions in 2028 if the uncrewed go well (otherwise +2 years). For terraforming, his named options are thousands of solar reflectors in orbit or thermonuclear detonations over the polar caps as “two little suns” to vaporize CO2 ice, thicken the atmosphere, and eventually produce liquid water oceans roughly a mile deep covering 40 percent of the planet. Cost of the entire civilization-insurance bet: less than 1 percent of Earth GDP.

    The 69 Core Musk Methods

    The bonus section compresses the entire book into 69 short maxims, intended as a copy-able reference. They are reproduced here near-verbatim.

    1. You are capable of more than you think.
    2. It is possible for ordinary people to choose to be extraordinary.
    3. You can teach yourself anything. Read widely. Talk to experts.
    4. Assume you are wrong. Aspire to be less wrong.
    5. Internalize responsibility.
    6. If we don’t make stuff, there is no stuff.
    7. Creating products and services creates wealth.
    8. A useful life is worth having lived.
    9. Don’t aspire to glory. Aspire to work.
    10. Take actions that increase the odds of the future being good.
    11. Every day, you either increase the rate of innovation or it slows down.
    12. Work on what is just becoming possible.
    13. Don’t wait for the world to want it. If it should obviously exist, go build it.
    14. Build what no one else is building.
    15. As you move forward, allies will assemble around you.
    16. Prototypes are proof.
    17. Start somewhere. Question assumptions. Adapt to reality.
    18. Reason from fundamentals, not from what others are doing.
    19. The magic-wand number. See the theoretically perfect and work toward it.
    20. Know the idiot index. Understand the cost of components.
    21. The Algorithm: Question Requirements, then Try to Delete, then Simplify, then Accelerate, then Automate.
    22. For critical items, run The Algorithm in 24-hour meetings to check progress.
    23. Stay as close to the actual work as possible. Do not separate yourself from the pain of your decisions.
    24. All requirements should be treated as recommendations.
    25. The only fixed laws are the laws of physics.
    26. The best part is no part. The best process is no process.
    27. Simplicity creates both reliability and low cost.
    28. Find the design necessity of every part and every process.
    29. Overdelete. Add back only the absolutely necessary.
    30. Push for radical breakthroughs.
    31. Be proactive. You will never win unless you take charge of setting the strategy.
    32. A maniacal sense of urgency is the operating principle.
    33. A factory moving at twice the speed of another factory is basically equivalent to two factories.
    34. Attack the bottleneck. The one thing that isn’t working sets the overall production rate.
    35. You’ll move as fast as your least-lucky or least-competent supplier.
    36. Do things in parallel.
    37. Give teams one key metric to focus on. Video games without a score are boring.
    38. Separating design, engineering, and manufacturing is a recipe for dysfunction.
    39. Speed of innovation is what matters.
    40. Beat competitors on speed, quality, and cost. Not anti-competitive behavior.
    41. Test the absurd. When something seems impossible, ask “what would it take.”
    42. Money is not the constraint. Exceptional engineers are.
    43. Get everyone thinking like the chief engineer.
    44. Get a clear, direct feedback loop with reality.
    45. Always be smashing your ego. Ensure ability is greater than ego.
    46. Ask “is this effort resulting in a better product or service.” If not, stop.
    47. Good taste is learnable. Train yourself to notice what makes something beautiful.
    48. Physics doesn’t care about hurt feelings. Make the rocket fly.
    49. Empathy is not an asset.
    50. Use simple, clear, humble terms.
    51. Go directly to the source of information.
    52. When hiring, look for evidence of exceptional ability.
    53. Combine engineering and financial fluency.
    54. To truly lead the product, lead the company.
    55. Lead from the front. Sleep on the factory floor.
    56. Physically move yourself to wherever the problem is. Immediately.
    57. All bad news should be given loudly and often. Good news can be said quietly and once.
    58. Failure is essentially irrelevant unless it is catastrophic.
    59. Fear of failure is the biggest cause of failure.
    60. Feel the fear and do it anyway.
    61. Double down. Push your chips back in.
    62. Work like hell. Every waking hour. Go ultra hardcore.
    63. Make sure you really care about what you’re doing, and take the pain.
    64. We should not be afraid of doing something important just because tragedy is possible.
    65. When something is important enough, do it even if the odds are not in your favor.
    66. Don’t ever give up. You’d have to be dead or completely incapacitated.
    67. Play life like a game.
    68. Go ultra hardcore.
    69. Humor is a differentiator.

    Thoughts

    The most underrated artifact in the book is The Algorithm, and the reason it is underrated is that it looks deceptively simple. Five steps. Anyone can recite them. Almost nobody runs them in order. The book’s central operational insight is that the sequencing is the whole game. People skip step one because it is uncomfortable to confront the fact that requirements they have spent years optimizing against came from somebody whose name they cannot remember. They skip step two because deletion creates risk that materializes immediately and the benefits show up later. They jump to step three because optimization feels like progress and is graded well in school. Then they jump to step five because automation looks impressive on a dashboard. Tesla’s $2M robotics step on the fiberglass mat would never have existed had the team run the steps in order. Most companies, at any scale, are sitting on enormous unrealized value the same way Tesla was, locked behind the simple act of asking “what is this part actually for, who told us we needed it, and would anything bad happen if we deleted it.”

    The second insight worth sitting with is the magic-wand number paired with the idiot index. These two ratios together turn first-principles thinking from a vague aspiration into an operational discipline. Any product you can buy or any process you run has a raw-material cost (the magic-wand number, the absolute floor) and a finished cost. The ratio between them tells you the upper bound on how much you can improve. A high idiot index is not a moral failing of the supplier; it is an unpriced opportunity that competition has not yet found. Once you train yourself to ask these two questions about every line item, the world rearranges. Rockets that cost 50x their steel become a problem to solve. Tunnels that cost a billion dollars per mile become an obvious target. Battery cells that cost 7.5x their materials become a startup. The discipline is not “be smart.” The discipline is “calculate both numbers.”

    The third theme is what the book calls “manufacturing is the moat,” and it is the part of Musk’s playbook that most observers, including most of his competitors, still underestimate. The book’s claim is not that design is unimportant. The claim is that production is between 1,000 and 10,000 percent more effort than design, and that nobody outside of practitioners understands the asymmetry. This is why Toyota and Daimler buy electric powertrains from Tesla rather than make them. It is why SpaceX spent 10 to 100 times more engineering on the Raptor manufacturing system than on the Raptor engine. It is why Apple’s contract manufacturers, not its designers, are the durable moat. The same logic now applies to AI infrastructure: the supercluster, the cooling, the power smoothing, the cabling at 3 a.m., the Megapack buffers, are the actual moat, and the model architecture is the visible-but-cheaper layer on top.

    The fourth theme is the way responsibility, ego, and feedback interact in Musk’s organizations. Most CEOs are insulated from the consequences of their decisions by layers of process and middle management. The result is a high ego-to-ability ratio, because the feedback loop between the ego’s prediction and reality’s response is intermediated to the point of uselessness. Musk’s defense is physical: sleep where the work happens, walk the factory floor at 3 a.m., personally answer the questions, run cabling himself if necessary. This is not theater. The epistemic claim is that decisions made by an insulated CEO are systematically worse than decisions made by a CEO whose body is in the same room as the problem. The cost is severe in personal terms (“the longest period of excruciating pain in my life”), but the alternative is making confident decisions on a model of reality that has drifted out of alignment with the actual machine. The same logic applies to engineers who do not see their designs in production, founders who do not talk to customers, and leaders who delegate the worst problems to people they did not pick.

    The fifth theme is the seven existential risks and why Mars sits at the center of them. The book’s framing is that any single risk is small, but compounded across centuries the probability of civilizational discontinuity is large. Hawking’s 1-percent-per-century estimate, repeated for 10 centuries, gives roughly a 10 percent cumulative probability. Across the timescales humanity has already survived, those odds are unacceptable for a species that can afford a backup. The Mars argument is not romanticism. It is a 1-percent-of-GDP insurance premium on the persistence of consciousness itself. The other six risks (war, regulation accumulation, energy exhaustion, misaligned AI, population collapse, asteroids) are presented in the same actuarial frame: each is independently survivable, but the cost of treating them as low-probability is precisely the cost a previous civilization paid by treating its own near-misses as low-probability until the one near-miss that wasn’t. The most uncomfortable specific risk in the book is population collapse, which is the only one where doing nothing is doing the wrong thing and where the demographic numbers are already locked in for decades regardless of policy response.

    The sixth and final point is the book’s underlying claim, which is also Naval’s claim in the foreword: Musk’s methods are copy-able. The book exists because Jorgenson believes that one million Musks would change the trajectory of the species. The 69 Core Musk Methods are not a personality cult. They are a starter kit. Most people will not pick the same problems, will not have the same tolerance for pain, and will not run the same companies, but anyone can apply The Algorithm to their own work, calculate the idiot index on their own product, demand requirements come from named people, attack the bottleneck on their own line, refuse to automate before deleting, and pick a problem that is on the path to the future. The book is best read as a manual, not a biography. If it ends up next to your laptop and you reread The Algorithm chapter every six months and the 69 Methods every quarter, that is the use Eric and Naval intended.

    Get The Book of Elon by Eric Jorgenson at elonmuskbook.org or wherever you buy books.

  • Shopify CEO Tobi Lütke: AI Is the Perfect Scapegoat for Layoffs, Canada Has Trump Derangement Syndrome, and 50% of Shopify Code Is Now AI-Generated

    TLDW

    Shopify CEO Tobi Lütke sat down with Harry Stebbings on 20VC for one of the most candid and controversial conversations of his career. Lütke argues that the current wave of mass layoffs has nothing to do with AI and everything to do with pandemic-era overhiring, but AI will be blamed because it cannot fight back. He blasts Canada for its “Trump Derangement Syndrome,” calls the climate cult “one of the most evil things wrought on the population,” reveals that over 50% of Shopify’s code is now AI-generated, and says many of his best engineers have not written a line of code since December when Claude Opus changed everything. He also introduces River, an AI engineer at Shopify that named itself, and explains why he believes context engineering will be the dominant role of the next five years.

    Key Takeaways

    • AI is not causing layoffs, COVID overhiring is. Lütke is blunt: “What you see right now is not AI layoffs. Those are just the companies that are really slow that overhired just like everyone else.” AI will get blamed for everything because it is the perfect Girardian scapegoat that cannot fight back.
    • Over 50% of Shopify’s code is now AI-generated and “converting to much higher numbers.” Many of Shopify’s best engineers have not written code this year. December 2025 and the release of Claude Opus changed everything.
    • Senior engineers became more valuable, not less. Lütke initially thought new grads with no priors would dominate the AI native era. He was wrong. Senior engineers steer agents better because steering is the new programming, and reps matter more than ever.
    • Context engineering will become the dominant role within 5 years. A new product builder role is emerging that subsumes engineering, design, and product management, focused on coordinating intelligent actors (humans and AI) to ship products.
    • “River” is Shopify’s AI engineer that named itself. Built first, then asked what name it wanted. River lives in Slack, ships engineering work, and learns publicly because it is steered through public Slack channels.
    • Builders are “eights” on the Enneagram and companies actively conspire against them. Eights call out nonsense, refuse fancy dressing, and are dangerous to colleagues’ careers. They rarely get promoted, often leave, and start companies. Shopify is “remarkably high on eights” because Lütke seeks them out.
    • Canada has “Trump Derangement Syndrome.” Over 60% of Canadians believe the United States is a bigger threat than Russia or China. Lütke calls this “stunning” and wrong. Canada’s only winning strategy historically has been “winning by helping America win.”
    • Canada should be the richest country on Earth. It has every resource the world needs for the next 20 years. Lütke wants pipelines built, industry built, refining done domestically, and an end to exporting raw resources to have other countries make end products.
    • Be deeply suspicious of “non-profit.” Lütke argues opting out of the only fitness function that has ever pulled people out of poverty (markets) and refusing to disclose your actual fitness function is a red flag. Non-profits replace merit with pull.
    • The climate cult is blocking civilization. Lütke called it “one of the most evil things wrought on the population” and pointed to anti-nuclear green parties and frog protection laws blocking factories as examples of policy capture.
    • The Chinese AI threat is real but misunderstood. The bigger concern is that if Western governments restrict children from using AI, kids will simply download Chinese open-weight models, train on collectivist worldviews, and stop ever writing high school essays about Tiananmen Square.
    • Markets are the most democratic system that exists. Every dollar spent is a vote. Capital allocation by hundreds of millions of consumers is more democratic than any election.
    • Friedrich List and the Prussian school over Adam Smith. Lütke prefers a model where governments define excellent games with positive externalities, then completely get out of the way and let competition do the rest.
    • Shopify’s biggest mistake was going into physical logistics right before AI got really good. Lütke initially defended the decision based on what he knew at the time, but later admitted he was probably just wrong.
    • Lütke does not look at the stock price. It has been at least 23 days since he last checked. He runs Shopify on product instincts, not market signals.
    • Great leaders must be exothermic. A CEO is a heat source for the company. Lütke prefers “temperature” to “chaos” because chaos has too negative a connotation.
    • Don’t go to university for university’s sake. Get a degree from somewhere hard to get into so you are surrounded by people who also fought to get in. Better yet, join a small company where you can actually be of value.
    • Entrepreneurship is the most AI-safe AND most AI-benefiting job. Lütke sees a coming golden age of entrepreneurship where priors no longer matter and AI co-founders eliminate the need to grow up around business.
    • “You can just do things” is the rallying cry Lütke wants to ingrain in the world. Action causes information. The cost of trying is lower than ever.
    • The demonization of wealth in America is misdirected. No one gets to a billion dollars by stealing. Builders create products that people vote for with their money, the most democratic act in any economy.

    Detailed Summary

    Harry Stebbings opens by asking Tobi Lütke whether entrepreneurs are motivated by fear of losing or hunger to win. Lütke says he is still figuring out his own answer, but argues that both extremes lead to short-term thinking. The real unlock is taking a long perspective, because compound advantages only accrue when you are willing to wait.

    Builders Are “Eights” and Companies Conspire Against Them

    Lütke explains the Enneagram personality framework and identifies himself as an “eight,” the type that refuses to accept that any organization’s output is acceptable just because it is dressed up nicely. Eights call out nonsense, are dangerous to careers around them, rarely get promoted in professionally managed companies, and often leave to start their own businesses. Shopify deliberately overweights eights in its hiring. Lütke also says people who build companies are “fundamentally crazy people” and that the public image of leadership comes from movies, not reality. He never wanted to be CEO but realized you cannot run a product driven company without controlling the company itself, because product needs and company needs only converge on a three-year horizon.

    The Luxury of Long-Term Thinking as a Public Company

    Stebbings asks if a public company can really afford long-term thinking. Lütke says trusted public companies are the best position to be in. The chasm to cross is from trusted private to untrusted public, which is why so many founders refuse to IPO. Shopify went public 11 years ago at a 1.67 billion dollar valuation when revenues were a fraction of today’s. The valuation is now roughly 100x higher. Lütke walks through the IPO mechanics: investment bankers serve the buy side, not the company, and Lütke priced his offering above range because he knew where his growth would come from. The first trade closed about 10 dollars higher, which he calls a “good performance” but a teaching moment about market price discovery.

    AI Is the Perfect Scapegoat for Mass Layoffs

    This is where the conversation gets explosive. Lütke says Shopify employs about 7,500 to 8,000 people today and his real hope is to have the same number in five years, but at 100x productivity. He argues that the layoffs sweeping the tech industry have nothing to do with AI. They are the result of pandemic-era overhiring catching up to slow-moving companies. But AI will get blamed for everything because it is the perfect Girardian scapegoat. It cannot defend itself, it has no PR team, and an entire industry of doomers is already trained to point at it. Lütke says his own industry has been “gaslighting everyone into AI fear” and science fiction did the same for 60 years before that.

    His own use of AI is what he calls utopian. Tasks that used to be hard are easy. Most jobs, he argues, are not actually good jobs to begin with. Being a human task queue is not a great job. Great jobs involve agency and creation. As AI gets cheaper, purchasing power explodes, and people will get options to do things on weekends that are vastly more productive than their day jobs ever were.

    Markets Are the Most Democratic Mechanism Ever Invented

    Lütke pivots into a long defense of capitalism as the most democratic system in existence. Every dollar spent is a vote, far more frequent and more granular than any election. He uses Elon Musk and Tesla as examples. Lütke owns a Model Y, did not touch the steering wheel that morning, and uses Starlink in the back to work on long drives. He posts on X and gets replies from Japan in real time. He calls Musk a “one man engine” who has captured a tiny percentage of the value he created. He extends this to Shopify itself: Lütke owns 6% of the company, which means 94% is owned by other people who all made money. Plus roughly 10 million people work in the broader Shopify ecosystem on customer fulfillment, web design, customer service, and more.

    Why “Non-Profit” Should Make You Suspicious

    Lütke targets the charity industrial complex. He argues that non-profits opt out of the only mechanism humanity has ever invented to lift people out of poverty (markets), and they fail to articulate what their actual fitness function is. The result is that “merit of organization is replaced with pull of individuals.” Smooth talkers, not builders, end up running these institutions. He acknowledges Carnegie’s libraries and a few exceptions but believes the ratio of charity dollars to good outcomes is dramatically off. He is far more enthusiastic about funders like MacKenzie Scott who give in unrestricted ways, and even more enthusiastic about Jensen Huang and Bloom Energy as compute and infrastructure investments that compound into civilizational gains.

    The Prussian School of Economics

    Asked about government intervention, Lütke pledges allegiance to Friedrich List and the Prussian school of political economy over Adam Smith and Lassalle. The job of government is to define excellent games where positive externalities accrue to society, then completely get out of the way. He calls the outsourcing of violence to governments “one of the most inspiring things humanity has ever done” because it created the conditions for personal property. But governments are extremely bad at doing things directly. The moment a government runs grocery stores, it costs 10x more, and entrepreneurs have to be enlisted to repair the damage.

    Canada’s Trump Derangement Syndrome

    Stebbings asks if Lütke is proud of Canadian Prime Minister Mark Carney for standing up to Trump. Lütke is unequivocal: no. He calls Carney’s stance “not a credible witness to the reality on the ground.” Canadians, he argues, are “massively overfit to niceness,” which leads to “unkind lies” and lying by omission. Over 60% of Canadians now believe the United States is a bigger threat than Russia or China, which Lütke calls “stunning” and clearly wrong. Canada is a small economy attached to a hegemon, and the only winning strategy in its history has been winning by helping America win.

    That said, he agrees with Carney on diversifying the economy, getting closer to Europe, and engaging Asia. But he wants Canada to also “build the [expletive] out of pipelines, build the [expletive] out of our industry, and start refining the stuff ourselves.” Canada has every resource the world needs for the next 20 years and the most educated workforce on Earth. The only obstacle is political will. Canada’s commercial story has been the same since the beaver pelt era: extract resources, ship them abroad, let other countries make end products. Canada Goose, Lululemon, Shopify, Miller Lite. That is the short list of products Canada actually makes.

    The Real Chinese Threat

    Lütke says the Chinese AI threat is both underestimated and overestimated. The bigger threat, he argues, is government overreach. If Western governments start dictating which AI models children can use, kids will simply download Chinese open-weight models. He notes that Chinese models, especially when prompted in Chinese, exhibit a clearly collectivist worldview. The risk is that an entire generation of students writes essays through models trained never to mention Tiananmen Square. He frames the broader political battle as collectivism versus individualism and says everything else is smoke screening.

    Fixing Europe and the Climate Cult

    Asked what he would do as president of Europe, Lütke begins by saying you have to “get rid of the climate cult.” He calls it “one of the most evil things wrought on the population,” citing green parties whose founding myth is that nuclear power is bad, and infrastructure projects blocked because of one frog breeding in one creek. He argues that very few people have the capability to truly build, and they need both enablement and accountability from the village. Beyond that, he wants Europe to follow the Prussian playbook: build excellent games, build infrastructure, and use the resulting wealth to sculpt the economy you want.

    Shopify’s Biggest Mistake

    Lütke says his biggest public mistake was Shopify’s full push into physical logistics and warehousing right before AI capabilities exploded. Initially he defended the decision as correct based on the information available at the time, but later admitted he probably just got it wrong. The hardest part was that real people lost their jobs when Shopify exited.

    Great Leaders Are a Heat Source

    Lütke previously talked about CEOs injecting “chaos” into organizations. He now prefers “temperature.” Heat is atoms jiggling. Great leaders must be exothermic, providing energy that flows through the organization. He says he hasn’t checked Shopify’s stock price in at least 23 days. Most public company CEOs are obsessed with their stock. Lütke runs on product instincts.

    Senior Engineers Don’t Write Code Anymore

    Lütke admits he was wrong about new grads having an AI native advantage. Some are exceptional (he hired a 13-year-old intern from Waterloo whose mother accompanies him to classes), but on the whole, senior engineers steer agents better than juniors do because they have done more reps. Programming is not gone. Programming has become higher level. Engineers massively underestimate how important steering is. Steering is just programming at a higher altitude.

    The Role That Will Dominate in 5 Years

    Lütke says context engineering, a term he had a hand in popularizing, will become a standard role within five years. It will likely subsume parts of product, design, and engineering management. The best AI programmers right now, surprisingly, are people from engineering management because they have been prompting intelligent agents (humans) for years. Good communicators are good thinkers because communication is distillation.

    River, the AI Engineer That Named Itself

    Shopify built an AI engineer that lives in Slack. They built it first, then asked it what name it wanted. The AI chose “River” because Shopify’s monolithic repository is called “world” and rivers shape worlds. River does an enormous amount of Shopify’s engineering, taking instructions through public Slack channels so that the entire company can learn from how others steer it.

    Over 50% of Shopify’s Code Is AI-Generated

    The number is “a fair deal over 50%” and “converting to much higher.” Many of Shopify’s best engineers have not written code this year, with the inflection point being December 2025 and the release of Claude Opus. Lütke himself still writes code occasionally, especially the data structure layer where he applies what he calls a “German school” of engineering: figure out how data persists on disk, then build everything else on top. Once that is right, the rest can be vibe coded by AI.

    Should His Kids Go to University?

    Lütke says he would not push his kids to attend university for its own sake. The value of a hard to enter program is being surrounded by people who also fought to get in. Better still: get into the room with people who are obsessed with the topic you care about. He thinks joining a small startup where you can actually be of value is often a superior path. He addresses nepotism directly. His instinct is that nepotism is bad. The gold standard is double-blind merit. But double-blind merit barely exists anywhere, and intersectional academic hiring criteria in Canada are arguably worse than nepotism.

    Final Reflections

    Lütke ends with what he calls the best advice he knows: “You can just do things.” The system exists to push everyone toward acceptable outcomes, but if you know what a good outcome looks like, you can step out of the system and try. Action causes information. The cost is lower than ever. The only constraint is that the experiment cannot have victims.

    He also addresses the demonization of wealth. No one gets to a billion dollars by stealing. Builders create products people vote for, the most democratic act there is. Buying from a local shop is voting for the welfare and future of local shops. Constructive criticism is itself something someone has to build, and Lütke welcomes it. Lazy criticism, hot takes, and bad faith arguments are corrosive and should be held in contempt.

    He is bullish on AI as a counterweight to information warfare. A council of AI models trained in different countries (Chinese, German, French, American) could fact check claims with multiple perspectives. The “@grok is this true” reflex on X is, he says, a primordial version of this. The information asymmetry that has favored bad faith actors for decades is about to flip.

    Thoughts

    This interview is a window into the operating philosophy of one of the most successful technical founders alive, and it is far more provocative than most of his public appearances. The headline claim, that AI is a scapegoat for layoffs caused by pandemic overhiring, deserves to be repeated until it sinks in. Every CEO who lays people off and then writes a memo about “AI driven efficiency” is taking advantage of a narrative that AI cannot push back against. The math is plain: if you doubled your headcount in 2021 and 2022 and now you are firing 15%, you are not net displaced by AI. You are correcting a hiring mistake.

    The 50% AI generated code statistic is the bigger story. Shopify is not a small company. 8,000 employees and 7 billion in revenue is enterprise scale. If a company that mature has crossed the 50% threshold and is “converting to much higher numbers,” the implication for the broader software industry is enormous. The senior engineer compounding observation is also subtle and important. If steering is the new programming, then the senior pool is more valuable, not less, and the pipeline problem for junior developers gets harder to solve. Companies that under invested in junior training during ZIRP will face an experience cliff in five years.

    Lütke’s Canadian commentary will offend many readers in his home country, which seems to be exactly the point. The “lying by omission” critique of Canadian niceness is sharp and accurate. The 60%+ of Canadians who view the US as their largest threat is genuinely a remarkable statistic, and it has implications for trade policy, capital flows, and immigration. Whether or not you agree with his political read, his prescription is unambiguous and pro-growth: build pipelines, refine resources domestically, stop being content as a feedstock economy.

    The non-profit critique deserves more public debate. The fitness function point, that markets reveal preferences and non-profits opt out of preference revelation while not disclosing what they optimize for, is a sharp economic argument. The pull versus merit observation about who ends up running large foundations rings true to anyone who has worked adjacent to the philanthropic sector.

    The introduction of River as an AI engineer that named itself is a small detail that signals where this is going. AI agents are going from tools to teammates with identities, channels, and reputations. The fact that River shapes the “world” repository is poetic, and the public Slack steering pattern is a real innovation in how organizations can scale agentic AI without creating siloed knowledge.

    Lütke’s “you can just do things” rallying cry is ultimately what ties the entire interview together. Whether he is talking about Canada, Europe, AI engineers, or his own kids, the through line is the same: action causes information, the cost of trying is lower than ever, and the only people who will benefit from the next decade are the ones who refuse to wait for permission. This is the most useful piece of philosophy in the entire conversation, and it applies far beyond entrepreneurship.

  • Brian Chesky on AI Founder Mode, the 11-Star Experience, and Reinventing Airbnb for the Age of AI

    Airbnb CEO Brian Chesky sits down with Patrick O’Shaughnessy on Invest Like The Best to talk about the next evolution of company building: AI Founder Mode. He covers the shift from founder to CEO, the lessons he learned from Steve Jobs through Hiroki Asai, why consumer AI is the next great frontier, and how he plans to change the atomic unit of Airbnb from a home to a person.

    TLDW

    Brian Chesky believes the next era of company building belongs to founders who refuse to delegate the soul of their company. He coined Founder Mode with Paul Graham after the pandemic forced him to take Airbnb back into his own hands. Now he is shaping what comes next: AI Founder Mode, where leaders work with on-demand context, fewer layers of management, asynchronous communication, and a new generation of hybrid manager-makers. He shares why most software companies have not been touched by AI yet, why consumer AI is about to explode, and why he is rebuilding Airbnb around people, not homes. The conversation also touches on the 11-Star Experience exercise, the power of small teams, why recruiting is the most important job a CEO has, and why every adult is still an artist underneath.

    Key Takeaways

    • Founder Mode is not micromanagement, it is having a steering wheel. Chesky woke up in 2019 feeling like the car had no steering wheel. After the pandemic, he reviewed every detail for two to three years before delegating again. Start hands-on and give ground grudgingly, not the other way around.
    • AI Founder Mode is even more intense. With AI, leaders can be in significantly more details because almost everything is on demand. Expect fewer layers of management, mostly asynchronous work, and the death of the pure people manager.
    • Two types of leaders will not survive AI. Pure people managers who only do one-on-ones, and rigid people who refuse to evolve. Everyone needs to be a hybrid manager-IC who can still touch the work.
    • Manage people through the work, not through meetings. Frank Lloyd Wright did it. Johnny Ive does it. You are not anyone’s therapist.
    • Consumer AI is the next great prize. 159 of the last 175 Y Combinator companies were enterprise. Almost every app on your home screen has not changed since AI arrived. That changes in the next 12 to 24 months.
    • Why consumer AI is hard. No proven business model, mature distribution, trend-chasing investor culture, and the simple fact that consumer is more hits-driven and requires excellence in design, marketing, culture, and press, not just technology and sales.
    • Project Hawaii is the new operating model. A 10 to 12 person Navy SEAL team, hands-on coaching from the CEO, crawl-walk-run-fly. The first project added roughly $200 million in year one and $400 to $500 million in year two.
    • Make the problem as small as possible. Airbnb spent 16 years failing to launch a second hit because it kept trying to scale globally on day one. Now: pilot in one city, expand to 10, then industrialize.
    • It is better to have 100 people love you than a million people sort of like you. Paul Buchheit shipped Gmail only after 100 Googlers loved it. The sample size of intense love is enough to predict mass adoption.
    • The 11-Star Experience is an imagination exercise. Push to absurdity (Elon takes you to space) so a 6 or 7-star experience suddenly seems normal. The gap between 5 and 6 stars is the gap between you and your competitor.
    • Simplicity is distillation, not subtraction. Hiroki Asai, Steve Jobs’s longtime creative director, taught Chesky that great design distills something to its essence. First principles is a design term too.
    • The score takes care of itself. Bill Walsh and John Wooden both taught that you do not focus on winning, you focus on making every input perfect. Wooden spent his first hour with new players teaching them how to put on socks.
    • Industrial design is the original product management. There are no PMs in industrial design. The designer is the PM, working alongside engineers and program managers to design through user journeys.
    • Recruiting is the CEO’s number one job. The more time you spend recruiting, the less time you spend managing, because great people self-manage. Build pipelines, not searches. Start with results, work backwards to people.
    • Co-hire the top 200 people, not just the executive team. Most CEOs hire executives and let them hire their teams. Chesky considers that fatal because most executives cannot hire well without help.
    • Bodybuilding is a metaphor for leadership. If you can change your body, you can change your life. Progressive overload, 1 percent a day, is how compounding works. Start with biology before therapy.
    • Founder-led companies build the deepest moats. Disney is still selling Walt’s playbook 60 years after he died. Apple is still selling Steve’s iPhone. The longer founders stay in founder mode, the more the company can endure when they leave.
    • Software is hyper fast fashion. Hardware ages well. Buildings get patina. Software always looks dated 10 years later. What endures is the community, the brand, the principles, the mission, and the network effect.
    • Apps are dying. Agents are coming. Chesky says we should let go of our attachment to apps because they are not what the future looks like.
    • Airbnb’s atomic unit is changing from a home to a person. Chesky wants to build the most authenticated identity on the internet, the richest preference library, a real-world social graph, and a membership program. Then expand to 50 to 70 verticals on top of that identity.
    • AI shifts attention from consumption to creation. Social media gave you a paintbrush only for opinions. AI gives everyone a real paintbrush and canvas. We are heading into a creative renaissance.
    • Founders are expeditionaries, not visionaries. They put one foot in front of the other and call it a vision later.
    • Detach from accolades. Chesky describes adulation as a cup with a hole in the bottom. Status is a drug. The path to durable creative work is doing it because you love it, the way Walt Disney, Da Vinci, Van Gogh, and Steve Jobs did until the very end.
    • The kindest gift is belief. The best way to activate a person’s potential is to see something in them they do not yet see in themselves.

    Detailed Summary

    From Industrial Design to the CEO Chair

    Chesky studied industrial design at the Rhode Island School of Design. He chose it on instinct after a department head told him industrial designers design everything from a toothbrush to a spaceship. He grew up enchanted by the Reebok Pump, the Game Boy, the Nintendo, and eventually by the late 1990s golden age of Apple. Raymond Loewy, the man who designed Air Force One and an enormous catalog of mid-century consumer products, became a touchstone, but Johnny Ive was the real hero.

    What he loved about industrial design was that it is technical, commercial, and empathetic. A building can win an architecture award and never be leased. A piece of industrial design that does not sell is a failure. So you have to think about manufacturing, distribution, marketing, and most importantly, user journeys. There are no product managers in industrial design. The designer is the PM. That training, he says, prepared him directly for the role of CEO.

    The Pandemic and the Birth of Founder Mode

    Chesky says no one is born a good CEO. People are born good founders. The job of CEO is counterintuitive in almost every direction. Founders are taught to learn by doing, but a CEO who learns by trial and error wastes years unwinding the empires of misfit hires.

    By 2019 he was running a 7,000 person company he no longer recognized. He felt he was driving a car without a steering wheel. He had a dream that he had left Airbnb for ten years and come back to find it had become a giant political bureaucracy. Then he realized he had been there the whole time. The pandemic hit and Airbnb lost 80 percent of its business in eight weeks. He shifted from peacetime to wartime, took control of every detail, worked 100-hour weeks, and reviewed everything for two to three years.

    The vision was never to micromanage forever. The vision was: I need to know what is going on before I can empower anyone. Hire people, audit their work, and only then give ground grudgingly. Most founders do the opposite, which is why they end up with executives building empires they later have to dismantle.

    AI Founder Mode

    Chesky says AI Founder Mode will be even more intense than Founder Mode because nearly everything will be on demand. He used to live in 35 hours of meetings a week to gather information, the same way Steve Jobs ran Apple. He held weekly, biweekly, monthly, and quarterly group reviews with the full chain of command in one room, anyone could speak, and he made the final call after listening last.

    In the AI era, that culture shifts from meetings to asynchronous work. He expects fewer layers of management. He cites the Catholic Church as a 2,000-year-old institution with only four layers and asks why most companies need seven, eight, or nine. Pure people managers will not survive. Every manager will have to be a hybrid IC, an engineer who still codes, a lawyer who still reads case law, a designer who still designs. You manage through the work, not through one-on-ones.

    He is also bullish that AI tooling will become consumer-grade simple very soon. The current tools, including Claude Code and Cowork, are not yet intuitive to the average person, but the economic incentive will force that to change.

    Why Consumer AI Is the Next Great Frontier

    Chesky points out that 159 of the last 175 Y Combinator companies were enterprise. Almost every consumer app on your phone, including Airbnb, has not fundamentally changed since the arrival of AI. He gives four reasons: investors feared ChatGPT would kill consumer companies; consumer AI has no proven business model because subscriptions hit a local max against free Claude and Gemini, ads are off the table for most labs, and e-commerce has been shut down via third-party app removals; distribution is mature; and Silicon Valley culture, while branded as rebellious, is in practice trend-following.

    The deeper reason is simply that consumer is harder. It is hits-driven, requires great design, marketing, culture, press, and you cannot easily start by selling to your dorm-mates the way enterprise YC startups sell to other YC startups. The prize is bigger. The risk is bigger. He predicts a consumer AI renaissance over the next 12 to 24 months.

    Project Hawaii and the Magic of Small Teams

    Inside Airbnb, Chesky tested a new operating model called Project Hawaii. He took 10 to 12 people, designers, engineers, product, and data scientists, treated them like a startup inside the company, and pointed them at one problem: improving the guest funnel. The system is crawl, walk, run, fly. First fix bugs, then add features, then re-imagine flows, then completely reinvent.

    The first team delivered roughly $200 million of internal revenue in year one and $400 to $500 million the next year, eventually contributing more than 600 basis points of conversion improvement on a base of $134 billion in gross sales. Then they took the same system to pricing, then to other problems, then to launching new businesses like Services and Experiences.

    The guiding lesson: make the problem as small as possible. Airbnb launched in one city, New York. Uber in San Francisco. DoorDash in Palo Alto. When Chesky launched Services and Experiences in 100 cities at once last year, it did not work. The fix was to dominate one city, expand to 10, then industrialize. Peter Thiel said it cleanly: better to have a monopoly of a tiny market than a small share of a big market.

    Underneath that is a Paul Buchheit insight Chesky calls the best advice he ever got. It is better to have 100 people love you than a million people sort of like you. Buchheit refused to ship Gmail until 100 Googlers loved it, and that took two years. Once 100 people loved it, 100 million people did.

    The Hiroki Asai Lessons: Simplicity and Craft

    Hiroki Asai, Steve Jobs’s quietly legendary creative director, taught Chesky two principles. The first is that simplicity is not removing things, simplicity is distillation, understanding something so deeply that you can express its essence. Steve Jobs called design the fundamental soul of a man-made creation that reveals itself through subsequent layers. Elon Musk’s first principles thinking is the same idea applied to physics.

    The second is craft. How you do anything is how you do everything. Chesky cites Bill Walsh’s The Score Takes Care of Itself and John Wooden’s first hour with UCLA players, an hour spent teaching them how to put on their socks. Walsh said the way you tucked your jersey was one of 10,000 details that decided whether you won. The lesson is to focus on getting every input right. The output follows.

    The 11-Star Experience

    The 11-Star Experience is one of Chesky’s most copied frameworks. Most Airbnb stays get five stars because anything else means something went wrong. So Chesky asked: what would six stars look like? Your favorite wine on the table, fruit, snacks, a handwritten card. Seven stars? A limousine at the airport and the surfboard waiting for you because they know you surf. Eight stars? An elephant and a parade in your honor. Nine stars, the Beatles arrive in 1964 with 5,000 screaming fans. Ten stars, Elon Musk takes you to space.

    The point is the absurdity. By imagining the impossible, six and seven star experiences stop seeming crazy. The gap between five and six stars is the gap between you and your competitor. If you can industrialize a sixth star, you may have product-market fit. The exercise also restarts your imagination, which Patrick noted has atrophied for many people in the era of consumption-only social media.

    AI as a Canvas for Creativity

    Chesky frames AI as the ultimate platform shift, the ultimate creative expression, and possibly the greatest invention in human history. Social media made us mostly consumers and gave creators only opinion-shaped tools. AI gives everyone a paintbrush. He believes far more people are creative than we recognize because most have never had craftsmanship or tools to express what is in their heads. Pablo Picasso said all children are born artists; the problem is to remain one as you grow up. Chesky thinks every adult is still an artist underneath.

    The Next Chapter of Airbnb

    Chesky describes four phases of the CEO journey: get to product-market fit, scale to hyper-growth, become a real profitable public company, and finally reinvent. Airbnb’s stock has been flat because the core idea is saturating. He is now squarely in phase four, with three priorities.

    First, change the atomic unit from a home to a person. He wants Airbnb to build the most authenticated identity on the internet, the richest preference library, a real-world social graph, and a membership program. Proof of personhood, he says, will be enormously valuable in the AI age. Second, industrialize the new-business engine to support 50 to 70 verticals (homes, experiences, services, eventually flights, and more) all built on top of that personal atomic unit. Third, navigate the AI transition without breaking the existing business or the livelihoods of hosts. He is also exploring sandbox apps that imagine a radically different Airbnb, the answer to “what is after Airbnb?”

    What Endures in the Age of AI

    Chesky is direct that software does not endure. Look at any software from 10 years ago and it looks dated. Hardware ages better. Buildings develop patina. Paris endures. So if you want to build something lasting, you cannot bet on the app. You have to bet on the community, the brand, the mission, the principles, the identity, and the network effect. Apps are going away, replaced by agents. Founders attached to apps need to let go.

    Founder-Led Moats: Disney and the Ham Sandwich Paradox

    Chesky reconciles Warren Buffett’s “buy a company a ham sandwich could run” with the venture capital truth that a founder’s ceiling is the company’s ceiling. The reconciliation is Disney. Most people cannot name a Paramount, Warner Brothers, Universal, or MGM film off the top of their head, but everyone can name Disney films. Walt Disney was a founder in founder mode for so long that he created enough IP and momentum that the company has been running on his playbook for 60 years after his death. Apple is similar with Steve Jobs and the iPhone.

    The counterintuitive lesson: if you want a company to last 100 years, do not delegate early to make it independent of you. Stay in founder mode for as long as possible so you can institutionalize the magic deeply enough that it endures after you. Tech is the industry of change, so founder mode matters even more there than in chocolate or insurance.

    Bodybuilding as Leadership Training

    Chesky was a 135-pound late bloomer who told his friends he would compete at the national level in bodybuilding by 19. He did. Two lessons came out of it. First, if you can change your body, you can change your life. Start with biology before therapy. Second, you cannot get in shape in one day. Progressive overload, discipline, consistency, and roughly 1 percent a day compound into massive gains. The visible feedback loop in bodybuilding taught him to break invisible problems (like the quality of a leadership team) into observable, measurable proxies (like the quality of the room at a twice-yearly roadmap review of the top 100 people).

    Recruiting as the CEO’s Number One Job

    Sam Altman told a 27-year-old Chesky he would spend 50 percent of his time on hiring. Chesky did not, and considers that his biggest mistake. He now starts and ends every day with his recruiter and spends two to three hours a day on hiring. The more time you spend recruiting, the less time you have to spend managing because great people self-manage.

    His system is pipeline recruiting, not search recruiting. He never starts with a search firm. He constantly meets the best people in their fields, asks each one to introduce him to the next two or three best, and builds a rolling rolodex. He starts with results, finds an ad he loves, and works backwards to the team that made it. He builds little mafias of top talent inside the company. He is the co-hiring manager for the top 200 people at Airbnb, not just executives, because most executives cannot hire well without help.

    Activating Talent and the Power of Belief

    You cannot teach motivation. You can only give people a problem and see if they have agency. The way to activate someone, Chesky says, is to show them potential they cannot yet see in themselves. He cites John Wooden, who said the secret to coaching was that he saw potential in players they did not see in themselves. People will climb mountains for that.

    The kindest gift anyone gave Chesky, he says, was belief. A high school art teacher named Miss Williams told his parents he was going to be a famous artist. He never became one, but the belief gave him the confidence to choose art school and to choose to be happy. Michael Seibel and the Justin.tv founders believed in him. Paul Graham made an exception to fund a non-engineer with what he thought was a bad idea. His co-founders Joe and Nate believed in him when he had no business being a CEO. The biggest gift you can give back, he says, is belief in others.

    Detaching from the Scoreboard

    Chesky describes adulation as a cup with a hole in the bottom. Status keeps draining out and you keep needing more to feel the same. The day Airbnb went public at a $100 billion valuation should have been one of the best days of his life. The next morning he put on sweatpants for a Zoom meeting and felt nothing. That triggered a re-evaluation. He stopped seeking accolades and started focusing on intrinsic work. He cites Rick Rubin: an artist is an artist when they make for themselves. He cites Vice President Obama, who told him to focus on what you want to do, not who you want to be.

    His four heroes are Leonardo da Vinci, Vincent Van Gogh, Walt Disney, and Steve Jobs. All four were working until the last week or day of their lives. Da Vinci carried the Mona Lisa with him until he died. Van Gogh sold one painting in his life. Disney was imagining theme parks in the ceiling tiles of his hospital room. Chesky says his motivation is the motivation of an artist. He calls being a CEO of a public company at his scale “almost a glitch in the system” that gave him one of the largest design canvases in human history.

    Thoughts

    What stands out about this conversation is how clearly Chesky has decoupled identity from outcome. He frames himself first as a designer, second as a CEO, and considers the resources he commands as a kind of accidental fortune for an industrial designer to be sitting on. That self-image is what lets him talk about disrupting Airbnb, killing the app paradigm, and changing the atomic unit of the company without flinching. Most public-company CEOs cannot afford that posture.

    The framework worth stealing is Project Hawaii. The pattern of taking a 10-person elite team, putting them under direct CEO coaching, and running them through crawl-walk-run-fly is a near-universal answer to the problem of innovation inside a large company. It works because it removes abstraction layers, creates direct contact with reality, and gives the founder a way to teach muscle memory before delegating. Anyone running a team of any size can borrow the pattern: pick one problem, staff it small, work with it weekly, then let go gradually. The golf-instructor analogy of teaching muscle memory before bad habits set in might be the most important management metaphor of the year.

    His prediction about consumer AI is the most economically interesting part of the talk. The fact that 159 of 175 recent YC companies are enterprise is a startling concentration. If he is right that the next 12 to 24 months bring a consumer renaissance, the opening is enormous. The hard part is what he names directly: there is no proven business model for consumer AI yet. Subscriptions cap out against free incumbents, ads are off-limits for the labs, and e-commerce has been throttled. Solving the business model is probably more valuable than building the next great consumer interface.

    The deeper philosophical thread, that AI is the transition from consumption to creation, is one that anyone building tools for makers should hold close. The 11-Star Experience also reads differently in the AI era. It used to be a thought exercise constrained by what you could plausibly build. AI compresses the gap between imagination and execution to minutes, sometimes seconds. The question is no longer “what is the most absurd version of this experience?” but “which six and seven star experiences can I now industrialize that were unthinkable a year ago?” The exercise has become operational.

    Finally, the meta-lesson on founder-led moats is worth taking seriously. The instinct in venture capital and at most public-company boards is to professionalize early. Chesky’s argument is the opposite: the longer the founder stays in founder mode, the deeper the IP and the longer the company endures after they leave. Disney is the proof. Apple is the proof. Whether Airbnb will be is the open question, and it is the question Chesky is using AI Founder Mode to answer.