PJFP.com

Pursuit of Joy, Fulfillment, and Purpose

Tag: will AI take my job

  • 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

  • Benedict Evans on Why AI Is Stuck in 1997: The Task vs the Job, Commodity Models, and Why the Jobs Apocalypse Is Overhyped

    Benedict Evans, the former Andreessen Horowitz partner and independent analyst behind the annual “AI Eating the World” presentation, sat down with Lenny’s Podcast for what the host calls the most rational take on AI you will hear this year. Instead of either doom or hype, Evans argues that AI is as big a deal as the internet or mobile, and only as big a deal as the internet or mobile, which means we are living through something closer to 1997 than to the singularity. The conversation moves through the jobs question, the difference between a task and a job, whether the model labs have any pricing power, the anti-AI backlash, and what people should actually do. You can watch the full conversation on YouTube here.

    TLDW

    Evans frames AI as a platform shift on the scale of the internet or mobile, with the crucial twist that almost nothing has been built yet, so we are in the 1997 moment where confident predictions about winners are usually wrong. He introduces his central tool, the distinction between the task and the job, to explain why “X percent of this profession is exposed to AI” studies are misleading, why the AI labs are paradoxically hiring forward deployed engineers and buying consultancies, and why accountants kept multiplying through every wave of automation (the lump of labour fallacy and Jevons paradox at work). On value capture he makes a deterministic bet that foundation models have no network effects, behave like a commodity, and will look more like cloud than like Windows, with the value moving up the stack to applications, much as it did in telecom, where a trillion-dollar industry grew data traffic thousands of times over while its stocks went nowhere. He covers distribution as the real moat, Apple Intelligence as the most compelling unshipped vision, the fuzzy anti-AI backlash (including the largely fake water panic and the very real harms of deepfakes), raising kids under radical uncertainty, and closes with the disarming admission that his own synthesis-heavy job is exactly the kind AI is currently worst at. His advice: presume radical uncertainty, dive in rather than sneer, and assume it will probably be okay.

    Thoughts

    The most useful thing in this conversation is a single question Evans keeps returning to: what is the task, and what is the job? A spreadsheet automated the arithmetic an accountant does, and the number of accountants went up for the next forty years. Claude Code can write the code, but deciding what to build, for whom, and why is the part nobody has automated. The reason the “this profession is X percent exposed to AI” studies feel hollow is that they assume a job is a neat stack of separable tasks. Evans argues, by analogy to the old expert-systems failure, that you simply cannot decompose a senior lawyer’s work that way. The 75-slide deck is the task. Walking your company, reading its politics, talking to your customers, and telling you the uncomfortable truth is the job, and that is what you actually paid McKinsey for.

    The boldest and most falsifiable claim is that the foundation-model companies look more like cloud than like Windows. No network effects means no winner-take-all, which means durable competition, which means commodity pricing and compressed margins, with the real value accruing up the stack in applications that nobody at the labs is going to build. His telecom analogy is the one to sit with. A trillion-dollar industry grew mobile data traffic by 1,500 to 2,000 times in fifteen years, and the stocks went nowhere for a quarter century, because it was a low-margin utility while all the interesting value moved to Apple and the people building apps on top. If he is right, the current token-burn economics, the person reportedly spending 1.5 million dollars a month on tokens, are the 2010 equivalent of a 50,000 dollar roaming bill, not the steady state. Evans flags openly that he could be completely wrong, which is the intellectually honest part and the part most forecasters skip.

    “It depends” and “it will probably be okay” sound like evasions, and Evans leans into that. But the 1997 framing is doing real work. The point is not that AI is small, it is that the things that will end up mattering have not been built, and that anyone confidently naming the winners today is repeating the 1997 mistake of betting on Excite over a search company with a weird logo. The discipline he is selling is to presume radical uncertainty and act anyway, because the alternative, declaring the whole thing slop and shouting about it online, buys a great feeling of moral superiority and nothing else. His repeated insistence that you can see the job that goes away but never the new job, because it does not exist yet, is the load-bearing idea under his optimism.

    The most disarming moment is the closing AI-corner answer, where the person whose entire brand is explaining AI admits he struggles to use it. His work is synthesis and precise information retrieval, and precise retrieval happens to be exactly what today’s models are worst at. He is, in his own words, the lawyer looking at VisiCalc: it is obviously transformative, and he just does not happen to make spreadsheets all day. That admission is worth more than any benchmark, because it locates the real variable. How much AI changes your life depends less on how good the model gets and more on whether your daily work sits on the part of the jagged frontier where it already works. That is a far more practical lens than arguing about whether AGI arrives in three years or thirty.

    Key Takeaways

    • Evans’s headline opinion is that AI is as big a deal as the internet or mobile, and only as big a deal as the internet or mobile. Both halves of that sentence matter.
    • If you make the internet comparison honestly, we are roughly in 1997: very exciting, most of it does not work yet, most of what people will build has not been built, and it is unclear how any of it will end up working.
    • Adoption is spread across a very wide distribution. Even among teenagers, only something like 15 to 20 percent are daily active users and another 20 percent weekly, with the majority saying they do not use it at all.
    • That spread maps onto the “jagged frontier” question of where AI works, where it does not, whether you can predict where it will work in advance, and whether you can even tell after the fact.
    • Software developers are the accountants seeing VisiCalc: for them everything has already changed. Most other professions are watching, intrigued but unsure what to do with it.
    • The AI labs are investing heavily in forward deployed engineers, consultancies, and professional services. Evans jokes that a forward deployed engineer is an Accenture outsourced developer who lives in San Francisco.
    • Companies do not have spare people sitting around to reimagine every internal workflow, so reinventing a business around AI is itself a project that needs consultants, which is why the most cutting-edge labs are funding exactly the firms everyone assumed AI would kill.
    • The central framework: separate the task from the job. Sometimes the task is the job (the elevator operator pressing a lever), and automating the task ends the job. Far more often, the task is only part of the job.
    • Amazon gets you the SKU once you know which SKU you want. Knowing which one to buy is a different job. Claude Code writes the code, but knowing what code and what features to build is the job.
    • A McKinsey or Bain engagement is not really about the deck. The deck is the task. The job is walking your enterprise, understanding the politics, talking to your customers, and telling you the truth.
    • The Jevons paradox is just price elasticity applied to labour. Make something cheaper to produce and you usually do far more of it, not the same amount with fewer people.
    • Excel did not give investment bankers shorter hours. iPhone SDKs did not shrink the number of engineers even though Apple writes 90 percent of the code for you. The number of accountants rose through every wave of automation.
    • The lump of labour fallacy: since 1800, each technology automates jobs and unlocks new ones. You can always see the job that disappears and never the new job, because it does not exist yet.
    • Evans is wary of argument from authority on jobs. He wants Dario Amodei’s view on where models go in the next 6 to 12 months, not necessarily his theory of labour markets and comparative advantage.
    • The doomer scenario of every company buying ChatGPT and firing everyone in two weeks misunderstands how enterprises work. Enterprise sales cycles run 18 months or more. Nobody is ripping out SAP overnight. The full transformation takes 3 to 10 years, sector by sector.
    • AGI and superintelligence are being quietly redefined to mean whatever works now. Larry Tesler’s theorem: AI is whatever machines cannot do yet, because once they can, people call it just software.
    • We have no theory of human intelligence, no theory of why these models work, and no theory of how much better they will get, so everyone is vibes-forecasting. Even if progress stopped tomorrow, what exists is already transformative and will roll out for a decade.
    • On value capture, Evans argues models show no network effects, so no single one runs away with the market. Persistent competition plus little real product differentiation means little pricing power.
    • Sam Altman’s pitch of selling intelligence on a meter like electricity ignores the brutal margin structure of utilities. Your TV maker does not pay the power company a cut of your bill.
    • The telecom analogy: a roughly trillion-dollar mobile industry spends 15 to 20 percent of revenue on capex, grew data consumption 1,500 to 2,000 times since 2010, and its stocks went nowhere for 25 years because it is a low-margin commodity utility.
    • The elemental question: does the model do the whole thing, or does it need thousands of different apps built by different people? If it needs apps, the labs cannot build them all, just as Microsoft did not, so it looks more like AWS than like Windows.
    • If the product is a commodity, distribution becomes the moat. Google pushes Gemini through its surfaces, Meta sprayed AI across its apps and quietly ranked between ChatGPT and Gemini in usage, and incumbents with distribution have a structural edge.
    • Browsers are the warning: Microsoft used distribution to win the browser war, then it turned out winning browsers did not matter because the value was further up the stack.
    • Apple Intelligence, as shown at WWDC 2024, was the most compelling vision of a personal AI assistant Evans has seen. Apple could not ship it, but neither could anyone else, because tool-using on-device agents with no hallucinations across thousands of apps is genuinely hard.
    • The model is “the dumb thing underneath” that powers a feature. The same commodity model can sit beneath both Gemini on Android and Apple Intelligence on iOS while the products and distribution differ entirely.
    • The anti-AI backlash is a big fuzzy mess. Some is real (local electricity bills, deepfakes, real job anxiety), some is sort of true, and some is simply false.
    • The data-center water panic is largely fake. A Livermore lab study put US data-center water consumption at about 0.017 percent of US water use. Local well conflicts are planning problems, not data-center problems.
    • We have shockingly little hard data. The model labs do not publish meaningful usage numbers. There is no public daily active user figure for ChatGPT, so economists are reverse-engineering effects from government surveys.
    • Real new harms do appear with each wave. A teenager could not use Photoshop to make explicit fakes of every classmate and send them to the whole school in an afternoon. Now they can, and turn them into video.
    • The UK Post Office Horizon scandal (buggy Fujitsu software wrongly showing cash shortfalls, leading to prosecutions, bankruptcies, and suicides) is a reminder that every technology brings new ways to ruin lives, by malice or by accident.
    • You cannot reliably predict what gets exposed. In 1997 people thought taxis were safe from the internet and newspapers would be fine. The opposite happened. Today, “AI-proof” jobs like personal trainer may not be as safe as they look.
    • Uber and Airbnb show that similar-sounding companies can have very different market impact. Uber demolished and then grew the taxi market, while Airbnb’s effect on hotels was fairly marginal because business travel still wants a hotel.
    • Every new technology first lets you do the old thing but more, then unlocks things that were not possible before. Recorded music revenue is U-shaped: first “what if I do not pay 15 dollars for a CD,” then “what if 15 dollars a month gives me all the music there is.” Spotify is not an online music store, it is something else.
    • Coding was supposed to be one of the last things automated, and instead it is the most transformed role of all, which is itself a lesson in how badly we predict exposure.
    • Practical advice: do not stick your head in the sand. Dive in, submerge yourself, and come out understanding what you can do with it. Going into a shrinking job market announcing you will never use AI is not the right posture.
    • Evans’s honest coda: he struggles to find AI use cases because his job is synthesis and precise retrieval, the things models are worst at. He uses it for proofreading, images, redecorating his apartment, and dictation. He is the lawyer looking at VisiCalc.

    Detailed Summary

    AI is as big as the internet, and we are living in 1997

    Evans opens with the opinion he calls his most controversial: AI is as big a deal as the internet or mobile, and only as big a deal as the internet or mobile. To some in tech that sounds dismissive, as if he is underrating a once-in-history event. His reply is that smartphones and the internet were themselves enormous, and we are talking over the internet right now. The deeper point is the comparison’s timing. If this is like the internet, then it is like the internet in 1997: thrilling, but most of it does not work yet, most of what will be built has not been built, and nobody knows how the pieces will fit. His latest 80-slide presentation, he jokes, is essentially 80 ways of saying “we do not know,” which is partly facetious and partly the entire point.

    The jagged frontier and the wide spread of adoption

    Adoption is not uniform, it is a wide distribution. Some people in tech have bought clusters of Mac minis and stopped using Google, while most people outside tech who use AI at all touch it once every week or two. Even among 13 to 18 year olds, daily active use sits around 15 to 20 percent, weekly use adds another 20 percent, and roughly 60 percent say they do not use it. That spread maps onto what Evans calls the jagged frontier: whether a given task works, whether you can predict in advance that it will work, whether it is intuitive, and whether you can even tell after the fact. Software developers are the accountants who just saw VisiCalc, living in a clear before-and-after. Everyone else is somewhere on the curve, picking it up to varying degrees and a little puzzled about what it is for.

    Why the AI labs are buying consultancies

    One of the most counterintuitive trends is that the leading labs are pouring money into forward deployed engineers and professional services, the very category many assumed AI would erase. Evans’s explanation is grounded in how companies actually operate. Firms do not keep spare people sitting around to redesign stores, hunt down churn, or rebuild a tech stack, which is exactly why they hire Bain, BCG, McKinsey, Accenture, or Infosys when a big project appears. Reimagining every internal workflow around AI, then actually plugging vertical and horizontal systems together and retraining people, is itself a multi-month project requiring people you do not have. So the work gets outsourced, and the most advanced labs are funding the firms that do it. His joke lands the point: a forward deployed engineer is a statistician, or an Accenture developer, who happens to work in San Francisco.

    The task versus the job

    This is the spine of the conversation. Ask what the hard part of a job really is. Sometimes the task is the job: the elevator attendant’s whole job was driving the car, the task got automated, the job ended. Much more often the visible task is only a slice. Amazon gets you the SKU once you know which SKU you want, but knowing what to buy is a separate job. Claude Code writes the code, but deciding what to build, for whom, and how to take it to market is the job. A consulting deck is the task, while the reason you pay Bain is for them to walk your company, understand its politics, talk to your customers, and tell you the truth. Evans notes you can already generate a bad McKinsey deck with AI, and the LinkedIn grifters who do are missing that the deck was never the thing you were buying.

    Jevons paradox and the lump of labour fallacy

    The Jevons paradox is just price elasticity applied to labour: make something cheaper to do and you usually do much more of it. Excel did not hand junior bankers their Friday afternoons off, it expanded the work. iPhone developers write a fraction of the raw code because Apple wrote the drivers and file system, and there are not a tenth as many engineers, there are far more. The count of accountants climbed through adding machines, punch cards, mainframes, databases, ERP, spreadsheets, and cloud. The lump of labour fallacy is the broader version: since 1800 every technology has removed jobs and unlocked new ones, the removed jobs usually look bad in hindsight, the new ones tend to be better, and GDP keeps rising. You can always see the job that disappears and never the one that does not exist yet.

    The jobs question, Dario, and the enterprise sales cycle

    On the coming jobs apocalypse, Evans is cautious about argument from authority. Running an AI lab makes Dario Amodei worth listening to on where models go in the next 6 to 12 months, not necessarily on labour economics and comparative advantage. The doomer image of companies buying ChatGPT and firing everyone within weeks misreads reality: enterprise sales cycles run 18 months or longer, nobody is tearing out SAP overnight, and the full transformation will take 3 to 10 years, sector by sector, as people slowly work out what to do. He points to the lag in software itself. Many SaaS companies founded the day before ChatGPT launched could have been built a decade earlier, and were not, because the delay was someone realizing a problem existed and that this was the way to solve it.

    Redefining AGI and superintelligence

    Evans is skeptical of the moving terminology. He cites Larry Tesler’s line that AI is whatever machines cannot do yet, because the moment they can, people call it just software. Machine learning, image recognition, and sentiment analysis all got reclassified as not really AI once they worked, the same way jet airliners were once high technology and are now just planes. AGI is now often quietly redefined as doing some percentage of economically valuable work, which a 1975 mainframe also did, rather than anything about consciousness or a soul. Whether we reach human-level intelligence is, in his view, genuinely unknowable right now. The reassuring point is that you do not need to resolve it. Even if models hit a brick wall tomorrow, what already exists is transformative and will take a decade to deploy.

    Where the value accrues: commodity models and the telecom analogy

    Here Evans makes his most deterministic argument. Foundation models appear to lack network effects, so no single model runs away from the pack, competition persists, and product differentiation as users experience it is thin. Without differentiation or lock-in, where does pricing power come from? He skewers Sam Altman’s image of selling intelligence on a meter like electricity by pointing out that utilities have terrible margins and nobody pays the power company a cut of their TV. His telecom career supplies the analogy: mobile is a roughly trillion-dollar industry that spends 15 to 20 percent of revenue on capex, grew data traffic 1,500 to 2,000 times since 2010, and whose stocks went nowhere for 25 years because it is a low-margin commodity utility while the value sits up the stack with Apple and the app makers. If models are commodities and the real product is thousands of apps the labs will not build, the outcome looks like cloud, not like Windows.

    Distribution as the moat

    If the product is a commodity, distribution decides the winners. The web browser is the cautionary tale: the browser product is a thin wrapper around a rendering engine, tab browsing was the last real innovation 20-plus years ago, Microsoft used distribution to win, and then winning browsers turned out not to matter because the value was elsewhere. Now Google drives Gemini through its surfaces and Meta sprayed AI across its apps and, in survey data, sat between ChatGPT and Gemini in usage despite tech writing it off. An adequate product with great distribution and brand becomes a big deal, which is why OpenAI spent last year trying everything to build a flywheel before the giants defaulted everyone onto their own offering. The power of the default and sheer inertia do a lot of work.

    Apple Intelligence and the model as the dumb thing underneath

    Evans calls the Apple Intelligence segment of WWDC 2024 the most compelling vision of a personal AI assistant he has seen: tool-using, on-device, agentic, with no prompt injection or hallucinations across a standardized API spanning thousands of apps. Apple could not ship it, but neither could anyone else, because that is genuinely hard. The episode illustrates his framing that the model is “the dumb thing underneath” that powers a feature. The same commodity model can sit beneath Gemini intelligence on Android and Apple Intelligence on iOS, with different products, different distribution, and different decisions about what the feature should be. Apple has a billion edge-capable devices, while Google’s “coming soon to our most powerful devices” really means it will not work on most Android phones.

    The anti-AI backlash, water, and real harms

    The backlash, Evans says, is a big fuzzy mess of very different things. Some is tangible, like a higher local electricity bill in a small number of places. Some is essentially fake, like the water panic. He dug into a Livermore lab study putting US data-center water use at about 0.017 percent of national consumption. Local well conflicts are planning failures, not data-center failures. The jobs piece is genuinely unresolved, with charts pointing both ways and a youth employment slowdown that shows up regardless of degree or AI exposure. He stresses how little hard data exists, since the labs publish no meaningful usage numbers and there is no public daily active user figure for ChatGPT. He compares the moment to the social media backlash, compressed, where some fears were true, some half true, and some simply false. The real new harms are real, though: deepfakes let a teenager generate explicit fakes of an entire school in an afternoon, and the UK Post Office Horizon scandal shows how buggy software plus institutional denial can destroy lives.

    You cannot predict what gets exposed, and what to actually do

    Evans dismisses the O*NET-style exercise of scoring what percentage of each profession AI can do as deluded, the modern version of the expert-systems problem, where you try to describe a job as 700 logical steps and it never works. You cannot say a senior partner’s work is 17 percent automatable. The history of prediction is humbling: in 1997 people thought taxis were safe from the internet and newspapers would simply save on printing, and both were wrong. Coding, supposedly one of the last things to automate, became the most transformed role of all. Personal trainers might be next once your phone can watch your form. His closing advice is to presume radical uncertainty and act anyway: do not retreat into sneering moral superiority, dive in, internalize what the tools can do, and make yourself a great hire. He ends with a candid admission that his own synthesis-and-retrieval job is exactly what AI is currently worst at, so he is the lawyer looking at VisiCalc, sure it changes everything while not personally making spreadsheets all day.

    Notable Quotes

    “My most controversial opinion is that I think that AI is as big a deal as the internet or mobile, and only as big a deal as the internet or mobile.”

    Benedict Evans, stating the thesis that frames the whole conversation

    “If you’re going to make the internet comparison, it’s like we’re in 1997. It’s very exciting. Most stuff kind of doesn’t work yet. Most of the stuff that people are going to do hasn’t been built yet.”

    Benedict Evans, on why confident predictions about AI winners are usually wrong

    “You can’t look at a senior partner at a law firm and say, well, 17 percent of their work could be automated. This is horseshit.”

    Benedict Evans, on why O*NET-style job-exposure scoring fails

    “Claude Code can write you the code, but what code do you want? It can make you the features, sure, but what features do you want? Who’s your customer? What’s the right product for that customer?”

    Benedict Evans, drawing the line between the task and the job

    “There’s this quote from Sam Altman where he said we’re going to be selling AI intelligence on a meter like water or electricity, and you look at this and think, my dear sweet child, you need me to explain the margin structure of the utility industry to you.”

    Benedict Evans, on why model labs may lack pricing power

    “The model is just the dumb thing underneath that powers the feature. The model is the commodity that powers different decisions about what the feature should be.”

    Benedict Evans, on why value moves up the stack to applications

    “Every time we have a new technology it automates away a bunch of jobs, and then that automation unlocks a bunch of new jobs, and you don’t know the new job because it doesn’t exist yet.”

    Benedict Evans, on the lump of labour fallacy and 200 years of automation

    “Don’t stick your head in the sand and say I hate all of this stuff. That gives you a great feeling of moral superiority, but that’s not going to help. What helps is you diving into this and coming out understanding what you can do with it.”

    Benedict Evans, on what to actually do about AI right now

    “AI is good at stuff that computers are bad at, and bad at stuff that computers are good at.”

    Benedict Evans, quoting an observation that explains why he struggles to use AI in his own work

    This is a curated set of pulls, not a transcript. To hear the full argument in context, including the telecom and recorded-music charts and the lightning round, watch the full conversation on YouTube here.

    Related Reading