The episode opens not with The Daily itself but with a promotional spot for The Interview, the New York Times podcast co-hosted by David Marchese and Lulu García-Navarro. The two hosts describe their mission — asking tough questions of the world's most interesting and influential people, revealing something real about those shaping our world — and invite listeners to find the show wherever they get their podcasts. It's a brisk, 35-second spot that sets the tone of the broader NYT audio network before Zolan Kano-Youngs takes over.

Zolan Kano-Youngs kicks off the episode with an unusually intimate opening — asking a series of real people to fill in the blank: 'I feel ___ about AI.' The answers tumble over each other: mixed, anxious, conflicted, love-hate. One voice captures the central tension perfectly: 'AI is amazing and it's making me a better writer, but it's also taking my job away.' Kano-Youngs quickly establishes the statistical backdrop — polling shows roughly 70% of Americans believe AI will lead to fewer jobs — before introducing today's guest, chief economics correspondent Ben Castleman, who will interrogate whether that fear matches reality. The framing is honest: for all the anxiety, what AI is actually doing to the economy 'remains pretty murky.'

As Zolan Kano-Youngs and Ben Castleman settle into their conversation, Castleman makes a striking claim: of all the economic shocks hitting the economy right now — tariffs, oil prices, geopolitical disruption — AI may be the one that defines the era in retrospect. 'If you and I are sitting here in 5 years or 10 years looking back on this period, the thing we'll be talking about is AI,' he says. He acknowledges deep uncertainty about what that conversation will look like, but insists this feels like the moment where it's all starting. The hosts also briefly joke about collecting Daily guest-hosts 'like Pokémon,' grounding the episode in a relaxed, conversational register before the hard questions begin.

Castleman delivers a quietly devastating diagnosis of America's economic measurement problem. The monthly jobs report — the gold standard of labor market data — doesn't even have a tech industry line item. Tech employment is scattered across the information sector (which also includes newspapers), professional services, and manufacturing. The categories were established decades ago and haven't kept pace with the economy. Want to know what's happening to recent college graduates month-to-month? That data doesn't exist reliably either. Castleman is careful not to call it an 'oversight' — economies move faster than data infrastructure, and you can't spin up new measures every time something changes — but the practical effect is the same: we are flying blind precisely when we most need to see clearly. ^{[1] — Ben Castleman "The U.S. monthly jobs report doesn't even have a tech industry category — that classification was set up decades ago. The result: we cannot…" 03:35 Play}

With government data inadequate, economists have turned to private-sector sources — ADP payroll data, LinkedIn job postings, Indeed hiring trends — to try to track AI's labor market impact in real time. The problem, Castleman explains, is that they're all telling different stories. Two serious economists, using careful methods and good data, can look at the same labor market and reach polar-opposite conclusions: one finding that entry-level workers in AI-exposed occupations are already losing jobs (the canary-in-the-coal-mine signal), the other finding that companies adopting AI most aggressively are hiring more people, not fewer. Castleman acknowledges this isn't entirely surprising given how fast the field is moving, but it means that anyone claiming certainty about what AI is doing to jobs right now is overreaching. ^{[1] — Ben Castleman "Credible economists using private-sector data from ADP, LinkedIn, and Indeed are reaching polar-opposite conclusions about AI's job impact.…" 06:10 Play}

Kano-Youngs presses on the string of high-profile corporate layoff announcements explicitly tied to AI — Amazon's 16,000 cuts, Block CEO Jack Dorsey's statement that AI has changed what it means to build and run a company. These feel like warning signs. But Castleman introduces a crucial piece of interpretive skepticism: companies are currently being rewarded by investors for making big AI claims, meaning any CEO who overhired during the pandemic boom and now needs to cut has a strong incentive to frame that decision as AI-driven productivity rather than a management error. He's careful not to be completely cynical — AI may genuinely be one factor among several — but the point stands: these headline announcements cannot be taken at face value as evidence of AI's direct labor market impact. ^{[1] — Ben Castleman "Investors are rewarding companies that make big AI claims, so CEOs who overhired and need to cut have a powerful incentive to frame layoffs…" 08:19 Play}

Castleman synthesizes the murky evidence into two firm conclusions. First, the technology is developing and being adopted with remarkable speed, and there is growing confidence among economists — not just Silicon Valley boosters — that it will have a real impact on the economy, the labor force, and daily life. A statement signed by around 200 economists warned it could be an unprecedented transformation, larger than the Industrial Revolution but unfolding over a vastly shorter period. Second, and seemingly contradictorily, the effect so far is subtle — subtle enough that if AI were already wiping out huge swaths of jobs, we would see it in the data. The key to squaring these two observations lies in how technology actually gets absorbed into economies. ^{[1] — Ben Castleman "Around 200 economists signed a statement warning that AI could be an unprecedented transformation of the economy — potentially larger than …" 10:00 Play}

The J-curve is Castleman's central analytical framework for understanding why AI seems both earth-shaking and imperceptible at the same time. Named for the shape of the letter J, it describes the pattern economists have observed repeatedly with transformative technologies: first, a dip, as companies struggle to figure out how to use the new tool (anyone remember the early, clunky days of video conferencing?), then a sharp upward surge once they crack the code and reorganize work around it. Kano-Youngs recognizes this in his own experience — he tries AI for a contact list, gets frustrated, and goes back to making calls. Castleman says that's exactly right: we are still in the scoop of the J, where AI may actually be making us less productive in the moment. New companies built from the ground up around AI will emerge; existing ones will figure out how to reorganize. Only then do the big economic impacts — productivity gains, and yes, job disruptions — really hit. ^{[1] — Ben Castleman "New technology always looks awkward before it looks transformative. Economists call it the J-curve: initial productivity drops as companies…" 11:16 Play}

In a mid-episode subscription pitch that lands thematically on-point, Jonathan Swan makes his case for independent journalism with characteristic directness. His job, he explains, is to dig out information powerful people don't want published, to take listeners into rooms they'd never otherwise access — and that requires human reporters with human sources. 'There's no robot that can go and talk to someone who was in the Situation Room and find out what was really said,' he says. It's a 90-second riff that functions simultaneously as a fundraising ask and as a pointed counter-argument to the idea that AI can replace investigative journalism. Listeners are invited to consider subscribing to The New York Times.

Castleman takes us back to the 1990s — dial-up, AOL, the early World Wide Web — as the closest historical parallel to the AI moment. It was an era when the internet was clearly going to be a big deal but nobody quite knew how. What the internet ultimately did to jobs is instructive: it wiped out entire occupational categories. Typing pools vanished when word processors became software rather than job titles. Travel agents became largely obsolete when Orbitz and Expedia appeared. Bank branches lost massive numbers of tellers. But none of this felt like a crisis because it happened gradually and spread across the whole economy. Workers later in their careers had time to retire; younger workers had time to redirect. The message Castleman draws: 'You had time to pivot.' That gradualness, not the internet's supposed job creation, is the real reason the 1990s technology revolution is remembered as a boom rather than a disaster. ^{[1] — Ben Castleman "The internet wiped out entire job categories — typing pools, travel agents, bank tellers — but it did it slowly enough that people had time…" 15:00 Play}

The China Shock — the wave of manufacturing job losses that swept through the American Midwest and Southeast after China's WTO entry — is Castleman's cautionary counterpoint to the optimistic internet story. The critical difference is speed and concentration. Hickory, North Carolina, once a global furniture manufacturing hub, was flooded with cheap Chinese imports and saw tens of thousands of jobs disappear in a matter of months and years. When that many jobs vanish from one community that quickly, the ripple effects are catastrophic: retail stores lose customers, schools lose funding, restaurants close. And you can't even move, because the housing market in a collapsed town has no buyers. Kano-Youngs connects the dots to consequences well beyond economics — addiction rates climbing rooted in unemployment, and the grievance-filled politics of a 'forgotten America' that has reshaped the national political landscape. Castleman's sobering observation: that all happened in a relatively small industry in a handful of places. Imagine something similar playing out across a much broader swath of the economy. ^{[1] — Ben Castleman "When trade opened with China in the early 2000s, furniture manufacturing towns like Hickory, NC lost tens of thousands of jobs in months. W…" 21:10 Play}

Having built both historical cases with care, Castleman now names the bet that everyone — workers, investors, policymakers, parents — is implicitly making. The long-run outcome, he suggests, is unlikely to be the sci-fi scenario of robots doing everything while humans idle on beaches; that's never how economic disruption has actually resolved. The near-term question is the critical one: if AI rolls out gradually and workers have time to see new career opportunities forming, if new industries emerge to absorb displaced labor the way the internet era created new jobs, the pain will be real but bounded. But if AI wipes out whole categories of jobs more or less overnight — if there's no clear 'pivot direction' because every adjacent career is also under threat — then the social, economic, and political fallout could be unprecedented. The speed of transition, not its ultimate scale, is the variable that matters most. ^{[1] — Ben Castleman "The single variable that will determine whether AI is a manageable transition or an economic catastrophe is speed. A slow rollout, like the…" 26:00 Play}

Kano-Youngs presses Castleman on whether policymakers are drawing the right lessons from history — and the answer is underwhelming. Discussions are beginning in Congress and state capitals, but 'grappling' with the lessons is not the same as acting on them. Economists are urging three tiers of response: first, build the measurement infrastructure so we actually know what's happening and which workers need help; second, shore up existing systems like unemployment insurance, which the pandemic showed is dangerously rickety; third, think seriously about whether entirely new programs are needed, from better-designed trade adjustment assistance to a sovereign wealth fund funded by government stakes in AI companies, or even universal basic income. None of those third-tier ideas are anywhere near actionable policy, Castleman notes — but the fact that they're being seriously discussed signals how extraordinary the stakes are considered to be by people who work in this space. ^{[1] — Ben Castleman "Congress and state capitals are having early conversations about AI's economic impact, but no party from either side has produced a compreh…" 27:14 Play}

The episode closes on its most personal and unsettling note. Kano-Youngs asks the question that listeners have been holding throughout: what is the individual supposed to do? What do you tell your kids? Castleman's answer is painfully honest. In the 1990s, there was a direction to point people in — go to college, pursue these new careers, even if it didn't work out for everyone. Today, there are no doubt going to be new jobs created through AI innovation, but we don't know what they look like. Castleman says he genuinely doesn't know whether to advise someone to go to college or not, what to major in, or which direction is safe. Because we can't offer clarity, people are inevitably going to feel lost. His closing line, in response to Kano-Youngs suggesting we'll 'have to embrace the uncertainty,' is three words: 'I don't know that we have much choice.' ^{[1] — Ben Castleman "During the internet boom, the answer was clear: go to college, pursue tech-adjacent careers. Today, Ben Castleman admits he genuinely doesn…" 30:57 Play}

Following the main conversation, The Daily cuts to a series of listener voices reflecting on how AI is affecting their work lives. One invokes the Office Space 'people person' meme — 'We're all just taking specifications from business people and feeding it to the AI.' Another, a worker who wasn't sure when they'd retire six months ago, says AI has accelerated that timeline because 'things are changing so much at work.' A third parent wonders aloud what their children in 5th and 7th grade will be doing when they reach college — 'Are certain choices gonna be gone?' It's a humanizing coda that puts flesh on the abstract economic arguments Castleman has just been making.

Kano-Youngs delivers the day's other news with The Daily's characteristic briskness. In Europe, firefighters are battling fast-moving and fatal wildfires in Spain and France, with another scorching heat wave expected — the fires have forced the evacuation of more than 300,000 people and witnesses describe walls of yellow flame five to six stories high. On the geopolitical front, The Times reports that the Trump administration opted against a major military escalation against Iran over the weekend, with President Trump motivated in part by concerns over dwindling military stockpiles — a vulnerability contested by U.S. Ambassador to the UN Mike Waltz, who insists the military has everything it needs and calls reports to the contrary 'nonsense' that 'deserves to be in jail.'

The episode wraps with a full production credit roll: Jack DeSidero, Diana Wynn, and Era Krupke as producers; Annie Minoff and Paige Cowett as editors; Patricia Willans providing additional editorial help; Susan Lee as fact-checker; Elisheba Itto, Marian Lozano, and Diane Wong for music; Wonderly for the theme; and Chris Wood as engineer. Zolan Kano-Youngs signs off as guest host, saying simply: 'See you tomorrow.'

The episode ends with a cross-promotion for The Wirecutter Show, teasing an episode on how to find and store great olive oil. The spot's memorable punchline — that heat, air, light, and time (the HALT acronym that degrades oil quality) happen to be the same forces that make people look old — offers a moment of levity after 36 minutes of economic anxiety.