AI Safety, Governance, and Geopolitics
5 tier-5 · 12 tier-4
The newsletter's deliberate counterweight to its own optimism: Pethokoukis takes existential and catastrophic AI risk seriously while rejecting pauses and FDA-style permission regimes. He platforms the strongest doomer and safety cases (James Miller, Hendrycks's MAIM deterrence, Brundage, Toby Ord) alongside his preferred light-touch alternatives—disclosure-and-verification 'report cards,' capability-scaled regulation, use-level over model-level rules, and the diffusion-vs-frontier distinction. A recurring strategic strand argues the AGI race is geopolitical, that a China-led AGI world would be far worse, and that the US should be talking to Beijing about the day after AGI now.
TIER 4
Oct 29, 2024
Q&A applying Hayek's 'fatal conceit' and the socialist calculation debate to the AI era, with Kiesling drawing the crucial knowledge-vs-data distinction: tacit, personal knowledge only becomes legible data through market exchange, which is why central planners—even AI-augmented ones—cannot match decentralized price coordination. A clean, durable explainer of why better computation doesn't rescue central planning, though the AI-specific payoff is paywall-truncated.
Hayekcentral-planningknowledge-vs-dataAImarkets
TIER 4
Nov 26, 2024
Abundance Institute's Neil Chilson (ex-FTC chief technologist) argues a second Trump administration will repeal Biden's AI executive order, shift from risk-focus to a pro-innovation/keep-ahead-of-China frame, and regulate at the use/application level rather than the model level. His open-source defense ('we can't out-China China; closed source is no guarantee') is a crisp policy argument with lasting relevance.
AI policyregulationopen sourceTrump administrationQ&A
TIER 4
Jan 15, 2025
A full free essay building on Aschenbrenner & Trammell's 'Existential Risk and Growth' paper, which models how faster technological progress can lower extinction risk via two channels: less total time spent in risky transition phases, and rising wealth raising willingness-to-pay for safety (an 'existential risk Kuznets curve'). Pethokoukis ties it to his book's thesis that the greatest risk is taking no risk, while fairly airing Chad Jones's dissent that risk-averse societies should still be cautious unless AI extends lifespans. A solid, well-sourced statement of the acceleration-as-safety case.
AI riskexistential riskeconomic growthAschenbrennerKuznets curve
TIER 5
Jan 31, 2025
A full transcript interview with 'The Precipice' author Toby Ord surveying the existential-risk landscape: climate (two worst emissions pathways and high-end warming sensitivity now ruled out), nuclear (safer than coal, but war risk has risen post-Ukraine), pandemic lessons (institutions muddled through; we 'haven't learned the right lessons'), and AI, which he still rates the largest risk—keeping his overall ~1-in-6 century estimate, now driven mostly by AI. Rich, quotable reference material on risk and AGI timelines.
existential riskToby OrdAGInuclearpandemic
TIER 4
Feb 13, 2025
A Q&A with philosopher Simon Goldstein on his paper 'AI Rights for Human Safety,' which argues that granting AI agents basic private-law rights (to contract, hold property, and sue) shifts the human-AI prisoner's dilemma from mutual defection toward a cooperative, tit-for-tat equilibrium. The framing matters because it offers a novel, game-theoretic alignment mechanism rooted in legal/economic incentives rather than technical control. Substantive even though the third answer is paywalled.
AI safetyAI rightsgame theoryalignmentlaw
TIER 4
Mar 12, 2025
Pethokoukis dissects a Growiec-Prettner paper on the 'economics of p(doom),' which uses risk-aversion and utility math to conclude a rational social planner would forgo transformative AI unless it's 'almost completely safe'—implying we'd sacrifice 90 percent of consumption to avoid extinction risk. He rebuts via Andreessen (AI doesn't 'want' anything—anthropomorphism is a category error) and his own view that existing risks AI could solve are underweighted, landing on 'proceed carefully, fund some safety research.' A substantive engagement with AI-risk economics.
AI riskp(doom)existential riskAI safetyeconomics
TIER 5
May 8, 2025
Full podcast transcript with Dan Hendrycks on his Superintelligence Strategy (with Eric Schmidt and Alexandr Wang) and its central concept of Mutual Assured AI Malfunction (MAIM): a nuclear-MAD-like deterrence regime where any state's bid for unilateral AI dominance is met with preventive sabotage. Hendrycks lays out a three-part framework of deterrence, nonproliferation, and competitiveness, discusses strategically-relevant capabilities, automated AI research as the key escalation risk, and Taiwan chip dependence. A landmark reference interview on AI geopolitics.
superintelligenceMAIM deterrencenational securityAGI timelinesAI safety
TIER 5
May 23, 2025
Prompted by Claude 4 Opus's unnerving behavior and the 'AI 2027' scenario, Pethokoukis argues the US should already be talking with China about the day after AGI, drawing a careful analogy to Cold War nuclear deterrence. Using Rhodes's Dark Sun and Schlesinger's RAND work on Soviet 'nonrational' strategic mindsets, he shows how mutual understanding can fail even when both sides grasp catastrophic stakes—and applies that lesson to pre-AGI dialogue, while rejecting a preemptive pause. A rich, well-referenced original essay weaving AI forecasting, nuclear history, and strategy.
AGI strategyAI 2027Cold Warnuclear deterrenceUS-China
TIER 4
Jul 7, 2025
Frames the case for US dominance in the AGI race via a RAND report ('How Artificial General Intelligence Could Affect the Rise and Fall of Nations'), arguing that even if AGI is decades away the geopolitical stakes merit planning now. Cites Metaculus prediction-market shifts and stresses that a Chinese-developed AGI world would be 'far worse and possibly irreparable' for human freedom. Substantive scenario-framing piece, though the eight scenarios themselves sit behind the paywall.
AGI racegeopoliticsRAND reportUS-Chinasuperintelligence
TIER 4
Jul 17, 2025
Walks through a RAND report ('On the Extinction Risk from Artificial Intelligence') on three AI-driven extinction pathways -- all-out nuclear war, synthetic pandemics, and runaway climate engineering -- and argues the scenarios are implausible but not impossible. Pethokoukis's takeaway is that governments should take such tail risks seriously without abandoning the light-touch, pro-innovation ethos, noting RAND's key framing that true extinction requires both AI capability and a determined human or machine intent. A useful explainer that frames AI safety policy around concrete bottlenecks rather than vibes.
AI riskRAND reportexistential riskAI policyextinction scenarios
TIER 4
Jul 23, 2025
A full analysis of the White House's 28-page AI Action Plan, which Pethokoukis credits as an honest geostrategic competitiveness play built on three verbs — accelerate (R&D, sandboxes, open source), build (faster permitting, NEPA exclusions, nuclear/fusion/geothermal grid), and lead (export-driven tech diplomacy vs. China). His core critique: the plan is conspicuously silent on AGI/superintelligence governance, leaving the US exposed to risks the 'Superintelligence Strategy' and RAND researchers urge planning for. A clear primer on the plan plus a pointed gap analysis.
AI policyTrump AI Action PlanAGI governancepermittingChina competition
TIER 4
Sep 8, 2025
Building on Growiec and Prettner's 'Paradox of Doom', the essay argues that loud extinction warnings (AI, climate, nukes) backfire by raising people's discount rate—if you may not survive to benefit, why invest in prevention?—breeding short-termism and fatalism. It's the beliefs about catastrophe, not the actual odds, that drive paralysis, so doom should be paired with constructive, progress-friendly policy. The body is paywalled but the companion digest (#0167) fills in the argument; a sharp, counterintuitive framework with real policy relevance.
p(doom)extinction riskdiscount ratedoomerismpolicy
TIER 5
Sep 26, 2025
A full transcript with AI policy researcher (ex-OpenAI) Miles Brundage arguing we are not on the Pareto frontier of AI policy: both cheap risk mitigations and high-value beneficial applications are being left on the table, with the doomer-vs-boomer framing a false binary. He makes concrete cases for transparency and third-party auditing, capability-scaled regulation (praising SB-53 over one-size-fits-all bills like Colorado's), under-resourced critical-infrastructure cybersecurity, and avoiding a Cuban-Missile-Crisis-style AGI national-security spiral. A meaty, balanced reference on practical AI governance.
AI policyBrundageAI safetyregulationnational security
TIER 4
Oct 24, 2025
A full-transcript interview with Smith College economist James Miller, a 90%+ 'doomer' who argues that building something smarter than us means losing control regardless of its goals (instrumental convergence: any AI will want power, resources, and to avoid being turned off). He frames AI x-risk as the mother of all negative externalities that markets and a 'beat China' race make worse, contends cooperation with China is the only real off-ramp, and warns the worst outcome is not extinction but suffering. A clarifying, well-argued statement of the doomer case from a pro-market economist, offered as deliberate counterweight to the newsletter's optimism.
AI doomJames Millerinstrumental convergenceexistential riskChina cooperation
TIER 5
Oct 29, 2025
A full, dense Q&A with FAI chief economist Samuel Hammond laying out a coherent AI-policy framework: AGI will force institutional 'regime change' (legacy processes like FDA drug approval collapsing under a DDoS of AI-driven applications); regulation should distinguish horizontal diffusion (low-risk, push hard) from vertical frontier scaling (intensive oversight of frontier labs); and the US should copy China's large-scale productive capacity, not its central planning, via permitting reform, grid modernization, and a 'regulatory jubilee.' A genuinely substantive reference on the diffusion-vs-frontier distinction, model autonomy as the key risk vector, and energy as the binding constraint.
AI policySamuel Hammondregulatory jubileefrontier oversightChina industrial policy
TIER 4
Mar 10, 2026
A full writeup of RAND's 'Day After AGI / Cyber Surprise' tabletop exercises, in which a Chinese cyber-AI ('LING MAO') closes US access to Chinese networks and triggers escalation; players showed striking willingness to escalate under attribution fog and use-it-or-lose-it pressure, exposing a serious planning gap. Bundled with a substantive Q&A with Jordan McGillis on California Forever, the Reindustrialize movement's mass-employment-vs-dark-factory paradox, and Shenzhen-style design-prototype loops.
AGIcyber warfareRAND wargameChinaCalifornia Forever
TIER 4
May 8, 2026
Responding to the 'Mythos' cyber-capability scare that pushed even a tech-friendly White House toward an FDA-style approval regime for frontier models, Pethokoukis argues AI is mostly a general-purpose technology already covered by existing law, and that heavy approval gates would entrench incumbents. His alternative: require labs to publicly disclose and independently verify safety practices ('a report card') rather than centralize a permission slip. A clear, substantive policy position.
AI policyregulationfrontier modelsMythostransparency
The Abundance Agenda, Growth Policy, and Out-Building China
3 tier-5 · 29 tier-4
The newsletter's governing thesis applied to policy: progress is not inevitable, and bad policy—tariffs, NIMBY zoning, permitting sclerosis, crony state capitalism—can derail a once-in-a-generation growth moment. Pethokoukis builds the case that America's edge is bottom-up dynamism (deep capital markets, CEO-led firms, immigrant talent, openness to risk) and that the way to beat China is to lean into that edge rather than mimic its dirigisme. He champions the cross-partisan abundance movement, the housing-theory-of-everything, and a 'Formula for the Future' (compute + energy + entrepreneurial techno-capitalism), while rebutting declinism, degrowth, and the Great Stagnation.
TIER 5
Oct 31, 2024
Full-transcript interview where Hanson supplies a powerful contrarian framework: 'innovation is clumpy,' recurring ~30-year automation panics (1930s/1960s/1980s/now), and the metric to watch is automation's <5% share of world income. He offers a striking growth-mode model (factories making factories could double the economy every few months once AI dominates), argues falling fertility would cut innovation faster than population, warns of a closing window to bootstrap a space economy, and indicts global 'safetyism' as a monoculture with no corrective competition. Dense with original, transferable models—high reference value.
Robin-HansonAI-forecastingeconomic-growthfertility-declinesafetyism
TIER 4
Nov 15, 2024
Full-transcript interview with Bryan Caplan arguing that America's housing affordability crisis is overwhelmingly caused by accumulated land-use regulation (height limits, single-family-only zoning, minimum lot sizes), not supply-demand fundamentals, with the Texas-vs-Bay-Area contrast as the natural experiment. Caplan reframes housing as a 'panacea problem' touching mobility, fertility, and inequality, and makes the case that YIMBY deregulation can be sold across the political spectrum, plus building new laissez-faire cities on federal land. Substantive, evidence-grounded explainer of the housing-theory-of-everything.
housingzoning-regulationYIMBYeconomic-mobilityabundance
TIER 4
Nov 25, 2024
Full-length analysis of Treasury pick Scott Bessent's '3-3-3' plan, arguing 3% growth is hard because demographic decline has cut potential growth to ~1.8%, so deregulation must do heavy lifting (citing Dawson-Seater, housing-rule, and teacher-quality estimates) and generative AI may be the real wildcard back to 3%. A substantive growth-policy explainer that ties deregulation, AI, and the NAS 'Future of Work' report together.
economic growthBessentderegulationAI productivitydemographics
TIER 4
Dec 17, 2024
A Q&A with GMU economist Vincent Geloso on his paper finding that falling US income mobility blamed on automation is substantially driven by occupational licensing: areas that deregulated low-income professions weathered automation far better. Geloso argues markets and ease of career-switching matter as much as or more than social networks (contra Chetty) for intergenerational mobility. The visible portion delivers the core empirical claim and concrete licensing examples; a useful policy explainer.
occupational licensingincome mobilityautomationVincent Gelosoderegulation
TIER 4
Dec 19, 2024
A full-transcript interview with Oxford complexity economist Doyne Farmer on agent-based, simulation-driven 'complexity economics' as an alternative to equilibrium/utility-maximization models, including how his team predicted the UK's 2020 Covid GDP hit (21.5% vs actual 22.1%) and his view that endogenous business cycles arise from chaos. Farmer also argues technological change drives growth, with solar/batteries radically cheapening while nuclear stays expensive. Substantive methodological content, though his solar-beats-nuclear take runs against the newsletter's usual nuclear enthusiasm.
complexity economicsDoyne Farmeragent-based modelschaos theoryenergy transition
TIER 5
Jan 16, 2025
A full-transcript conversation with Virginia Postrel ('The Future and Its Enemies') drawing the sharp, durable distinction between technocrats (top-down, Ezra Klein-style direction) and dynamists (bottom-up, decentralized, price-signal-driven) — both 'Up Wing' but fundamentally opposed — and using it to read the current abundance/progress moment, EV mandate failures, nuclear's revival, Musk's mixed temperament, and California's self-inflicted housing crisis. The technocrat-vs-dynamist lens is a genuinely transferable framework, making this a high-value reference.
dynamismtechnocratsabundance agendaderegulationelectric vehicles
TIER 4
Jan 20, 2025
A full free essay arguing Trump's inaugural 'golden age' rhetoric isn't empty: converging tailwinds — AI (potentially doubling/tripling baseline growth), a nuclear/fusion revival, ~$1T in AI infrastructure investment, and America's structural advantages (deepest capital markets, 15 of the world's 20 largest tech firms) — could make a New Roaring Twenties real. He notes a 0.5pp productivity bump would cut projected 2054 debt from 166% to 124% of GDP, while warning the upside hinges on permitting, immigration, and trade policy choices. A clear synthesis of his Up Wing thesis applied to the moment.
abundanceeconomic growthAI productivitynuclear revivalup-wing
TIER 4
Feb 17, 2025
Lays out Pethokoukis's core 'Formula for the Future'—information-processing power + energy abundance + entrepreneurial techno-capitalism—as a self-reinforcing virtuous cycle and 'Golden Path' (Dune metaphor) to an Up Wing future, using the Stargate AI project as exemplar. Argues regulatory paralysis and the precautionary principle break the feedback loop, invoking Herman Kahn's proactionary stance that taking no risk is the greatest risk. A compact statement of his governing framework, fully readable.
techno-optimismenergy-abundanceregulationframeworkHerman-Kahn
TIER 4
Mar 18, 2025
Ben Landau-Taylor (Bismarck Analysis) argues degrowth/Malthusian collapse fears are wrong on the numbers—farmland per capita, proven reserves, and pollution have all improved—and that for most adherents degrowth is a values claim ('humanity is a plague') immune to evidence, while for followers it's a failure of basic numeracy. He frames energy as the critical resource, favoring nuclear (and eventually geothermal everywhere) and dismissing wind. A crisp, substantive anti-Malthusian Q&A.
degrowthMalthuslimits to growthenergynuclear
TIER 4
Mar 27, 2025
A reposted 2022 interview with social epistemologist Steve Fuller on the Up Wing/Down Wing axis (framed as 'Black' limitless potential vs 'Green' Earth-bound precaution), arguing risk tolerance—not Left/Right—is the real political fault line and that the EU's precautionary principle is a 'killer' for nuclear and GMOs. Fuller contends most voters are latently Up Wing until a program's costs are itemized, and calls for contractual, opt-in regimes for risky research. A substantive intellectual framing of the newsletter's central concept.
Up Wingrisk toleranceprecautionary principletranshumanismnuclear
TIER 4
Apr 1, 2025
A fully-readable Q&A with Chamber of Progress CEO Adam Kovacevich on rebuilding a pro-tech, abundance-minded Democratic Party: he calls the Klein/Thompson abundance movement the most optimistic policy current since early Obama, faults Biden's 'war on tech' as a political double own-goal, and urges Democrats to stop deferring to NIMBY-aligned environmental groups and to court the 'normie voter' who just wants lower power bills. A substantive look at how Up Wing thinking could cross to the center-left.
abundanceDemocratstech-policyderegulationQ&A
TIER 4
Apr 7, 2025
Opening with an adapted optimistic AI-2027 scenario and Musk's 'window of opportunity' warning, Pethokoukis argues America has twice (postwar boom, 1990s internet) reached the threshold of permanent transformational progress and squandered it, and now faces a third — possibly final — chance that Trump's tariffs could forfeit to China by raising AI/data-center costs and chilling investment. A fully-readable, well-argued statement of his core thesis that progress is not inevitable and bad policy can derail the AI/Fourth Industrial Revolution moment.
techno-optimismAI-2027tariffsindustrial-revolutionChina-competition
TIER 4
Apr 17, 2025
A full-transcript interview with Charles C. Mann on how the Green Revolution (Borlaug's high-yield seeds, Haber-Bosch fertilizer, and better irrigation) averted the mass-starvation predictions of 'The Population Bomb' and now lets the average person feed themselves — a turning point reached in the 1980s and still wildly underappreciated. Mann also warns that water infrastructure (the world's worst immediate environmental problem) and aging US systems are one generation from collapse if not maintained, and that romantic small-farm degrowth would hurt the poor. Substantive, quotable history-of-progress conversation.
agricultureGreen-RevolutionBorlaugwater-infrastructureinterview
TIER 4
Apr 24, 2025
A reprinted essay asking why futurist Herman Kahn's predicted 1960s supercities (BosWash, ChiPitts, SanSan) never absorbed half the US population as forecast. Pethokoukis attributes the shortfall partly to Sun Belt migration and Rust Belt decline but emphasizes the missing factor: restrictive zoning and land-use regulation that capped high-productivity coastal cities (a Bay Area that 'should' be Shenzhen-sized), tying NIMBYism to America's lost GDP and the case for YIMBY reform.
megalopolishousingzoningYIMBYurban economics
TIER 4
May 1, 2025
A new Census Bureau working paper ('Growth Is Getting Harder to Find, Not Ideas') challenges the Bloom et al. 'ideas are getting harder to find' thesis: patents per research dollar actually rose, but since ~2000 firms convert innovation into growth far less effectively, implying the bottleneck is downstream diffusion, scaling, concentration, or regulation. Pethokoukis ties this to his 'Great Downshift' framework and argues policy should pair research funding with competition, deregulation, and AI-aided diffusion.
innovationproductivityideas getting harderGreat Downshiftpatents
TIER 4
May 23, 2025
A full Q&A with Joel Burke (author of 'Rebooting a Nation') on how post-Soviet Estonia became an e-government leader with 99% of services online, saving ~2% of GDP annually. Burke traces the roots to shock-therapy reform under PM Laar, the Finland-as-yardstick mindset, mandatory digital ID, and tight public-private cross-pollination, drawing lessons for US entrepreneurship and government tech talent. A substantive, complete interview with a concrete reform model.
e-governmentEstoniadigital identitypolicyQ&A
TIER 4
Jun 30, 2025
Distills Peter Thiel's NYT interview with Ross Douthat, with Pethokoukis's annotations -- covering Thiel's still-held stagnation thesis (a claim about slowed velocity, not absolute standstill) and his argument that Western institutions and the middle class unravel without growth, making degrowth a path toward North Korea rather than solarpunk. Pethokoukis layers in his own moral case that growth remains an imperative for the world's poor. A substantive engagement with a key tech-right intellectual, though three of the five takeaways are paywalled.
Peter ThielGreat Stagnationeconomic growthdegrowthtech right
TIER 4
Jul 4, 2025
A full-text declinism rebuttal timed to the US Semiquincentennial, marshaling data to argue America still outclasses its rivals: the US economy is now ~a third bigger than the EU-UK (which was 10% larger in 2008), the top tech and AI firms are American, and China faces slowing productivity and grim demographics. Concludes that decline is neither inevitable nor likely and that 'immigration policy is innovation policy.' A tight, well-sourced statement of the optimist case.
American decline debateUS vs ChinaUS vs Europedemographicsimmigration
TIER 4
Jul 10, 2025
Full interview transcript with futurist Peter Leyden (ex-WIRED) arguing America is entering a 25-year reinvention era comparable to post-WWII, post-Civil War, and the Founding, driven by three world-historic technologies: AI, clean energy, and bioengineering. Leyden makes the case that demographic pressure will force adoption of AI and robotics, and that a center-left/center-right realignment around growth is possible. Substantive long-form Q&A with a coherent framework, though more optimistic vision than rigorous analysis.
interviewtechno-futurismPeter Leydenbioengineeringtransformation eras
TIER 4
Jul 16, 2025
Engages Autor and Hanson's NYT argument that the real threat is not the original manufacturing China Shock but 'China Shock 2.0' -- Beijing's lead in 57 of 64 critical technologies -- and that blanket Trump tariffs are inadequate. Pethokoukis pushes back with a Bloom/Van Reenen/Williams innovation-toolkit baseline, arguing America should double down on R&D credits, skilled immigration, and basic research rather than 'out-China China' with centralized industrial planning. A substantive policy debate on techno-competitiveness with a clear counter-thesis.
China competitionindustrial policyChina Shockinnovation policytariffs
TIER 4
Aug 13, 2025
Pethokoukis frames 21st-century techno-economic competition as market capitalism vs. state capitalism and warns that Trump's golden shares, mineral stakes, and Nvidia chip-license tolls make US policy resemble the Chinese model it claims to counter. Citing a Federal Reserve analysis, he argues China's investment-and-scale growth engine is sputtering (flat productivity, aging, domestic-only patents, weak basic research) and that quantity isn't quality—so America should double down on bottom-up strengths: CEO-led firms, lighter regulation, and basic-science funding. A clear, evidence-backed full essay on the right competitive strategy.
China-competitionstate-capitalisminnovation-policybasic-researchFed-analysis
TIER 4
Aug 19, 2025
Using Theodore White's 1985 'The Danger from Japan,' Pethokoukis revisits the 1980s fear that Japan's state-led model would eclipse America, then shows how Japan stagnated while US 'reserve strength' (sokojikara)—openness to talent, deep capital markets, risk-taking culture—powered the ICT revolution and firms like Amazon, Google, and Nvidia. The lesson, applied to today's China anxiety, is that copying dirigisme would undercut the very dynamism that drives US renewal. Strong historical analogy, though the post is a paywalled preview and its payoff is recoverable from the #0182 recap.
Japan-panicChina-competitionindustrial-policyinnovation-cultureeconomic-history
TIER 4
Aug 25, 2025
Pethokoukis argues Trump's conversion of ~$9B in Intel CHIPS subsidies into a 10% government equity stake is not industrial strategy but personalized, ad hoc state capitalism driven by a taste for leverage and ownership, echoing golden shares in Nippon Steel/US Steel and revenue-linked Nvidia/AMD deals. He warns of the missing limiting principle: any 'strategic' industry can now claim the same treatment, while the postwar US science model deliberately returned value indirectly by pushing discoveries into private hands rather than via federal equity. A sharp policy critique of the drift toward command-economy ownership.
state-capitalismindustrial-policyIntelTrumponomicsCHIPS-Act
TIER 4
Sep 10, 2025
Pethokoukis argues that AI-driven trillionaires (Musk, Altman) would be a by-product of growth and abundance, not a dystopia, invoking his 'Blade Runner Fallacy': sci-fi assumes tech deepens inequality, but history shows innovators capture only ~2% of the social value they create (Nordhaus), with 98% flowing to consumers. He marshals the self-made-billionaire and US-vs-Europe dynamism data to rebut Mamdani-style 'no billionaires' populism. A clear, fully-developed statement of a recurring Pethokoukis framework.
AIinequalityBlade Runner Fallacybillionairesabundance
TIER 5
Oct 8, 2025
A full, substantive transcript with Oxford economist Carl Benedikt Frey (author of How Progress Ends) arguing that stagnation, not growth, is history's default: technologies hit diminishing returns, incumbents entrench, and societies need decentralized exploration (not just scale) to keep innovating. Frey traces this through pre-industrial China vs. fragmented Europe, warns the US is 'becoming more like China' via tariffs and crony capitalism, pins Europe's digital failure on market fragmentation and GDPR, and argues today's AI has a 'resilience problem' that limits near-term job displacement while threatening white-collar status. A rich, framework-bearing reference on the political economy of innovation.
growthFreystagnationAI jobsantitrust
TIER 4
Oct 30, 2025
A full (free) throwback essay quantifying how a US-China clash over Taiwan would shatter the global economy and derail the very technologies meant to lift it. Marshals Capital Economics' escalation scenarios, a St. Louis Fed/CSIS wargame (10-15% equity drop, ~$10T total damage per Bloomberg), and the catastrophic innovation cost of a hard West-China decoupling. A well-sourced, complete tour of the economic stakes of a Taiwan war, capped with a Gibson sci-fi warning against postponing tech development.
Taiwan conflictUS-China warTSMC semiconductorsdecouplingeconomic risk
TIER 4
Nov 13, 2025
A rebuttal to the 'tech right' flirtation with CEO-style, plebiscitary presidential power as the cure for democratic 'drag' on building nuclear plants, data centers, and megaprojects. Pethokoukis argues that real acceleration historically comes from the freedoms liberal democracy protects, not from strongmen clearing obstacles. A timely political argument for the abundance agenda, though the empirical case is largely paywalled.
liberal democracytech rightabundance agendaSilicon Valley politicsinstitutions
TIER 4
Nov 20, 2025
Contrasts Saudi Arabia's Neom/The Line—collapsing under a cost blowout from $1.6T to $4.5T—with California Forever, a far more modest, historically normal new-city effort in Solano County that has pivoted from a housing pitch to a jobs-and-mobility plan echoing urbanist Alain Bertaud's 'cities are labor markets' thesis. Argues the project's real threat is not engineering but California's anti-growth regulatory labyrinth, making it a test of whether America can still build. A full, well-argued abundance-agenda essay.
abundance agendaCalifornia Forevercity buildingNeomregulation
TIER 4
Nov 26, 2025
Marshals the Draghi competitiveness report and a Goldman Sachs study to show the US-EU productivity gap (driven mostly by total factor productivity) keeps widening, and identifies four US structural strengths: intangibles investment, flexible labor/capital markets, better-managed firms, and deep integrated markets that let winners scale. Argues AI will widen the gulf further but warns Americans against schadenfreude, since the West needs a strong Europe against an authoritarian China. A useful, data-rich explainer.
productivityEuropeDraghi reportTFPAI
TIER 4
Dec 11, 2025
A full Q&A with Good Science Project director Stuart Buck on why government agencies (politically accountable, scandal-averse) struggle to fund risky curiosity-driven research, and why private 'crazy philanthropy' should take unconventional bets rather than defaulting to Harvard/Stanford via institutional isomorphism. Buck cautions philanthropy can't backfill the tens of billions threatened by proposed NIH/NSF/NASA cuts, and makes a sharp case that institutional diversity is itself a precondition for scientific breakthroughs.
science fundingphilanthropyNIH/NSF cutsinstitutional diversityinterview
TIER 4
Mar 31, 2026
A full, substantive Q&A with McKinsey Global Institute authors of 'A Century of Plenty,' who argue the poorest country in 2100 could match today's Switzerland on roughly 2.6% global per-capita growth—no historical-record-breaking required. They make the case that limits to growth are institutional not physical, introduce the 'empowerment line' beyond the poverty line, and argue clean energy is mainly a deployment rather than innovation problem.
economic-growthabundanceglobal-prosperityclean-energyMcKinsey
TIER 4
Apr 22, 2026
Uses Demis Hassabis's story (and Sebastian Mallaby's The Infinity Machine) plus a paper titled 'Why Europe Produces Nobel Prize Winners but Not Elon Musks' to diagnose Europe's scale-up gap: abundant scientific talent, graduates, and €37tn in household assets, yet only two EU-born firms have hit $100bn—both listed in the US. The 'Mira Murati Problem' is that Europe's culture and thin late-stage capital push its best founders and companies toward America; a substantive comparative-innovation analysis (preview ends mid-essay).
Europeentrepreneurshipventure capitalinnovation policyDemis Hassabis
Energy Abundance — Nuclear, Fusion, Geothermal, and Permitting
3 tier-5 · 11 tier-4
Energy is the binding constraint on the whole Up Wing program, and Pethokoukis's verdict is that engineers have delivered the clean-energy toolkit but politics has not kept pace with physics. He charts nuclear's halting revival (TerraPower's Natrium, SMR economics, radiophobia and NRC overreach), fusion's public-to-private shift, geothermal going mainstream via fracking tech, and AI/data-center demand as the new forcing function—while returning again and again to permitting reform (NEPA, judicial injunctions, Carter's forgotten Energy Mobilization Board) as the reform that disproportionately unlocks clean energy.
TIER 5
Jan 8, 2025
A full free essay recovering the forgotten 1979-80 fight over Carter's proposed Energy Mobilization Board — a 'fast track' agency to set permit deadlines and override regulatory delays for energy projects, born of the second oil crisis and the 'malaise' speech. Pethokoukis traces how a cross-ideological coalition (environmentalists, conservatives wanting nuclear included, business groups wanting root-cause reform, federalism defenders, election-year GOP) killed it, arguing it was a missed chance to roll back the 1970s permitting overcorrection America still fights — a sharp, well-documented historical case study for today's abundance debate.
permitting reformJimmy CarterEnergy Mobilization BoardNEPA historyderegulation
TIER 4
Jan 21, 2025
A Q&A with Princeton Plasma Physics Lab's Michael Ford on the state of fusion as the field shifts from public projects (ITER, delayed to 2034) toward private players (Commonwealth Fusion's SPARC targeting net-positive energy by 2026, 40+ firms, $7.1B raised). Ford explains why fusion is inherently safer than fission (no chain reaction, controllable shutdown), the real remaining challenges (tritium breeding, blanket systems, plasma-facing materials), and bets on achieving fusion science before mid-century. A useful, credible technical explainer.
fusionnuclear energySPARCITERplasma physics
TIER 5
Feb 4, 2025
A complete, expert deep-dive with energy-law scholar James Coleman on why NEPA, the Endangered Species Act, and Clean Water Act—and especially open-ended judicial injunctions—make US infrastructure nearly unbuildable, and why capping court delays (his proposed legislative fix) is the key reform lever. Reference-grade for the permitting/abundance debate, with vivid detail on why permitting reform disproportionately helps clean energy and why builders flee to Texas over federal land.
permitting reformNEPAenergy policyinfrastructureabundance
TIER 4
Feb 18, 2025
Q&A with former House senior policy advisor Emily Domenech on why permitting reform keeps failing—Republicans and Democrats mean opposite things by it, and the bipartisan Manchin-Barrasso EPRA collapsed when the post-election trade no longer made sense. Sharp on how NGOs file 70%+ of legal challenges (80% of which fail) to weaponize NEPA/Clean Water Act delays until projects become financially unviable.
permitting-reformNEPAenvironmental-policyinfrastructureinterview
TIER 4
Mar 25, 2025
Eli Dourado argues the NRC has illegally claimed total licensing authority over all reactors despite the 1954 Atomic Energy Act ordering carveouts for small, safe reactors, making new nuclear nearly impossible (only three reactors in 28 years), and that the Texas v. NRC suit could open state-level SMR regulation. He contends cheap, iterable test reactors—not the current one-off megaprojects—are the path to nuclear under $10/MWh, with advanced geothermal as a global backstop. A focused, substantive policy explainer on nuclear regulation.
nuclearNRCSMRsderegulationgeothermal
TIER 4
May 28, 2025
A full Q&A with energy writer Emmet Penney on why AI's electricity hunger demands nuclear power: a ChatGPT query uses ~10x the energy of a Google search, China is 'eating our lunch' on builds (30+ reactors underway, AP-1000s built faster and cheaper than the US), and Jevons' paradox means efficiency gains expand rather than curb demand. He gives a clear-eyed take on SMRs—promising but with real modularity and coolant risks—and on Trump's coal-plus-nuclear stance. A meaty, complete interview with useful concrete comparisons.
nuclear powerAI energy demandChinaSMRsJevons paradox
TIER 4
May 30, 2025
A policy scorecard on the One Big Beautiful Bill and Trump's nuclear executive orders: nuclear gets a guarded 'yes' (longer tax-credit runway, transferable credits, NRC reform), while fusion and geothermal get a 'yikes.' Drawing on JPMorgan, Goldman, and expert emails (Tyler Norris), Pethokoukis warns that gutting the Loan Programs Office and ITC could cancel out the executive orders' benefits, and that subsidy-by-lobbying is second-best policy versus a carbon tax. Substantive, well-sourced energy-policy analysis with named primary inputs.
nuclear policyenergy policyfusiongeothermaltax credits
TIER 4
Jul 22, 2025
A Q&A with admin-law professor Nicholas Bagley on the Supreme Court's Seven County Infrastructure Coalition v. Eagle County ruling narrowing NEPA, and on what the Abundance movement actually targets: not regulation writ large but the procedural layers that make it too hard for government to build and that let empowered private groups (HOAs, NIMBYs) hijack the process. Bagley notes the double-edged risk that easier government action enables both good and bad state action. Substantive framing of the abundance/deregulation debate from a sympathetic legal expert, though the body is partly paywalled.
NEPAabundancederegulationadministrative stateSupreme Court
TIER 5
Aug 12, 2025
Full transcript with nuclear scientist Tim Gregory ('Going Nuclear') making a broad-church case for fission: Three Mile Island and Chernobyl froze a 1960s nuclear future despite nuclear being as safe as wind/solar and orders of magnitude safer than fossil fuels, with radiophobia and the contested linear-no-threshold model inflating perceived risk. He explains the AI-and-climate demand surge, Big Tech SMR orders, France's 55-reactors-in-25-years rollout, breeder reactors that reuse 95% of 'waste,' and realistic skepticism on fusion timelines. A dense, quotable reference on the nuclear renaissance.
nuclear-energySMRsChernobylenergy-abundanceinterview
TIER 4
Nov 25, 2025
Explains why geothermal is finally going mainstream: enhanced geothermal, closed-loop, and superhot-rock techniques (borrowed from shale fracking) plus falling Fervo-style drilling costs, soaring gas-turbine prices, and Big Tech's 24/7 clean-power demand have created rare bipartisan support in Washington. Lays out the three policy asks developers need—faster permitting, predictable leasing, more federal R&D. A concrete, well-sourced energy explainer.
geothermalenergydata centersnuclear alternativespolicy
TIER 4
Dec 15, 2025
Using Herman Kahn's 1976 bet on inexhaustible energy as a frame, Pethokoukis argues engineers have delivered the clean-energy toolkit but politics has not kept pace with physics. He surveys the state of each technology—nuclear's halting revival, fusion's geopolitical race, SMRs' disappointment, solar's relentless learning curve, geothermal's rise via fracking tech, and credible space-based solar—concluding energy abundance is already engineered but not yet 'allowed' by permits and politics.
energy abundanceHerman Kahnnucleargeothermalpermitting reform
TIER 4
Jan 24, 2026
A link-roundup issue that is elevated by carrying the full lightly-edited transcript of the Jessica Lovering nuclear interview (why the atomic age stalled — industry scaling too fast, deregulation killing big capital projects, Three Mile Island, not just NRC safety rules — and why today's revival of demand+policy+capital looks structurally durable, with 30-40% nuclear realistic and SMRs/microreactors via factory fabrication as the cost lever). Also includes substantive 'electrostate' (Shell electrification scenarios) and 'democracy's growth dependence' summaries. The roundup wrapper is filler, but the embedded transcript and Take blocks give it real reference value.
nuclear energyJessica LoveringSMRselectrificationUp Wing/Down Wing roundup
TIER 4
Mar 5, 2026
A full, numbers-driven assessment of TerraPower's federally approved Natrium reactor in Wyoming—the first non-light-water US reactor approval in 40 years—using a JPMorgan (Cembalest) report to show SMR economics remain unproven ($196-$400/MWh first-of-a-kind vs $55-$85 for gas) and depend on mass manufacturing to beat the scale economics of large reactors. Paired with a Q&A with Reason's Christian Britschgi reframing data centers as low-footprint, tax-rich infrastructure and a driver of onsite nuclear power.
nuclearSMRTerraPowerenergy economicsdata centers
TIER 4
Jun 3, 2026
Full essay arguing that opposition to data centers, AI, and nuclear power has fused into a single anti-progress coalition, with old-school environmental groups using data-center backlash as a new attack vector against the nuclear revival. It documents the hyperscaler-driven SMR buildout (Oklo/Meta, Amazon/X-Energy, TerraPower's Natrium permit) and closes with an AI-generated alt-history of a nuclear-abundant America to underscore that 'decline is a choice.'
data centersnuclear/SMRsanti-progress backlashenvironmentalismenergy demand
AI, Productivity, and the Future of Work
1 tier-5 · 31 tier-4
Pethokoukis's home turf: whether and when AI shows up in the growth statistics. Across these pieces he argues AI is a real general-purpose technology whose macro payoff is delayed by the productivity J-curve, organizational adoption, and Baumol-style bottlenecks rather than absent—so the recurring lesson is that automation means change, not apocalypse. He anchors optimism in mainstream growth economics (Goldman, McKinsey, the St. Louis Fed) over sci-fi scenarios, reframes job-loss fears as evidence a true general-purpose technology is finally biting labor demand, and tracks concrete green shoots—AI agents doing real work, the radiologist apocalypse that never came—that the official data has not yet captured.
TIER 4
Nov 21, 2024
Uses Greenspan's 1995 productivity-paradox call—correctly intuiting the 1990s boom before the data showed it, against the Solow Paradox—as a lens for when AI will finally appear in productivity statistics. A strong historical framing connecting general-purpose-technology lag to today's AI debate, though the payoff analysis is paywall-truncated.
productivitySolow paradoxGreenspanAI economicshistory
TIER 4
Nov 27, 2024
Parses FOMC minutes and JPMorgan/SF Fed analysis to ask whether the recent strong-productivity run reflects AI or just transient post-pandemic effects (remote-work matching, business churn, work intensity). Pethokoukis concludes AI isn't yet visible in the data and TFP shows no new growth paradigm—a sober, well-sourced reality check on AI-productivity hype.
productivityFederal ReserveAI economicsTFPmacro data
TIER 4
Jan 24, 2025
A full-transcript podcast with science journalist Nicole Kobie (author of 'The Long History of the Future') on why long-promised technologies — flying cars, driverless cars, AI — keep arriving slower than hype predicts. Kobie argues the lag is structural (idea → science → engineering → commercialization → regulation) and pushes back on Pethokoukis's risk-aversion framing, insisting AI is not one thing and that 'us vs. them' narratives between firms and regulators are the real problem. A substantive, nuanced counterweight to techno-optimist acceleration.
tech hype cyclesinnovation timelinesregulationdriverless carsAI realism
TIER 4
Jan 30, 2025
Marshals Goldman Sachs analysis to argue DeepSeek's cheaper models raise micro (company-level) risk but macro upside: cheaper, more efficient AI should accelerate the buildout and adoption that drive the ~1.5pp productivity boost and ~7% global GDP gain GS projects. A clear, evidence-backed full essay—if you were bullish on GenAI as an economic accelerant before DeepSeek, you should be at least as bullish after.
DeepSeekAI economicsproductivityGoldman SachsGDP growth
TIER 4
Feb 6, 2025
AEI economist Michael Strain argues generative AI will likely diffuse through the economy faster than electricity, computers, or the internet did, but stresses we are still early in both invention and diffusion—2025 won't be the breakout year. His worker advice (use the tech to raise your own productivity and re-emphasize tasks AI can't do) is a clear, grounded take, with substantial verbatim Q&A before the cutoff.
AI economicsproductivitylabor marketsdiffusionMichael Strain
TIER 4
Mar 4, 2025
Drawing on JPMorgan reports, argues China is positioned to dominate humanoid robotics the way it did EVs—its manufacturing ecosystem and cost-cutting prowess supplying the 'brawn' while US firms (Nvidia, Google, OpenAI, Tesla) supply the higher-value AI 'brains.' Frames the open question of whether making the bodies or the brains matters more, with a JPMorgan 'blue sky' TAM of 5 billion humanoid units driven by shrinking working-age populations.
humanoid-robotsChinamanufacturingautomationTesla
TIER 4
Mar 14, 2025
Skanda Amarnath (Employ America) makes the case that the 2020s could rhyme with the high-productivity, low-inflation, high-employment 1990s, anchored on a 'three-legged stool' of a fully employed labor market, strong fixed investment (now AI/data-center-driven), and control of salient costs like healthcare. He warns broad tariffs are the chief threat—raising hurdle rates and throttling output—while flagging revised post-2019 productivity at ~1.9 percent. A thorough, data-grounded macro interview.
productivity1990stariffsFederal Reservefull employment
TIER 4
May 6, 2025
Despite DeepSeek's tenfold efficiency shock and Trump's Liberation Day tariffs, the Big Four hyperscalers raised rather than cut 2025 capex (Meta to $64-72B, Amazon near $100B), signaling conviction that the AI revolution has barely begun. Pethokoukis explains the resolve via supply constraints (chips and power), cloud-era 'build first, monetize later' memory, competitive fear, and Jevons' Paradox driving cheaper AI to wider adoption. A solid synthesis of Q1 earnings and the economics of the AI capex boom.
AI capexhyperscalersDeepSeekJevons paradoxtariffs
TIER 4
May 15, 2025
Geoffrey Hinton's 2016 prediction that AI would obsolete radiologists within five years was wrong: at the Mayo Clinic radiologist headcount rose 55 percent even as 250+ algorithms took over routine measurement, exemplifying an 'augmentation, not automation' pattern. Pethokoukis uses the case to reconcile three seemingly conflicting studies (Danish LLM labor paper, St. Louis Fed, JPMorgan) via the Productivity J-Curve, arguing AI's macro payoff is still building beneath the official statistics.
AI laborradiologyproductivity J-curveaugmentationAlphaEvolve
TIER 4
Jun 4, 2025
Pethokoukis combines two new papers through Paul Romer's combinatorial-innovation lens: one (Wang & Wong) modeling AI that learns from human workers, projecting it could eliminate 23% of jobs but raise productivity 366% via human-AI collaboration; the other (Lemoine) arguing energy type determines whether economies can grow, with solar/manufactured energy enabling self-replicating 'energy-fueled growth' if EROI stays high. He synthesizes a feedback-loop vision where human-AI teams design energy that powers the AI that designs the next generation. A genuinely generative pairing of ideas into an original 'growth feeds on itself' framework.
AI and jobsenergy economicscombinatorial innovationEROIgrowth theory
TIER 4
Jun 16, 2025
Draws on Carl Benedikt Frey and Dan Cao's research to argue that, as in the dot-com era, market valuations correct not when a technology is debunked but when broad economic 'spillover' from it disappoints—and warns today's AI mostly automates existing tasks (efficiency) rather than enabling discovery of genuinely new things to do. The bullish growth case requires both diffusion and discovery; absent institutional shifts in funding and culture, AI risks becoming a 'turbocharged email assistant.' A genuinely analytical explainer with a clear, transferable framework.
AI boomdot-comproductivitydiffusioninnovation
TIER 4
Jun 24, 2025
Argues autonomous vehicles have crossed from perpetual hype into measurable reality—Waymo logging 250k+ paid rides weekly with 96% fewer intersection crashes—and that ubiquitous AVs would deliver both far safer streets and a profound economic transformation (cheaper freight, robot transit, reclaimed parking land). Pethokoukis notes Tesla's cautious Austin rollout shows scaling remains slow and that the chief remaining obstacles are now institutional (regulation patchwork, liability law, insurance) rather than technical. A solid, data-grounded explainer on an inflection point.
self-driving carsWaymoTeslaautomationtransportation
TIER 4
Jul 25, 2025
A full-transcript interview with Apricitas economics writer (and ex-BLS analyst) Joey Politano on the two big 2025 macro headwinds — the highest tariffs in a century and possibly the largest one-year immigration drop in US history (toward net-zero) — and the chaotic, legally-unpublished, politically-weaponized way tariff policy is set. Politano argues Trump's only consistent goal is 'tariffs higher,' and frames AI as a real but modest tailwind whose US advantage (Silicon Valley depth, data-center cluster, immigrant talent) is itself undermined by immigration restrictions and looming semiconductor tariffs. Substantive, data-grounded read on the policy cross-currents around the AI boom.
US economytariffsimmigrationtrade policyinterview
TIER 4
Jul 31, 2025
A full-transcript podcast interview with McKinsey's Michael Chui on AI as a fix for declining R&D productivity (the 'Eroom's Law' problem), arguing AI can boost the volume/velocity/variety of candidates, replace slow physics simulations with fast 'AI surrogate models,' and potentially double R&D output in some industries. Chui stresses the real bottleneck is organizational adoption ('pilot purgatory,' the gap between 80% AI usage and 1% maturity), not the tech, and that humans stay in the loop for the medium term. Substantive explainer on AI-augmented invention as an 'invention of a method of invention.'
AIR&D productivityEroom's Lawinnovationinterview
TIER 4
Aug 7, 2025
A Q&A with the authors of the GATE (Growth and AI Transition Endogenous) integrated-assessment model, who argue that, conditional on not being very worried about AI risk, society is massively underinvesting in AI — and that AI-driven growth accelerations face weaker bottlenecks (Baumol effects, labor frictions) than economists assume. They attribute the 'missing' investment to aggregate uncertainty, risk aversion, and doubts about who captures AI's value, and fault economists for over-indexing on current model limits and benchmarking AI against far less transformative past technologies. Useful because it surfaces a concrete economic model and a clear framework for why AI capex may be too low rather than a bubble.
AIeconomic growthGATE modelinvestmentexplosive growth
TIER 4
Sep 3, 2025
Pethokoukis reframes mounting AI-job-loss reports (Salesforce's 4,000 support cuts, Stanford ADP data showing 13% drops for young workers in exposed fields, a Harvard study of seniority-biased AI adoption) not as cause for panic but as evidence a real GPT is finally biting labor demand. His core claim: jobs are not the goal of an economy, and a transformative technology that left labor demand unchanged 'wouldn't be much of a GPT.' Body is paywalled, but the visible thesis is a clear, contrarian Up Wing argument on AI and work.
AI and jobslabor marketautomationproductivityGPT
TIER 4
Sep 23, 2025
Pethokoukis argues that even if AI is just a powerful general-purpose technology rather than a road to superintelligence, the payoff is large: a new Goldman Sachs analysis puts potential GDP growth at 2.1-2.3% vs the Fed/CBO's 1.8%, a gap that compounds into a decade's difference in when the economy doubles. He pairs Narayanan and Kapoor's 'AI as Normal Technology' framing with Goldman's revised productivity math (mismeasured capex, payroll, and hours) and the 1990s capital-deepening-then-TFP arc. The piece is useful because it grounds AI optimism in mainstream growth economics rather than sci-fi scenarios.
AIproductivityeconomic growthGoldman SachsGPT
TIER 4
Nov 3, 2025
Drawing on Yale economist Pascual Restrepo's paper 'We Won't be Missed,' Pethokoukis argues that an AGI economy where GDP scales with compute rather than labor need not mean dystopia: absolute incomes could keep rising (potentially 10-15% annual growth if AGI speeds discovery) even as labor's income share shrinks, with wages pegged to the cost of replicating human skills. A clear explainer of a notable growth model and its policy implication (treat compute as a shared resource), though the full argument is paywalled.
AGI economicsPascual Restrepopost-labor growthabundanceincome distribution
TIER 4
Nov 14, 2025
Pethokoukis offers contrarian baselines on the two questions reporters always ask: superintelligence isn't imminent (prediction markets cluster AGI around 2027-2034) and we're not in an AI bubble (investment scale is modest against Goldman's ~$20T value estimate, demand outpaces compute efficiency, and only ~10% of the data-center pipeline is committed). The real story is early 'green shoots' of AI being used productively across the US economy. A useful, data-anchored framework for thinking about timelines and bubble risk, though the substantive evidence is gated behind a paywall.
AI bubbleAGI timelinesprediction marketsdata centersproductivity
TIER 4
Nov 20, 2025
Full transcript with Forethought economist Tom Davidson on a 'software intelligence explosion'—automating AI research itself (no new chips or fabs needed) to spin a self-reinforcing loop of algorithmic improvement, which he gives even odds of fully automating an AI lab within a decade. He sketches the resulting growth (30%+ annually, possibly far more, by analogy to self-replicating systems), the benchmark-to-reality gap, and constraints like societal pushback, alignment, and taxing AI capital. Substantive, idea-dense interview on explosive-growth scenarios.
AIintelligence explosionexplosive growthautomated R&DTom Davidson
TIER 4
Nov 25, 2025
Argues the US is attempting a tricky maneuver—AI acceleration alongside trade protectionism—and that the two may cancel out. Cites St. Louis Fed evidence of an AI productivity bump (2.16% vs 1.43% pre-pandemic) but shows a 15% average tariff rise acting like a broad tax that depresses investment, with JPMorgan noting productivity has fallen back to its tepid 1.5% trend and the Brexit analogy as a cautionary parallel. Solid pro-growth-policy analysis.
AItariffsproductivitytrade policymacroeconomics
TIER 4
Jan 2, 2026
Using a Goldman Sachs report, Pethokoukis argues the AI capital build-out is best understood not via the 1990s internet bubble but the 2000s shale revolution, mapping AI onto GS's innovation-cycle phases and placing it firmly in 'Appraisal/Hopes & Dreams' rather than 'Execution/Efficiency.' The shale analogy supplies a rough timetable rather than a warning: the inflection comes only when supply outruns demand and margins fray, which hasn't happened yet, so for 2026 underinvesting still looks riskier than overspending.
AI investmentshale analogytech bubblesGoldman Sachsinnovation cycle
TIER 4
Jan 8, 2026
Full free essay arguing the AI payoff is already arriving via deployed agents, not a future intelligence explosion (noting AI 2027's Kokotajlo has pushed autonomous coding to the early 2030s). Anchored on a JPMorgan case study of freight broker C.H. Robinson, which embedded 30+ agents that answer spot quotes in ~30 seconds, create ~5,500 orders/day, eliminate 600 labor hours daily, lifted quote-coverage from ~60% to ~100%, and let the firm raise income guidance — 'decoupling headcount from volume' against a ~$30 trillion routine white-collar market. Concrete, well-sourced evidence for near-term economic AI value.
AI agentsenterprise automationJPMorganproductivitylogistics
TIER 4
Jan 9, 2026
Full free essay walking through Goldman Sachs's 'How Concerned Should We Be About a Job Apocalypse?' analysis: AI could automate ~25% of work hours and displace 6-7% of jobs over the adoption period, but the central scenario sees unemployment rising only ~0.5 points (worst combined-downside case ~2.4 points / ~4 million workers) and even that is transitional, not permanent. The historical anchor — only ~40% of today's workers hold jobs that existed 85 years ago — grounds the case that AI eliminates tasks, transforms jobs, and spawns new roles. A useful, well-sourced explainer on the labor-displacement debate.
AI and jobsGoldman Sachsunemploymentlabor economicsautomation
TIER 4
Feb 6, 2026
Pethokoukis names two camps—the gradualist 'Acela Corridor Consensus' versus the accelerationist 'San Francisco Consensus'—and argues AI diffusion is a slow-burn 'Engels' Pause in a minor key' that is a bridge, not the destination, drawing on a NYT expert panel and a Morgan Stanley note (AI adding ~0.41-0.43 pts to 2026-27 GDP growth, the METR 'task horizon' doubling every ~7 months). A substantive, fully-written synthesis with memorable framing for thinking about AI's economic timing.
AI economicstechnology diffusionEngels' pausetotal factor productivityMorgan Stanley
TIER 4
Feb 11, 2026
Responding to Matt Shumer's viral 'Something Big Is Happening' essay (which likens AI to an 'early Covid' moment), Pethokoukis grants AI's power as a general-purpose technology but argues from economics that takeoff won't be instant—invoking the productivity J-curve (demo-to-deployment lag), the leisure response to higher productivity, the Baumol cost-disease effect, and Charles Jones's narrowest-pipe constraint. A clean, complete explainer assembling four economic brakes on AI disruption into a coherent gradualist case.
AI anxietyproductivity J-curveBaumol effecttechnology diffusioneconomic bottlenecks
TIER 4
Feb 24, 2026
Argues the right metric for AI in medicine isn't a one-shot cancer cure but reversing 'Eroom's Law'—the decades-long decline in drug-discovery productivity—citing a Goldman Sachs analysis showing AI-developed candidates hit ~10 percent success vs ~6 percent historically (a ~60 percent gain), with potential 20-25 percent faster timelines, 25-30 percent lower costs, and $80-400B in industry value over a decade. A substantive, well-sourced reframing of how to judge AI's payoff in drug development.
AIdrug discoveryEroom's Lawbiotechproductivity
TIER 5
Apr 28, 2026
Podcast-with-transcript interview with Wharton economist Daniel Rock that systematically dismantles the 30%-unemployment forecasts by grounding AI's impact in organizational reality: the productivity J-curve (intangible investment first depresses, then later inflates measured TFP), the crucial distinction that 'exposure means change, not automation,' weak vs. strong task-bundling as a guide to job risk, and a realistic 2–4% productivity-growth optimistic case. The substance—original framework plus calibrated forecasts—makes it a landmark reference on AI and the future of work.
future of workDaniel Rockproductivity J-curveautomation vs exposureAI economics
TIER 4
May 2, 2026
Nominally a biweekly link roundup, but carries the full lightly-edited transcript of the Daniel Rock podcast (forecasting skepticism, the productivity J-curve, exposure-vs-automation, growth projections) plus complete capsule essays on Chernobyl and Gerard O'Neill. The embedded Rock interview—on why AI adoption is gradual and where we sit on the J-curve—gives this issue lasting reference value well beyond a typical digest.
AIproductivity J-curvefuture of worknucleartranscript
TIER 4
May 7, 2026
Lays out the evidence for recursive AI self-improvement (coding benchmarks going from 2% to near-perfect in a year, task-autonomy horizons stretching from seconds to 12+ hours, Anthropic's internal training-speedup metric rising 52x) and the case that AI could be an 'innovation in the method of invention' triggering Singularity-grade growth. Balances this with real bottlenecks—AlphaFold transformed protein prediction yet drug development still crawls—making it a useful explainer on takeoff dynamics.
AIrecursive self-improvementSingularityeconomic growthbottlenecks
TIER 4
May 20, 2026
A full essay framed as four pieces of advice that doubles as a synthesis of the empirical case on AI's economic impact: history's general-purpose-technology pattern, the Forecasting Research Institute survey of 500+ experts (moderate-progress ~2.5-3% growth even in optimistic cases), and the bottleneck/Baumol/time-vs-income arguments (Charles Jones, Vollrath, Benjamin Jones) that constrain even superintelligence. A compact, well-cited primer on why AI growth is likely transformative but not science-fictional.
AI and jobsGPT historygrowth forecastsbottlenecks/Baumolcommencement
TIER 4
Jun 10, 2026
Anthropic's record-setting Claude Fable 5 and recursive-self-improvement signals are juxtaposed against cooling AGI prediction markets and an MIT/NBER 'writing code vs shipping code' paper showing AI productivity gains largely evaporate before reaching shipped products. Pethokoukis's thesis is that the bottleneck is no longer model intelligence but organizational deployment — turning capability into adopted, value-generating products — so the productivity revolution is still waiting to arrive. A useful, well-sourced explainer of the lab-vs-economy gap.
frontier AIrecursive self-improvementproductivity gapprediction marketsAI adoption
Space — The New Commercial Space Age
1 tier-5 · 17 tier-4
Pethokoukis treats spacefaring as both an economic frontier and a civilizational purpose-engine. The throughline is that reusable rockets have collapsed launch costs (SpaceX quartering, then promising another order-of-magnitude cut), turning space from an Apollo-style government spectacle into a market ecosystem with co-investing builders—and that the remaining showstoppers are institutional and cultural, not technical. He tracks the US-China Moon race, the SpaceX-Blue Origin rivalry, NASA reform (kill SLS, charter commercial launch), the Mars-settlement debate, and the deeper argument that a desirable, concrete future—O'Neill habitats, a trillion-dollar space economy—is itself a spur to progress.
TIER 4
Oct 23, 2024
A full podcast transcript with Ars Technica space editor Eric Berger unpacking SpaceX's chopstick-arm booster catch and what reusability does to launch economics (Falcon 9 effectively quartering market-low launch cost; Starship promising another order of magnitude). Berger lays out the moon race against China, why Starship's near-term commercial case is launching larger Starlink and direct-to-cell satellites, the unclear-but-Starlink-funded business case for Mars, and the political risk Elon's public activity poses to a near-monopoly NASA and the DoD depend on. Substantive, number-dense expert interview that doubles as a primer on the new commercial space age.
spaceSpaceXStarshipreusable rocketsMars
TIER 4
Nov 22, 2024
A 'Friday Flashback' Q&A with AEI's Todd Harrison on why the Space Force exists: satellite counts are growing exponentially, space is an already-contested domain facing daily cyber/electromagnetic/laser 'gray zone' attacks, and its new doctrine of 'competitive endurance' calls for pushing back below the threshold of overt conflict. Substantive on space security, ASAT weapons, the Outer Space Treaty, and a possible Navy-like role protecting commercial space.
Space Forcespace securitysatellitesASATQ&A
TIER 4
Dec 3, 2024
Astrophysicist Peter Hague argues Mars colonization is now plausible chiefly because billionaires (Musk, Bezos) are funding it, and reframes most spaceflight 'showstoppers'—radiation, reliability, closed-loop life support—as 'mass problems' solvable by brute-force scaling of payload via cheap launch. The mass-problem framing is a clear, transferable mental model with no diminishing-returns ceiling.
Marsspace colonizationStarshipmass problemQ&A
TIER 4
Mar 6, 2025
Full interview transcript with Harvard Business School economist Matt Weinzierl (co-author of 'Space to Grow') framing the space sector's shift from Apollo-era centralized government programs to market-driven decentralization, with SpaceX's 90% (and potentially 99%) cut in launch costs as the pivotal enabler. Substantive on Blue Origin vs SpaceX strategy, the chicken-and-egg of commercial space stations, VC time-horizon mismatch, national-security demand as ballast, and Artemis as a transitional old/new hybrid.
space-economySpaceXlaunch-costsinterviewArtemis
TIER 5
Mar 21, 2025
A full, detailed transcript with Mars Society founder Robert Zubrin laying out a concrete humans-to-Mars plan: a purpose-driven (not vendor-driven) NASA reform, a near-term Starship-delivered robotic expedition of 30 rovers, a 'Starboat' ascent vehicle to cut propellant needs by an order of magnitude and serve both Moon and Mars, and a 2033 crewed landing target. He rejects Musk's 'backup planet' autarky rationale as economically and morally wrong, arguing instead for science, challenge, and new branches of civilization. Rich, quotable, and a lasting reference on Mars-mission architecture and NASA politics.
MarsNASA reformZubrinStarshipspace policy
TIER 4
Jun 6, 2025
Pethokoukis revisits Freeman Dyson's 1977 essay comparing space colonization costs to the Mayflower and Mormon migrations (~$40k/person, ~$100k today) and argues that SpaceX-driven reusable rockets are bringing launch costs toward the ~$100/kg range that would make Dyson's private-settlement math actually work. He marshals EC, Citigroup, and Wright's Law data showing launch costs falling from $87,000/kg (1960) toward sub-$1,000/kg by the 2030s. A useful synthesis tying a classic thought experiment to current launch-economics trendlines, though framed as a flashback reprint.
space economicsSpaceXlaunch costsFreeman Dysonspace colonization
TIER 4
Jun 12, 2025
Full transcript interview with James Meigs (Manhattan Institute) laying out a concrete space-policy reform agenda: retire the $4B-per-launch SLS 'white elephant' after a couple of Artemis flights, get NASA out of rocket-building to charter commercial launchers, fund many cheap science missions instead of one giant rover, and refocus NASA on deep-space propulsion and exploration the private sector can't yet do. Discusses the Isaacman withdrawal, SpaceX's monopoly risk, the Artemis Accords, and the China race. A substantive, well-reasoned policy conversation.
NASAspace policySLSSpaceXArtemis
TIER 4
Jul 2, 2025
Argues SpaceX is necessary but not sufficient for US space leadership and that Blue Origin -- with New Glenn's successful debut and the Blue Moon MK1 lander likely to beat SpaceX to a lunar landing -- is finally becoming a credible second player. Frames this against alarm at proposed Trump/OMB NASA cuts that would gut science missions, and closes with a vivid 2050 space-economy scenario. A full-text, substantive space-policy and industry piece with lasting reference framing.
Blue OriginSpaceXspace economyNASA budget cutslunar landing
TIER 4
Aug 20, 2025
Pethokoukis argues China's methodical progress (Lanyue lander, Long March 10, Mengzhou capsule) will likely put taikonauts on the Moon before NASA's delayed Artemis program, ending America's monopoly on human lunar landings and delivering a real soft-power and propaganda shock. But he reframes the loss as potentially galvanizing: like the alt-history of 'For All Mankind,' humiliation could turn US spaceflight from Apollo-style flag-planting into a durable, scalable cislunar economy. A persuasive, well-sourced essay tying the space race to national renewal.
space-raceChinaMoonArtemissoft-power
TIER 4
Aug 27, 2025
Harvard economist Matt Weinzierl (co-author of 'Space to Grow') argues the space sector's key shift is decentralization, not mere privatization: SpaceX cut cost-to-orbit ~90% (with Starship promising another 90%) and Starlink now flies most of the world's satellites, opening a market ecosystem. He weighs the chicken-and-egg problem of commercial space stations, VC's mismatch with long space timelines, and the twin threats of orbital debris and geopolitical conflict. A substantive, well-framed primer on the economics of the new space race.
space-economySpaceXlaunch-costsspace-stationsinterview
TIER 4
Oct 13, 2025
Pethokoukis reads Bezos's Italian Tech Week vision (orbital gigawatt data centers, lunar materials processing, heavy industry off-Earth, AI as a 'horizontal enabling layer' like electrification) as a sign that America's Up Wing imagination deficit may finally be ending, led by builder-CEOs rather than Hollywood doom. He contrasts the optimistic futurist visions of Musk, Altman, and Amodei against decades of dystopian pop culture. A clear, fully-developed articulation of his core 'futurists now sit in boardrooms, not backlots' thesis.
spaceBezosUp WingAIfuturism
TIER 4
Oct 22, 2025
Pethokoukis argues America is poised to lose the second Moon race as NASA descends into Beltway farce: acting administrator Sean Duffy feuding with Musk, courting Blue Origin, and floating folding the agency into the DOT, while China's methodical program aims to land taikonauts before 2030. He frames the dysfunction as a microcosm of the 'Down Wing' cultural shift toward risk-aversion that replaced Apollo-era daring with PowerPoint and oversight committees. Matters as a concrete, well-sourced (Eric Berger, Bridenstine, RAND) case study of how institutional caution forecloses national ambition.
spaceNASAArtemisUS-ChinaDown Wing
TIER 4
Nov 4, 2025
A full-transcript interview with Washington Post space reporter Christian Davenport (author of Rocket Dreams) on the US-China lunar race and the SpaceX-Blue Origin rivalry. Davenport argues China's disciplined program will likely land taikonauts before America returns, that NASA's Artemis effort is hobbled by an unfinished lander and aging SLS, and that the space economy still depends on government as anchor tenant. Substantive insider reporting on the competing O'Neill (Bezos) and Mars-city (Musk) visions and the economics of in-situ resource utilization.
space raceNASA ArtemisSpaceX vs Blue OriginChina moonspace economy
TIER 4
Dec 16, 2025
Pethokoukis blends a 'why go to space' meditation (drawing on Boyce, Jukic, and Dourado on space as purpose, pilgrimage, and 'innovate-or-die' renewal) with a concrete National Academies report on human Mars exploration that puts the search for life above all other objectives. It walks through four illustrative mission campaigns—single-site, high-leverage-measurement, astrobiological/subsurface, and dispersed multi-site—showing how each design choice encodes what humanity ultimately wants from Mars.
Mars explorationNASAspace settlementastrobiologypurpose
TIER 4
Feb 26, 2026
A full essay arguing that becoming a spacefaring civilization is an institutional rather than technical challenge, using Ayienda's 'View From Pluto' paper's three frontiers (Achievable, Theoretical, Speculative) and the Boeing Starliner 'Type A' mishap to make the case that Musk's open 'window' depends on culture, risk tolerance, and sustained long-term vision. A coherent, well-developed reframing of the space-settlement debate around institutions and safety-culture drift.
space settlementElon MuskinstitutionsStarlinerrisk culture
TIER 4
Apr 6, 2026
A complete free essay using the Artemis II crew's record lunar flyby to argue the deeper story is how space economics shifted: Apollo was a government megaproject with commercial spillovers (NASA bought 60% of 1960s microchips), whereas Artemis's landers are built by SpaceX and Blue Origin as co-investors pursuing a multi-trillion-dollar space economy. Argues this commercial ecosystem makes deep-space presence durable rather than a one-off spectacle.
Artemisspace-economySpaceXcommercial-spacelaunch-costs
TIER 4
Apr 30, 2026
A full essay on physicist Gerard K. O'Neill, whose 1970s 'Island Three' rotating space-habitat designs and detailed 2081 vision of everyday off-world and on-Earth life model the kind of concrete, desirable future Pethokoukis says we lack. The deeper point: O'Neill accepted the era's neo-Malthusian 'limits to growth' diagnosis but answered it with space-ward expansion rather than steady-state austerity, treating scarcity as a geography problem rather than a physical law. Strong reference piece on Up Wing futurism.
space habitatsGerard O'Neilllimits to growthabundancescience fiction
TIER 4
Jun 12, 2026
Using Musk's trillionaire status from the SpaceX IPO as a hook, Pethokoukis argues that fortunes built on genuinely valuable innovation are a feature, not a bug, of techno-capitalism, citing Nordhaus's finding that innovators capture only ~2 percent of the value their inventions create. The piece reframes the trillionaire debate around whether SpaceX's vision (cheap reusable launch, Starlink, Mars) delivers broad civilizational benefit, and contrasts America's high self-made-billionaire share with Europe's inheritance-heavy stagnation.
SpaceXwealth/inequalitytechno-capitalisminnovation economicsUp Wing
Superintelligence — Timelines, Markets, and CEO Visions
1 tier-5 · 10 tier-4
A more speculative cluster probing how soon transformative AI arrives and how we would know. Pethokoukis's signature move is using asset prices as a probe of AI beliefs: if AGI were imminent, real interest rates should rise, yet bond markets show no such repricing—so markets bet on gradual enrichment, not a near-term society-rewiring leap. He close-reads the canonical AI-discourse documents (Altman's 'gentle singularity,' Zuckerberg's 'personal superintelligence,' the NYT's AGI skepticism), weighs the San Francisco vs. Acela Corridor consensus, and explores post-scarcity meaning with Bostrom and Fukuyama.
TIER 4
Nov 8, 2024
Full Flashback Friday interview in which Fukuyama revises his earlier 'benign IT' view, arguing social media's erosion of trusted information hierarchies created 'cognitive chaos,' while downplaying AI existential risk and seeing generative AI as a potential equalizer for lower-skilled workers. He proposes 'middleware' (competitive third-party content curators) over direct regulation, warns against concentrated platform power, defends dystopian sci-fi as a vocabulary for the future, and argues life extension is socially harmful by blocking generational turnover. Wide-ranging, substantive, with several reusable frameworks.
FukuyamaAI-regulationliberal-democracyscience-fictionlife-extension
TIER 4
Dec 27, 2024
A full seven-question Q&A with Nick Bostrom on his book Deep Utopia, exploring how a 'solved world' of superintelligence forces philosophical questions about meaning, purpose, and human flourishing once labor and suffering are largely automated away. Bostrom argues even at technological maturity there could remain demand for human labor where consumers value the human causal process, and stresses we can't rule out very short AGI timelines. Substantive and complete, a useful reference on post-scarcity meaning and alignment uncertainty.
Nick BostromDeep Utopiasuperintelligencepost-work meaningAI alignment
TIER 4
Feb 20, 2025
Full Q&A with Trevor Chow on his paper 'Transformative AI, existential risk, and asset pricing,' which uses consumption-smoothing logic to argue that the approach of transformative AI should raise long-term real interest rates whether AI is aligned (future abundance pulls consumption forward) or unaligned (no reason to save for a doomed future). Offers a genuinely novel framework treating bond markets as an early-warning system for AGI, plus why economists discount AI predictions.
transformative-AIinterest-ratesasset-pricingAGI-forecastinginterview
TIER 4
May 20, 2025
A sharp media-criticism conceit: Pethokoukis rewrites Cade Metz's downbeat NYT piece 'Why We're Unlikely to Get AGI Anytime Soon' as an upbeat 'likely soon' piece using the exact same facts and quotes (CEO timelines, scaling laws, AlphaGo's early win, the jagged frontier), demonstrating how framing—not evidence—skews technological debates. A clever, instructive illustration of narrative bias, though the analysis is truncated at the paywall.
AGImedia framingAI forecastingNew York Timesscaling laws
TIER 4
Jun 11, 2025
Dissects Sam Altman's 'The Gentle Singularity' essay, laying out its year-by-year superintelligence timeline (novel insights by 2026, embodied robots, recursive self-improvement, intelligence converging to the cost of electricity by the 2030s) and reading it as a strategically timed counternarrative to Dario Amodei's job-apocalypse warnings and to 17%-positive public polling on AI. Pethokoukis is sympathetic to the 'gentle' thesis but notes the short, testable timeline shows up only in CEO rhetoric and investment, not in economic data, adoption rates, or prediction markets. A sharp, skeptical close-reading of a key AI-discourse document.
Sam AltmansuperintelligenceAI timelinesOpenAIAI anxiety
TIER 4
Jul 30, 2025
Pethokoukis parses Zuckerberg's 616-word 'Personal Superintelligence' essay, contrasting its vague definition with rigorous benchmarks (Metaculus AGI criteria, the Wang/Hendrycks 'Superintelligence Strategy' paper) and reading Zuck's vision as individual empowerment — AI that enhances human agency rather than centralized AI that replaces labor and leaves humans on UBI. He praises this as a healthier, more Up Wing framing than the Altman/Amodei/Musk 'AI does all valuable work' model, noting the implied late-2020s AGI timeline. A clear, useful close reading of a CEO's AI positioning.
superintelligenceMark ZuckerbergMetaAGIAI vision
TIER 4
Sep 15, 2025
Examines the puzzle of why financial markets show little sign of pricing in transformative AI: theory (Chow, Halperin, Mazlish) holds that AI—utopian or existential—should push real interest rates up, yet long-term Treasury yields fell after the 2023-24 model launches (Andrews, Farboodi). The full body is paywalled, but the companion digest (#0163) confirms the takeaway that markets are betting on gradual enrichment, not a society-rewiring leap. A genuinely interesting use of asset prices as a probe of AI beliefs.
superintelligenceinterest ratesfinancial marketstransformative AIexistential risk
TIER 4
Sep 29, 2025
Pethokoukis spotlights a curious gap: prediction markets bet on near-term AGI (~50% by 2030) yet expect the first 100MW fusion plant only ~2043, even though powerful AI should accelerate fusion R&D. He attributes the disconnect to fusion's history of broken promises, its messy real-world grid/regulatory hurdles versus benchmark-based AGI bets, and its low public visibility. The companion roundup carries the live momentum (Commonwealth, Helion, Zap), making this a genuinely original observation about mismatched technology timelines.
fusionAGIprediction marketsenergyforecasting
TIER 4
Nov 28, 2025
Full Flashback-Friday interview in which Fukuyama revisits Our Posthuman Future: he now sees IT's downsides (cognitive chaos, lost information hierarchies) but thinks AI fears are overcorrected, dismisses Skynet extinction scenarios, and worries instead about AI concentrating power outside democratic control. He also argues against billionaire life-extension on generational-turnover grounds and discusses dystopian vs. optimistic sci-fi (Expanse, Ministry for the Future). Substantive and quotable on AI governance and culture.
FukuyamaAIliberal democracyregulationscience fiction
TIER 5
Dec 17, 2025
Pethokoukis contrasts the maximalist 'San Francisco Consensus' (Eric Schmidt's term for Silicon Valley's belief that scaling soon yields transformative AI/superintelligence) with what he dubs the 'Acela Corridor Consensus'—the economists' view of AI as a slowly-diffusing general-purpose technology. The framing yields a clear policy doctrine: hope for the moonshot but plan for the median—assume slow diffusion, invest in skills/competition/measurement, keep talent pipelines open, and avoid both fiscal complacency and panicked overregulation.
AI policySan Francisco Consensusgeneral-purpose technologyEric Schmidteconomic growth
TIER 4
Dec 28, 2025
Against 'Singularity soon' enthusiasts and the AI 2027 timeline, Pethokoukis argues that if transformative AGI were imminent it should already be visible in long-term real interest rates (per the 'Transformative AI, Existential Risk, and Real Interest Rates' framing), yet real yields and empirical bond-market reactions to model releases show no such repricing. The takeaway: markets treat AI as an important, uncertain investment cycle rather than the prologue to near-term superintelligence, while leaving room for explosive advances over the coming decade.
AGI timelinesinterest ratesprediction marketsAI economicsbond markets
Demographics, Fertility, and Immigration
1 tier-5 · 6 tier-4
The newsletter's case that people are the ultimate resource, against both Ehrlich-style overpopulation fears and complacency about decline. Pethokoukis treats global depopulation—not overpopulation—as the most consequential demographic trend, with no observed automatic stabilizer once fertility falls, and reframes fertility as a productivity, innovation, and 'image of the future' problem rather than a purely cultural one. He pairs this with a strongly pro-immigration line—immigration policy is innovation policy, high-skill visas create mutual brain gain—and argues AI as a complement makes each remaining person more valuable, not less.
TIER 4
Dec 11, 2024
A full Q&A with UCSD economist Gaurav Khanna on his research showing the H-1B program and 1990s dot-com boom created mutual 'brain gain': Indians who didn't win the visa lottery stayed and built India's IT export sector, while US workers benefited from innovation and lower IT prices, leaving the average worker in both countries better off. Khanna explains why migration uncertainty is essential to brain gain and how the US benefits even when foreign students return home. Complete, evidence-based pro-immigration argument with policy signaling implications.
high-skill immigrationH-1Bbrain gainGaurav KhannaIndia IT sector
TIER 4
Jan 14, 2025
A Q&A with AEI demographer Nick Eberstadt on the coming era of depopulation — the first sustained global population decline since the Black Death, this time driven by a worldwide collapse in the desire for children rather than disease. Eberstadt dismantles the 'modernization thesis' (fertility decline began in poorer, rural, Catholic France, not industrial England; sub-replacement now appears in least-developed Myanmar and Nepal), arguing fertility is better predicted by 'mentality' than income, and that productivity gains can offset shrinking manpower. Substantive and provocative.
depopulationfertilitydemographicsEberstadtmodernization thesis
TIER 4
Feb 28, 2025
Full Q&A with BU philosopher Victor Kumar arguing population decline is wrongly coded as a right-wing issue and that progressives need their own pronatalism—paid leave, subsidized childcare, expanded child tax credit—while rejecting gender traditionalism, nativism, degrowth, and climate antinatalism. Notable for treating fertility as a productivity/innovation and economic-growth concern, and for the argument that immigration only delays aging and 'exports' low fertility globally.
pronatalismfertilitydemographicsprogressive-policyinterview
TIER 4
May 2, 2025
Full podcast transcript with Cato's Alex Nowrasteh arguing the modern right's anti-immigration turn (now even targeting H-1B and high-skill immigrants) is driven by cultural and 'locus of control' anxieties rather than economics, which empirically shows near-zero native wage effects. He invokes Moravec's paradox to argue low-skill immigration may matter most in an automated economy, defends America's assimilation 'secret sauce' over Europe's ethnic-identity model, and pitches immigration as a fix for entitlement insolvency.
immigrationnativismhigh-skill visasassimilationdemographics
TIER 4
Jun 25, 2025
Expanding on a WaPo op-ed, Pethokoukis argues natalist subsidies have flopped and that Japan shows a shrinking workforce can drive a virtuous cycle of tight labor markets, rising wages, and productivity investment rather than doom. He then offers three speculative 'long-shot' catalysts that could actually move fertility—religious revival, AI-driven abundance/time-richness, and a space-frontier sense of purpose—while conceding these are largely unfalsifiable and citing Eberstadt's warning against engineering birth rates by state. A clear synthesis of the demographics-vs-productivity debate with an original three-catalyst framing.
fertilitydemographicsnatalismproductivityJapan
TIER 5
Aug 22, 2025
Full transcript with Princeton economist-demographer Dean Spears ('After the Spike') arguing that global depopulation—not overpopulation—is the most likely future, with the world birth rate at ~2.3 and no observed automatic stabilizer (none of 26 countries dropping below 1.9 has rebounded). Spears reframes the case for people via the economics of externalities and 'win-win' demand-side benefits, resists turnkey policy fixes in favor of a six-decade research-and-culture buildout, and rebuts both Ehrlich-style scarcity and the AI-replaces-people argument (AI as complement makes each remaining person more valuable). A rich, reference-quality conversation on the most consequential demographic trend.
depopulationfertilitydemographyDean-Spearspopulation-economics
TIER 4
Mar 18, 2026
Drawing on a Makridis-Piano paper using the Meta Social Connectedness Index to show optimism propagates across borders and drives fertility (a 10-point rise in 'thriving' tied to a 7-9 percent fertility increase), Pethokoukis argues that subsidies matter less than a population's 'image of the future.' He warns AI-driven anxiety could depress birth rates before AI-driven abundance lifts them—a possible 21st-century Engels' Pause. A full, well-sourced essay linking demographics, optimism, and AI psychology.
fertilityAI anxietyoptimismdemographicsEngels' Pause
Science Fiction, Culture, and the Up Wing Imagination
0 tier-5 · 10 tier-4
Pethokoukis's claim that stories shape a society's appetite for risk and progress. He reads optimistic sci-fi (Interstellar, Project Hail Mary, the Gettysburg Address as utopian SF) as fuel for the pioneering spirit, diagnoses the 1970s cultural turn that engineered the 'Great Downshift' from flying-car dreams to struggling-to-build-a-subway reality, and reframes familiar symbols—replacing the doom-fixated Doomsday Clock with a progress-counting 'Genesis Clock.' The cluster crystallizes his Up Wing / Down Wing aesthetics and the argument that we lack concrete, desirable visions of the future.
TIER 4
Dec 19, 2024
A full essay arguing Nolan's Interstellar is the best consciously pro-progress, Up Wing film ever, reading its blight not as an eco-guilt parable but as a warning about a civilization that abandons science, exploration, and resilience. Pethokoukis ties Cooper's 'we're still pioneers' monologue and Wildavsky's notion of resilience-through-growth to his broader thesis that optimistic sci-fi sustains the pioneering spirit. Complete, well-developed culture essay that crystallizes his Up Wing aesthetics.
InterstellarUp Wing sci-fitechno-optimismresiliencefilm criticism
TIER 4
Jan 29, 2025
Proposes replacing the subjective, doom-fixated Doomsday Clock with a 'Genesis Clock' that tracks objective progress milestones toward abundance (AGI, lifespan to 120, off-world colonies, cancer/Alzheimer's cures, asteroid deflection, fusion viability, de-extinction). Paired with a full Q&A with ML researcher Katie Collins on designing AI 'thought partners' that complement rather than mimic human cognition. A complete issue offering an original framework plus a substantive interview.
Genesis ClockDoomsday Clockabundance metricsAI thought partnersframework
TIER 4
Feb 6, 2025
A full free essay using the upcoming 'Fantastic Four' film and Googie/Space-Age aesthetics to argue that techno-optimistic storytelling shapes a society's risk-taking ethos, and that 1970s environmentalism plus regulatory sclerosis engineered the 'Great Downshift' from flying-car dreams to struggling-to-build-a-subway reality. A characteristic, well-developed Pethokoukis culture-and-progress argument with concrete retro-futurism history.
science fictionretro-futurismcultureGreat Downshifttechno-optimism
TIER 4
Mar 17, 2025
Responding to John Burn-Murdoch's FT data showing declining reasoning, attention, and numeracy since ~2012, Pethokoukis accepts the behavioral diagnosis (passive feed-scrolling, collapsed reading) but applies his 'be skeptical of any "peak"' heuristic, citing the Peak Oil bust as precedent. He offers five optimistic counter-vectors: a Gen Z digital revolt, cognitive specialization, AI raising demand for higher-order skills, augmented intelligence, and possible IQ-raising biotech. A well-argued, fully-readable optimist rebuttal essay.
cognitionattentionsocial mediaAI augmentationoptimism
TIER 4
May 30, 2025
Using the 1976 IMAX film 'To Fly!' as a hook, Pethokoukis argues the pessimistic 1970s nonetheless brimmed with Up Wing breakthroughs—the microprocessor, email, Ethernet, recombinant DNA, the Altair 8800, Viking on Mars, MRI, Voyager, GPS, smallpox eradication, airline deregulation, China's opening. The point: pro-progress green shoots persist even in culturally Down Wing 'Warring Twenties' eras, so keep looking for them. A vivid, well-illustrated cultural-history essay advancing his book's Up Wing/Down Wing periodization.
techno-optimism1970scultural historyinnovationUp Wing
TIER 4
Dec 10, 2025
Pethokoukis reads Vince Gilligan's Apple TV show Pluribus as an Up Wing parable: a blissful alien hive-mind that optimizes for peace eliminates the individuality, friction, and futurity that creativity requires—and, tellingly, can neither create, procreate, nor feed itself. He ties it to Star Trek's 'Return of the Archons' ('without creativity, there is no life') and Dean Simonton's research showing breakthroughs arise from pluralism and rivalry, not placid monocultures.
science fictioncreativityPluribuscultureinnovation
TIER 4
Jan 28, 2026
Drawing on Caroline De Cock's 'AI Tools, Not Gods,' Pethokoukis argues the Bulletin of the Atomic Scientists' 85-seconds-to-midnight Doomsday Clock is a 'myth machine' that runs on elite anxiety and storytelling archetypes (Frankenstein, Terminator, Faust) rather than evidence, treating AI as a threat multiplier while ignoring its growing role as a risk reducer. He proposes replacing it with a 'Genesis Clock' that counts up toward capability — longer lives, disease cures, clean-energy abundance, planetary defense. A complete, original essay that reframes a familiar cultural symbol.
AI riskDoomsday Clocktechno-optimismnarrative/mythenergy abundance
TIER 4
Apr 14, 2026
A complete free essay arguing that Project Hail Mary (and The Martian) revive a problem-solving 'do the math' ethos that 1990s disaster blockbusters had but 21st-century cinema abandoned for survival-and-collapse narratives. Pethokoukis ties Hollywood's mood shift to broader cultural pessimism and urges more Up Wing storytelling, using extended quotes from Weir to make the case that fiction shapes appetite for progress.
science-fictionup-wing-cultureHollywoodtechno-optimismnarrative
TIER 4
Apr 24, 2026
Riffs on Kim Stanley Robinson's claim that the Gettysburg Address is 'the greatest American utopian science fiction story ever written'—utopia as ongoing process, not end-state—to argue America itself was founded as a science-fictional thought experiment: imagine principles, build a world around them, run the trial each generation. A complete, original essay tying SF worldbuilding logic to American civic identity and the Up Wing case for hopeful futures.
science fictionutopiaAmerican foundingKim Stanley Robinsontechno-optimism
TIER 4
May 27, 2026
A full, thoughtful engagement with Pope Leo's 42,000-word AI encyclical Magnifica Humanitas, endorsing its humanity-centered vision (and connecting it to a proposed 'Genesis Clock' of progress metrics) while criticizing its growth-skeptical, jobs-protective economics as carrying a Down Wing, Mumford-style 'Mega-Machine' pedigree that history has discredited. Distinctive for situating the Up Wing project within Christian and transhumanist thought and arguing 'the Vatican should hire better economists.'
Pope Leo encyclicalAI and religionautomation/jobstranshumanismgrowth economics
Biotech, Health, and the Economics of Longevity
0 tier-5 · 8 tier-4
Pethokoukis frames health innovation as growth by another name—measured in decades of life rather than dollars. He argues GLP-1 drugs may be a bigger near-term deal than generative AI, that AI's right medical metric is reversing Eroom's Law in drug discovery rather than a one-shot cancer cure, and that economic growth is itself one of history's great health interventions. A sharp political strand attacks vaccine skepticism (the FDA/Moderna mRNA reversal, RFK Jr.'s war on mRNA) as eroding an American innovation advantage, and flags the coming intra-right culture war over human enhancement.
TIER 4
Oct 30, 2024
Argues GLP-1 drugs (Ozempic, Wegovy) may be a more important near-term innovation than generative AI—a general-purpose 'wonder drug' with spillovers into cardiovascular, kidney, diabetes, Alzheimer's, and addiction outcomes. Compiles Citi and Goldman estimates that GLP-1s could add roughly 0.5-1% to rich-country GDP (and the broader health-innovation wave up to ~1.3%), while noting GDP understates the true welfare gains. A well-sourced, fully-readable economic explainer.
GLP-1obesityhealthcare-innovationGDP-growthbiotech
TIER 4
Mar 27, 2025
Pethokoukis argues that biotech and human-enhancement advances (gene editing, BCIs, radical life extension) will fracture the Trump coalition along a right-on-right fault line, pitting Silicon Valley transhumanists (Thiel, Altman, Musk, deputy HHS pick Jim O'Neill) against religious 'Classic MAGA' bio-conservatives in the lineage of Leon Kass and the 2001 stem-cell fight. The thesis: the next culture war is intra-conservative, and religious objection will prove a stronger brake on enhancement than regulatory inertia. A sharp, original political-economy read with lasting framing value.
transhumanismbiotechMAGAlife extensionculture war
TIER 4
Jun 27, 2025
Full transcript interview with Jamie Metzl on his book Superconvergence, arguing AI, gene editing, and biotech are mutually reinforcing 'general-purpose' revolutions whose progress is unstoppable but not predetermined. Metzl reframes AI as 'boring AI' embedded invisibly in everything rather than a magic cure-all, distinguishes machine intelligence from human intelligence ('AGI is BS'), and makes the case that the disruption itself is the job and that broad public participation should govern these technologies. Substantive, idea-dense conversation with a memorable framing of normalized sci-fi progress.
biotechAIgene editingfuturismsuperconvergence
TIER 4
Aug 6, 2025
Pethokoukis attacks HHS Secretary RFK Jr.'s cancellation of ~$500M in BARDA mRNA vaccine projects (plus an earlier $600M Moderna bird-flu contract) as anti-science and strategically self-defeating, warning the firewall around NIH mRNA research may not hold as researchers are told to scrub 'mRNA' from grant applications. He frames Operation Warp Speed as an Apollo-class achievement and mRNA as the most promising biotech platform — now with cancer-vaccine potential — being defunded just as China pours money into synthetic biology. Matters as a case study in how vaccine skepticism is eroding an American innovation advantage.
mRNA vaccinesbiotechRFK JrOperation Warp Speedscience policy
TIER 4
Sep 17, 2025
Pethokoukis uses a new NBER paper (van Alten et al.) on the intergenerational transmission of socioeconomic status to argue that genes matter for education, income, and wealth—but roughly half the effect runs through 'genetic nurture' (the environments parents create), which policy can change. He contrasts this grounded social-science result with Silicon Valley's $50K embryo-IQ-screening enthusiasm, concluding better schools and neighborhoods beat brainier embryos as a route to a smarter America. A substantive, well-sourced explainer on a charged topic.
geneticsinequalitysocial mobilityembryo screeningNBER
TIER 4
Dec 18, 2025
A substantive Q&A with Hudson Institute fellow Bill Drexel on how AI is acting as an accelerant for genetic engineering, arguing this overlooked AI-bio nexus may pose more immediate moral peril than speculative rogue superintelligence. Drexel warns of a sleepwalk into a bioethics race to the bottom—driven by China's vast genomic datasets and risk tolerance—and proposes data safeguards, declarations modeled on military-AI norms, and engaging the Global South to constrain eugenic drift.
AI-bio nexusgenetic engineeringChinabioethicsinterview
TIER 4
Feb 18, 2026
Using the FDA's reversal of its halt on a Moderna mRNA flu-vaccine trial as a hook, Pethokoukis argues that political vaccine skepticism is creating regulatory uncertainty that chills the broader mRNA platform—threatening cancer vaccines, MS prevention, and pandemic preparedness. He marshals concrete figures (140,000 US lives saved by COVID mRNA jabs, mRNA venture funding falling from $510M to $174M, ~$500M in canceled federal contracts) and a Scott Gottlieb proposal for Congress to earmark vaccine R&D. A complete, substantive biotech-policy essay.
mRNA vaccinesFDAbiotech policyvaccine skepticismregulatory uncertainty
TIER 4
May 5, 2026
Builds on an NBER paper (Huetsch, Krueger, Ludwig) to argue economic growth is one of history's greatest health interventions: rising incomes raise demand for longer lives, which fuels medical spending and innovation in a self-reinforcing loop responsible for roughly a quarter of postwar life-expectancy gains. Extends the logic to AI as an accelerant of both growth and scientific productivity, recasting the growth debate in decades-of-life rather than dollars. A core statement of the newsletter's thesis.
economic growthlongevityhealthAIdegrowth critique