AIPE Atlas
AIPE Atlas
Twelve openings on the path — each anchored to a need from the ledger and a force from the inversion. Near-term plays ride the transition; bridge plays build the rails between horizons; endgame plays construct what a post-labor society will need most. Filter, then expand.
AI levels up novices most: +14% average, +34% for novices in the landmark support-agent study; the bottom of the skill distribution gains most in writing and coding. The augmentation path is a choice firms are actively making — or failing to make.
Build tools that make mid-skill workers perform expert-reserved tasks safely: clinical decision support for nurses and aides, code review for junior engineers, paralegal-to-counsel lifts. Autor's "rebuild the middle class" scenario is a product roadmap.
Autor (Noema, NBER w32140) · Brynjolfsson, Li & Raymond (NBER w31161) · Brynjolfsson's Turing Trap warning as the design constraint.
Entry-level employment in AI-exposed jobs down ~13% for ages 22–25; big-tech new-grad hiring down >50% in three years. Firms have stopped buying junior labor — but they still need future seniors. Nobody has solved how juniors become seniors when AI does the apprentice work.
Simulation-based apprenticeships, AI-supervised deliberate practice, credentialing that proves judgment rather than output, employer consortia that share the cost of training the cohort AI skipped.
Stanford "Canaries in the Coal Mine" (2025–26) · Yale Insights on pre-career job destruction · the expertise-pipeline worry in the KB's workforce file.
Gartner's 2026 framing: brands have entered an era of trust scarcity — credibility signals, expert proof, and human points of view now determine discoverability and preference. Harari's Nexus warning is the civilizational version: democracy itself runs on shared, verifiable reality.
Provenance and human-made certification, deepfake detection, reputation rails, "proof of human" identity, audit and accountability tooling for AI outputs. The scarce asset is not information — it is the warrant behind it.
Gartner (May 2026) on trust scarcity · Harari, Nexus · the human-made premium studies (Psychology & Marketing, 2025).
Demographics are destiny: US healthcare adds ~2M jobs this decade, more than any sector; home health aides are the single largest occupational increase; nurse practitioners grow at 13× the national average. Baumol's cost disease guarantees these services get relatively more expensive as everything else gets cheap — because the human is the product.
Augment the carer, never replace the care: admin-killing copilots for aides and nurses, family-coordination platforms, eldercare logistics, training that lifts aides into practitioner tracks. WEF's largest absolute growth is frontline and care roles.
BLS Employment Projections 2024–34 · WEF Future of Jobs 2025 · Baumol (UNESCO analysis) · Autor on care as the most-cited growth area.
Across every source in the KB, the human capabilities that retain value converge: expert judgment on high-stakes one-off cases, and the bearing of responsibility no machine can carry. When generic competence is free, the scarcity value of accountable expertise rises — the "expertise paradox."
Human-in-the-loop sign-off as a product category: AI-era audit, fiduciary review, safety certification, "accountable counterpart" services for medicine, law, finance, and engineering. Sell the signature, not the draft.
The KB's cross-source synthesis of retained human capabilities · PwC's 56% wage premium for AI-skilled roles · the accountability scarcity argument.
Experimental evidence: labeling work AI-generated lowers willingness-to-pay and perceived creativity; human-made commands a premium. The Culture's composer got the same answer from a Mind: listeners value that a member of their species was the creator.
Human-made certification and marketplaces, craft provenance (the "vinyl revival" pattern generalized), live performance, artisan food and goods, human-taught masterclasses — the premium tier of every category machines commoditize.
Jin (Psychology & Marketing, 2025) · ECU and RSM consumer studies · Banks, Look to Windward · Amodei's adoption of the same logic.
US consumers already spend ~$2.1 trillion on experiences; spending on physical experiences surged 65% from 2019–2023; most buyers say experiences beat possessions on value-for-money. Pine & Gilmore's sequence — commodities → goods → services → experiences → transformations — is the consumer-side mirror of the inversion.
Live events, pilgrimage-grade travel, transformation programs (fitness, retreats, skill quests), "the real" as luxury. Banks's citizens go camping; Star Trek's crew runs the holodeck but values the vineyard. Scarcity of the authentic is permanent.
Pine & Gilmore (HBR 1998; transformation economy) · Empower/Morning Consult 2025 spending data · the KB's "scarcity re-emerges in post-scarcity" sci-fi theme.
The single point of near-universal agreement in the KB: meaning must be deliberately reconstructed — through relationships, mattering, contemplation, craft, play, faith — and will not arrive automatically with abundance. Lessin forecasts a resurgence of religion; Jahoda's five latent functions are an unmet product spec.
Institutions that supply time structure, social contact, collective purpose, status, and activity without a job attached: secular congregations, guilds and leagues, civic service corps, contemplative education. Susskind's "leisure policy" implies a whole public sector that doesn't exist yet.
Jahoda's latent-deprivation theory · Frankl · Lessin's meaning-crisis essay · Pieper and the Sabbath tradition · Amodei's "relationships, not labor."
Suits: game-playing — "the voluntary attempt to overcome unnecessary obstacles" — is the ideal of existence once machines do all instrumental work. Storr: status drives are wired in and migrate to new arenas. Positional scarcity is the one demand automation cannot satisfy, only relocate.
Competitive crafts and leagues, exploration challenges, esports' physical successors, mastery academies, games engineered for genuine mattering rather than escapism — designing against the pessimists' warning that empty play deepens the void.
Suits, The Grasshopper · Danaher, Automation and Utopia · Storr, The Status Game · Hirsch on positional goods · the WALL-E warning as the design anti-pattern.
The most telling trend of the 2020s: AI insiders migrating from cash UBI to ownership — Altman's American Equity Fund and "universal basic compute," Berggruen's universal basic capital, Sanders's AI sovereign-wealth fund proposal. The Alaska Permanent Fund proves the mechanism at state scale.
Sovereign and social wealth fund administration, fractional equity distribution, dividend rails, employee-and-citizen ownership tooling, the registry and governance layer for "who owns the machines." Weyl's critique is the design constraint: passive income without decision rights is not agency.
Altman, Moore's Law for Everything · Berggruen & Gardels · Alaska Permanent Fund · Korinek & Stiglitz (NBER w24174) on why ownership is the durable channel.
Data-center electricity demand grew 17% in 2025; AI-focused consumption surged 50%; power is now "the defining constraint" on AI growth (IEA). Noah Smith's own caveat bites here: if humans must outbid AIs for energy and land, the comparative-advantage defense of wages breaks.
Generation, storage, grid software, siting, efficiency — plus the governance layer: compute is "detectable, quantifiable, and supply-chain concentrated," making it the practical lever for both profit and policy. Whoever widens the energy bottleneck shapes who gets abundance.
IEA Energy & AI reports (2025–26) · CNAS on compute concentration · Smith's scarce-inputs exception · Hassabis's "energy zero-carbon and free" as the bull case.
The Intelligence Curse's grim logic: when elites no longer need people as workers or soldiers, the incentive to invest in them erodes — unless people acquire new leverage first. Buterin's d/acc names the counter-strategy: we choose which technologies to build.
Personal AI that answers to its owner, decentralized coordination tools, citizen-auditable government AI, defensive security for individuals and small institutions, democratic-infrastructure tech. The normative checklist: regular people can create value, disrupt elites, and control their destiny after AGI.
Drago & Laine, The Intelligence Curse · Buterin's d/acc · Bullock et al. on the narrow corridor under AGI · Hoffman's superagency as the optimistic frame.
Where the evidence and the arguments converge, design principles fall out. These seven are the load-bearing beams — each one traceable to the atlas above, each one actionable now.
The augmentation path is a choice, not a law. "So-so automation" cuts labor cost without creating offsetting growth; levers for expertise rebuild the middle. Every product decision is a vote for one path. (Brynjolfsson's Turing Trap; Acemoglu & Johnson; Autor.)
Ordinary people's power came from being needed — as workers and soldiers. That foundation is eroding on both fronts at once. The replacement is ownership and access, and the window to install it is while labor still has bargaining power. (Intelligence Curse; Altman's pivot from cash to equity; predistribution.)
Time structure, social contact, collective purpose, status, activity: a job supplied all five invisibly. A post-work society must supply them deliberately — like sanitation, not like luck. (Jahoda; Frankl; the near-universal convergence of the meaning literature.)
"Democracy is likely a necessary precondition for breaking the intelligence curse." AGI amplifies both state power and non-state chaos; the narrow corridor between Despotic and Absent Leviathans has to be defended on purpose. Comfort without power is the recurring nightmare — even under a benevolent master.
Trust, judgment, care, attention, authenticity, the real: demand is migrating there measurably, now. Markets, institutions, and careers built on the new scarcity stack are aligned with the century's direction of travel. (Baumol; the trust-scarcity turn; the human-made premium.)
Act on Horizon A's evidence — reskill, repair ladders, augment care — while building Horizon B's institutions before they're needed. Stiglitz holds both positions without contradiction: no UBI yet; sovereign funds eventually. Scenario-plan like Korinek: prepare for the branch you can't rule out.
Manna's two endings run on the same robots. The Engels' Pause was "a function of collective choices." The deciding variable in every scenario — fictional and economic — is never the technology; it is distribution, ownership, and who controls the machines.
The economic crux. Unbounded: wages rise forever. Bounded: the horse analogy. Watch: whether AI starts generating genuinely novel problems, not just solving given ones.
The timing crux that separates the horizons. Watch: robotics cost curves in care settings and trades — Autor's refuge either holds for decades or it doesn't.
Individuals have always found purpose beyond work. No civilization has done it at scale. Watch: the experiments — UBI cohorts, post-work communities, the religion resurgence Lessin predicts.
Thermodynamics, planetary boundaries, and the energy bottleneck on one side; artificial scarcity and rent preservation on the other. Watch: the data-center power crunch as the first real test.
The honest bottom line of the entire literature: the outcome is not yet determined. The economics decides whether there is abundance; the politics decides who claims it; the distribution decides how; and the meaning decides whether it was worth it. All four are still being written — which is the most important point of agreement of all.
The operational sequel: WORKBank × JobBench — where delegation actually works today, measured and priced.