The headline data look contradictory. Inflation is easing, shelter inflation is finally bending, national rents are now negative year-over-year in multiple series, and yet the unemployment rate is edging higher while labor productivity keeps rising. That’s not a conventional slowdown; it’s a structural rift. Automation and AI are decoupling output from broad-based labor demand, creating a two- (or three-) speed economy where a high-skill, automation-complementary core accelerates as large swaths of routine or easily-replicable work get squeezed.

We’ve written the plumbing side of this repeatedly in our dollar-liquidity pieces (The Dollar Isn’t Collapsing — It’s Evolving) and the adaptation side in our AI essays (Why Most People Won’t Adapt to AI — and Why Polymaths Will and The Age of Algorithmic Influence). This piece ties those threads into a transition framework for policy and society.

Personal aside: the rental market’s direction matters a lot to me as an operator. We’ll keep this focused on the macro mechanics—but recognize that shelter is the most important transmission channel into real lives and local balance sheets.

Signals: Inflation ↓ | Rents ↓ | Unemployment ↑ | Productivity ↑

• Inflation easing into October: The BLS schedule shows the October CPI due mid-November; recent prints were drifting near ~3% y/y (BLS CPI; BLS calendar).
• Rents are sliding: Apartment List reports the national median rent fell 0.8% m/m in October, −0.9% y/y, and now sits ~4.2% below the 2022 peak.
• Unemployment ticking up: With official releases disrupted, the Chicago Fed’s estimate rounded October unemployment to 4.4%, the highest in four years (Reuters).
• Job openings cooling: JOLTS show openings drifting lower from the 2022 peak (latest JOLTS PDF; series home: BLS JOLTS).
• Productivity rising: BLS revised Q2 2025 nonfarm productivity up to 3.3% annualized (BLS Productivity & Costs, Q2 2025 (revised)).

Translation: prices are cooling (especially shelter), hiring demand is less hot, but firms are getting more output per hour. That’s exactly what you’d expect from a world where enterprise AI, workflow automation, and early-stage robotics are spreading through operations. It’s not a paradox. It’s the beginning of a new equilibrium.

The Split: A/B/C Economies

Think in segments:

1. Segment A — Automation-Complementary Core: Capital-dense, model-driven domains where AI/robotics augment output per worker. Value capture concentrates here.
2. Segment B — Routine & Middle-Skill Squeeze: Roles with repeatable task structure face substitution or ruthless standardization—classic “routine-biased technological change” (Brookings).
3. Segment C — Residual/Under-Employed: Regions and cohorts that automation passes by or displaces without quick re-absorption—fertile ground for social stress unless demand is stabilized.

The macro literature now openly acknowledges this exposure: the IMF estimates roughly 60% of jobs in advanced economies are exposed to AI; the OECD warns benefits concentrate without policy; the FSB flags system-level implications.

We’ve mapped this split across power, chips, and capital formation (see Big Tech’s Fusion Bet).

This Isn’t 2008 or 1999

Every feed this morning is framing the economy as if it’s 2008 or 1999. That lens is comforting because it’s familiar—but it’s wrong. Neither 2008 nor 1999 had an at-scale robotics + AI stack about to deploy into physical workflows. Benchmarking today against those eras guarantees you’ll misprice both risk and upside.

The reversion fantasy—“we’ll go back to the 1990s economy”—assumes a society-level breakdown. Absent that, progress does what it always does: it compounds. We’ve endured wars, crises, shutdowns, political and ideological convulsions. We didn’t reverse; we redirected and resumed. That’s the empirical shape of modern history.

The Robot Wave: 2025–2035 Outlook

This is the piece most people miss. You can argue recession odds; you can argue soft-landing versus hard-landing. But underneath the cycle is an adoption curve that doesn’t care about your nostalgia: millions of robots across factories, logistics, retail back-rooms, and eventually light-industrial and service workflows. The first wave will be uneven and awkward—then platform effects kick in (supply chains, safety standards, developer tooling, maintenance networks), and adoption accelerates.

• Factory capacity exists now: Agility’s RoboFab is designed to scale toward ~10,000 humanoid units per year at peak (Agility RoboFab).
• Live industrial pilots: Figure has a humanoid running on a BMW line for extended daily shifts per founder disclosures; coverage and interviews note 10-hour production stints (coverage; see also Time).
• Broader OEM testing: Mercedes-Benz took a stake in Apptronik and is testing humanoids for repetitive/hazardous tasks in multiple plants (Reuters).

That’s not 2050 sci-fi; that’s 2025–2030 commercialization. Multiply pilots; add rival platforms; layer in vertical-specific form factors (arms, mobile bases, bin-picking cells). We won’t replace everything—humans remain unmatched in dexterity and general social intelligence—but repetitive, hazardous, monotonous tasks are squarely in scope. The macro implication: more output with fewer human hours in a growing share of workflows.

Mechanism: Automation → Labor Decoupling

In the classic model, output ↑ ⇒ employment ↑ ⇒ wages ↑. Under automation, the slope flattens: output can rise without proportional hours. That’s exactly what the data combo implies—higher productivity, rent disinflation, and a softer labor market. It’s not a paradox; it’s a regime shift.

The academic backdrop foreshadowed this years ago—job polarization as routine tasks get automated, with gains accruing to the upper end and a long tail of low-paid in-person work (overview: Brookings). Add Gen-AI to back-office, support, sales ops, QA, documentation, and code-adjacent tasks, and the adoption curve steepens.

Winners, Losers, and the Middle That’s Hollowing Out

The split isn’t just about sectors; it’s about functions. Anything repeatable, spec-bound, and data-rich can be automated or aggressively standardized. The prize for firms is margin defense and resilience when demand wobbles. The cost to society is a widening split between those who can work with the machines and those who are displaced by them.

The middle class—especially roles defined by compliance, coordination, and routine transformation of information—faces the ugliest pressure. Some will upskill and move to the complement side. Others won’t, either because the training window closes, or because the new ecosystem rewards different kinds of cognition and temperament. Without a policy bridge, this is how you get an economy that’s “fine” on aggregate and failing in household reality.

Policy Pivot: Why MMT-Style Tools Fit the Terrain

If machines raise the frontier while labor demand narrows, income tied strictly to “having a job” becomes a scarce gate to the economy’s output. To keep demand stable and social cohesion intact, we need fiscal tools that route purchasing power and purpose to displaced cohorts—without blocking automation’s gains.

• Job Guarantee (JG): A standing, federally funded offer at a living wage that anchors prices and absorbs slack (Levy Institute; Tcherneva; Kelton).
• Targeted transfers & public investment: Finance real capacity—skills, grid/nuclear build-out, data stewardship, the care economy—rather than propping up zombie roles (Wray).
• Inflation control via buffers + taxes: Use buffer-stock employment and counter-cyclical taxes instead of unemployment by design (Levy Institute).

Even mainstream bodies now price the disruption: the IMF and OECD warn of distribution risks if policy doesn’t adapt. You don’t have to “be an MMTer” to accept the operating logic in an automation regime.

Related Pattern Nexus frames: Dollar Liquidity Architecture, Adaptation and Polymaths, Exponential Minds & Simulated Realities.

Transition Playbook: Make Work a Choice, Not a Gate

1. Guarantee a floor, not a fiction: Stand up a Job Guarantee that produces useful public goods (care, climate adaptation, community resiliency, digital public works like open-data labeling for public AI) instead of subsidizing pointless roles (Levy Institute).
2. Pay for capacity, not just claims: Aim fiscal firepower at reskilling for model-adjacent trades, grid hardening, small-modular nuclear, and compute-era infrastructure. (See OECD Employment Outlook.)
3. Let shelter normalize: Don’t fight the disinflation in rents while supply catches up (Apartment List).
4. Target the pain, not the engine: If CPI cools while joblessness rises, bridge displaced workers with income/support without capping Segment-A productivity or throttling capex.
5. Measure what matters: Track output per hour (BLS productivity), leading shelter indicators (Apartment List), JOLTS for openings/layoffs (BLS JOLTS), and alt labor gauges during data gaps (e.g., Chicago Fed estimate via Reuters).

This isn’t about “handouts.” It’s about wiring a macro system where automation’s dividend shows up in household stability and social cohesion instead of just margins and buybacks. If we do it right, work becomes a choice—a domain for human creativity, care, and complexity—rather than a gate to basic survival.

2025→2035 Timeline: What Changes When

• 2025–2026: Enterprise AI saturates back-office and knowledge workflows. Robotics pilots proliferate in auto, logistics, and component handling. Rents continue to normalize from the 2021–2022 overshoot (Apartment List).
• 2027–2029: Platform-ized robots (common hardware + app layer) unlock more stations per plant; duty cycles improve; total cost of ownership drops. Labor-light expansions become feasible. Policy debates harden around JG vs UBI; some regions pilot hybrid approaches.
• 2030–2032: “Semi-autonomous everything” in routine industrial tasks; high-reliability human-in-the-loop systems in service sectors; meaningful penetration in retail back-rooms, depots, and fulfillment centers. Measurable decoupling of hours from output in official statistics.
• 2033–2035: Multi-vendor robot ecosystems and standardized safety/regulatory frameworks; public-facing deployments in limited service contexts; education/employment models adjust to a world where “role = portfolio of capabilities” rather than “job = fixed task bundle.”

Risks, Misreads & What Could Break

Misread #1: “Highest unemployment ever” equals 2008 front-run. The Chicago Fed’s modeled jobless rate ~4.4% is a four-year high, not a Great Recession analog. Context matters; composition matters (Reuters).

Misread #2: “Robots will do nothing for a decade.” Capacity and pilots exist now (Agility RoboFab; Figure at BMW; Mercedes/Apptronik). The first waves will underwhelm—and then they won’t (Agility; Figure coverage; Reuters).

Risk #3: Policy paralysis. If we refuse to modernize fiscal architecture, we will get the worst of both worlds: disinflation, profit concentration, and a politically explosive, under-employed class.

Risk #4: Over-tightening into supply-led disinflation. If rents and goods are normalizing while productivity rises, blanket suppression worsens the split. Use targeted support and investment instead of economy-wide brakes.

What to Track (A Practical Dashboard)

• Productivity: BLS Nonfarm Business Productivity (quarterly) — inflects when automation bites (BLS PDF).
• Shelter: Apartment List m/m changes; Zillow ZORI for leading trend (Apartment List; Zillow ZORI).
• Labor demand/turnover: JOLTS openings, quits, and layoffs (BLS JOLTS).
• Unemployment (when official data are delayed): Chicago Fed estimates and private trackers (Reuters).
• Robotics adoption: OEM pilot announcements; factory capacity disclosures (Agility RoboFab; Mercedes/Apptronik; Figure/BMW) (Agility, Reuters, coverage).
• CPI schedule: Know when the next shelter print hits (BLS CPI calendar).

Sources

• BLS — Consumer Price Index (latest release page)  |  BLS — CPI Release Calendar
• Apartment List — National Rent Report (Oct 2025)
• Zillow Research — Housing Data (ZORI)  |  ZORI Methodology
• Reuters — Chicago Fed estimate puts Oct unemployment ~4.4%
• BLS — Productivity & Costs, Q2 2025 (revised)
• BLS — JOLTS (latest PDF)  |  JOLTS Home
• Agility Robotics — RoboFab capacity  |  RoboFab announcement
• Figure AI at BMW — coverage  |  Time — Figure 03 feature
• Reuters — Mercedes-Benz stake in Apptronik; factory tests
• IMF — AI will transform the global economy
• OECD — AI: Productivity, Distribution & Growth
• FSB — The AI Adventure: Economic & Financial System Impacts