LaSalle Browne · AlphaSage · Strategic Framework · 2026

Industrial
Decoherence

The unified diagnosis beneath five separate crises

America does not have an innovation problem. It does not have a capital problem. It does not even, at its core, have a manufacturing problem. It has a capability formation problem — the progressive loss of the individual, organizational, and institutional capacity to build hard things at scale, reliably, over time. And until that is understood, every solution will feel directionally correct and still fall short.

The Symptom
Manufacturing Gap
The Disease
Capability Formation Failure
01

The Symptom Map

Five domains. Six root causes. One capability formation failure expressing itself in different materials. Click any domain or root cause to explore the connections.

INDUSTRIAL DECOHERENCE THE ROOT THESIS ROOT CAUSE DOMAIN
Interactive Diagnostic Map
Select a Node to Explore
Click any of the five domain nodes (outer ring) to see its symptoms, shared root causes, and essay coverage.

Click any root cause node (inner ring) to see which domains share that cause and how the same failure manifests across each.
■ EDUCATION   ■ AI / TECH   ■ MANUFACTURING

■ INFRASTRUCTURE   ■ NAT'L RESILIENCE
02

The Evidence Trail

Twelve sources mapping the capability formation failure from different angles — supply chain cascades, cultural psychology, institutional design, and on-the-ground proof points. Read together they form a complete diagnosis and a partial map of the response.

Essay 01
The Trouble with Transformers
Rob L'Heureux
InfrastructureManufacturing
The US has one GOES producer, a 10–30% transformer deficit, and a regulatory environment that tried to ban the only available material. The grid is pre-industrial. The escape hatch from the bottleneck — solid-state transformers — runs directly through GaN, which is itself a dependency we don't control.
Essay 02
The US Is Repeating Its Silicon Mistake with Gallium Nitride
Pradyot Yadav
ManufacturingAI / Tech
GaN is critical for next-gen power systems, defense electronics, radar, and space. America is ceding the substrate layer again while holding the application crown — repeating the silicon mistake at the power semiconductor layer.
Essay 03
The Great Blue Frontier
Packy McCormick & Will O'Brien
Nat'l ResilienceManufacturing
70% of the planet — carrying 99% of internet traffic, 80% of trade, and more critical minerals than all land reserves — is pre-industrial and essentially unmonitored. The ocean stack has converged. The question is who builds it and how fast.
Essay 04
How to Build American Shenzhen, Fast
Aaron Slodov
Manufacturing
The most rigorous publicly available blueprint for American reindustrialization. Density — co-located HMLV manufacturing clusters — produces the iteration velocity that makes Shenzhen competitive. Slodov gets the stack right. The question is what organizing principle sits underneath it.
Essay 05
Don't Build American Shenzhen
LaSalle Browne · AlphaSage
ManufacturingEducation
Density is the mechanism. Metabolic rate is the prize. A model-defined, distributed manufacturing system compounds learning at machine speed. China can copy a factory. It cannot copy what the factory has learned or how that learning propagates.
Essay 06
Education Convergence Architecture (Parts I–III)
LaSalle Browne · AlphaSage
EducationManufacturing
The human substrate layer is decohering in parallel with the physical substrate. Operator populations are replacing architect populations. The developmental sequence cannot be compressed. The 15-year pipeline must start now, simultaneously with physical infrastructure buildout.
Essay 07
Strategically De-Risking U.S. Drug Supply Chains
Jordan Despanie & Timothy Marler · RAND Corporation
ManufacturingNat'l Resilience
30% of FDA-approved drug workflows offshored to Chinese CDMOs; 74% of preclinical and clinical services. Replacing capacity takes 8 years. China controls 23% of global innovative drug development. The hollowing sequence has progressed into R&D itself. RAND recommends strategic taper over decoupling: allied nearshoring as near-term countermeasure while domestic capability is rebuilt through targeted incentives and advanced market commitments.
Essay 08
Strategic API Sourcing: Offshoring vs. Nearshoring
Ash Gude · Pharmaceutical Online
ManufacturingInfrastructure
The second-order dependency most analyses miss: Indian API manufacturers depend on Chinese key starting materials upstream. "China plus India" dual-sourcing doesn't resolve the KSM concentration. Nearshoring is economically viable for complex and biologic APIs — not for commodity APIs without policy subsidy. The binding constraint is upstream, not downstream. Mirrors the rare earth argument precisely.
Initiative 09
Re:Build Manufacturing
Founded 2020 · $545M raised · 1,100+ employees across US
ManufacturingEducation
The most sophisticated existing Track 1 institutional initiative. Integrated design, engineering, and manufacturing from concept to full-rate production. Slate platform delivers automation, data capture, and process control. Manufacturing Constrained Design™ evolves product and production systems simultaneously. The gap that defines Track 2's opportunity: Slate is facility-specific automation — process knowledge doesn't propagate across subsidiaries at machine speed. Re:Build is a natural Phase 2 MDS pilot partner.
Initiative 10
Proto-Town
Farahzad & Nye · Lockhart TX · $100M valuation
ManufacturingNat'l Resilience
1,200-acre campus: autonomous bulldozers, basketball court drones, water systems, a nuclear reactor. Oklo: greenfield to fission in 10 months — Manhattan Project speed. Caldwell County's minimal-regulation deal is a working prototype of the manufacturing zone sandbox. The gap: co-location produces iteration velocity; no intelligence layer connects the twelve companies. Proto-Town proves builder identity is real and speed is achievable. It does not yet prove the metabolic rate architecture.
Essay 11
The "Boring" Industrial Shortage That's About to Break Everything
Jay Martin · Jay's Letter on Substack
InfrastructureManufacturingNat'l Resilience
The sulphur cascade: Persian Gulf closure → China sulphur shortage → Chinese industrial chemical shortage → mining operations across DRC, Chile, Indonesia, and Peru rationing production → copper, cobalt, and battery material shortages everywhere. China manufactures ~45% of global industrial chemical supply. "Oil keeps the engine running. But it does not build the engine." Correctly diagnoses the symptom cascade; stops short of the capability formation architecture that constitutes the response.
Essay 12
Why America Can Build After All (And: The One Chart That Explains Everything)
Howard Yu · IMD Center for Future Readiness
EducationManufacturing
America can build with terrifying speed when the builder is rich enough and powerful enough — data centers prove it. The question is whether America builds in ways that make ordinary people feel invited into the future. Two Boxes: Box One is current core competency; Box Two is the capability stack for five years from now. Kodak mastered Box One. The US has maximized Box One (intelligence layer) while hollowing Box Two (substrate capability stack). Trust, broken by platform extractivism, is what makes cultural rewiring harder than narrative alone can fix.
Essay 13 · Most Critical
The Real Rival to Silicon Valley Is Not a Chinese City. It Is a Chinese System.
Leon Liao · China as a System · April 2026
ManufacturingAI / TechNat'l Resilience
China is not building a single Silicon Valley equivalent. It is building a closed loop: research → engineering → manufacturing → deployment → data → capital → research. Of 16,000 humanoid robots deployed globally, 80%+ are in China. CATL holds 42% of global battery share. Chinese shipyards took 53% of global new ship orders vs. fewer than 10 US commercial vessels built last year. China's manufacturing robot density: 470/10,000 workers vs. 295 in the US. What China is challenging is not one industry, nor one city, but an entire method of organizing technological progress. This is the most important external validation of the capability formation thesis currently available — described from the adversary's perspective.
The Sulphur Cascade: How a shipping disruption propagates through Chinese chemical dependency into global mining, metal supply, infrastructure timelines, and economic output — Jay Martin's cascade mechanism visualized.
Supply Chain Cascade The Sulphur Cascade: How a shipping disruption propagates through Chinese chemical dependency into global mining, metal supply, infrastructure timelines, and economic output — Jay Martin's cascade mechanism visualized.

Read individually they describe separate crises and partial solutions. Read together — including Liao's account of China's closed loop — they map a single capability formation failure and its strategic consequence: the adversary is already building Track 2. The US private market response is necessary and not yet sufficient.

— The Unified Thesis
03

Six Root Causes — One Pattern

The same capability formation failure appears across all five domains through six structural drivers. Fixing domains independently without addressing these root causes produces expensive patches over a systemic condition. Root Cause 06 is the root of the roots.

Root Cause 01
Time Horizon Compression
Every institution governing critical systems operates on 2–7 year cycles. Every critical system requires 15–30 year investment horizons. The gap is where all five domains fail simultaneously. No rational actor inside current institutions can hold the commitment the problem requires.
EducationManufacturingInfrastructureAI / TechResilience
Root Cause 02
Efficiency Over Capacity
Thirty years of rewarding optimization within existing systems over construction of new ones. Shareholder primacy, quarterly earnings, regulatory cost-benefit analysis — all point at efficiency. The institutional machinery that could choose capacity has been systematically atrophied.
EducationManufacturingInfrastructureAI / TechResilience
Root Cause 03
Financialization of Maintenance
Private equity ownership of infrastructure systematically defers maintenance to maximize cash extraction within the hold period. Infrastructure degrades invisibly until it fails visibly. Colonial Pipeline, Boeing, aging dams, nuclear atrophy — the same pattern across domains.
InfrastructureManufacturingResilience
Root Cause 04
Regulatory Fossil Layer
In every domain, the regulatory architecture is a fossil of the world it was written for — actively obstructing adaptation. Nuclear regulation for 1970s LWRs. Transmission permitting before renewable geography mattered. Maritime law before autonomous systems existed. The counter-proof: Oklo built a nuclear test reactor at Proto-Town TX in 10 months under a minimal-regulation county deal. Vogtle took 16 years under standard process. The technology didn't change — the regulatory environment did.
InfrastructureAI / TechResilience
Root Cause 05
Substrate Offshoring Sequence
Manufacturing moves offshore for cost → Engineering follows → R&D follows → Design capability atrophies because there is nothing to design for domestically. The hollowing sequence runs in one direction and does not reverse easily.
ManufacturingAI / Tech
Root Cause 06 · The Root of Roots
Intelligence–Substrate Decoherence
Howard Yu's Two Boxes: the US maximized Box One (intelligence layer — software, algorithms, finance) while hollowing Box Two (the substrate capability stack). Leon Liao shows the consequence: China has built a closed loop — research feeding engineering feeding manufacturing feeding deployment feeding data feeding research — across AI, EVs, batteries, robotics, drones, and shipbuilding. The US must build a competing closed loop. A better Box One alone will not be sufficient — civilizational scale. Dan Wang's Engineering State vs. lawyer-financial state is the cultural expression.
EducationManufacturingInfrastructureAI / TechResilience
04

The Full Diagnostic

The transformer shortage and the GaN dependency are representative samples — not anomalies. The same structural signatures appear across every critical domain. Breadth and depth both exceed what the public discourse acknowledges.

1
US producer of
grain-oriented steel
232×
China/US ratio
in annual shipbuilding
85%
Rare earth processing
controlled by China
27%
Ocean floor
mapped to date
D+
ASCE grade for
US water infrastructure
17yr
SunZia transmission
permitting timeline
Energy & Grid
Transformer Supply
10–30% deficit by type. 1 US GOES producer. 15+ year lead times for large power transformers. Regulatory own-goals froze all new investment.
Semiconductors
GaN & Advanced Materials
China-dominant GaN substrate. ASML EUV monopoly — only source of leading-edge lithography globally. 90% of PCBs manufactured in Asia.
Defense Industrial
Shipbuilding & Munitions
China builds 232× more tonnage annually. Single Pennsylvania plant produces majority of 155mm shells. Virginia-class subs 2+ years behind per boat.
Materials
Rare Earth Magnets
85% processing, 90% magnet production in China. Mountain Pass ore ships to China for processing. Every precision munition, every EV motor, every radar depends on this chain.
Ocean Domain
Maritime Presence
73% of seafloor unmapped. 100–200 cable breaks/year unmonitored. 2 US heavy icebreakers vs Russia's 40+. China building Underwater Great Wall.
Infrastructure
Nuclear & Grid
Zero new reactors built. Vogtle: 16 years, $35B, 4× over budget. No ports in global top 50. 240,000 water main breaks per year.
Industrial Chemicals
Sulphur Cascade
China manufactures ~45% of global industrial chemical supply. Mining in DRC (cobalt), Chile (copper), Indonesia (nickel), Peru (copper) all depend on Chinese sulphuric acid and processing chemicals. Persian Gulf disruption → Chinese sulphur shortage → global mining slowdown → copper, cobalt, battery material shortages everywhere.
Pharmaceutical
API & CDMO Dependency
30% of FDA-approved drug workflows offshored to Chinese CDMOs; 74% of preclinical & clinical services. Replacing capacity takes 8 years. China controls 23% of innovative drug development globally. Indian API suppliers are themselves dependent on Chinese key starting materials — dual-sourcing doesn't resolve the upstream concentration.
Food Systems
Fertilizer & Processing
Russia/Belarus control ~40% of global potash. Four companies control 80% of US beef processing. Agricultural equipment chip dependency exposed in 2021 shortage.
Systemic Chokepoints & Decoherence: From Visible Problems to Hidden Misalignments — Ten critical bottlenecks mapped from surface symptom to the deeper structural misalignment beneath each.
Diagnostic Visual Systemic Chokepoints & Decoherence: From Visible Problems to Hidden Misalignments — Ten critical bottlenecks mapped from surface symptom to the deeper structural misalignment beneath each.

These are not five separate problems. They are one capability formation failure in different materials — the progressive loss of the capacity to build hard things at scale, expressing itself simultaneously across every domain the US must compete in over the next thirty years.

— The Diagnostic Conclusion
05

Metabolic Rate Is the Prize

Jay Martin correctly argues that old factories that lost in 1995 can't win in 2026. He's right about the old model. He doesn't address the new one. And Leon Liao shows that China is already building its own closed loop — which means Track 2 is not insurance against being copied. It is the only viable competing architecture. Density is the mechanism. Metabolic rate is the prize.

The factories are not the constraint. The capability to form the people and institutions that build, evolve, and govern the factories — at the pace the moment requires — is. China is already building its closed loop. The US response is not to copy it, but to build a competing one: faster-learning, more distributed, governed by architects rather than directed by the state.

— The Closed Loop Response
DimensionDensity-First (American Shenzhen)Metabolic Rate (Distributed)
Speed advantageHigh in normal conditionsModerate initially, accelerates as nodes compound
ResilienceSingle point of failure risk under adversarial stressAntifragile — disruption at one node strengthens others
Knowledge transmissionGeographic proximity, human apprenticeshipModel propagation at machine speed across distributed nodes
Long-term trajectoryPlateaus at local optimumCompounds — each node makes all others smarter
Adversarial targetingHigh-value concentrated targetDistributed — no single node is decisive
Replication difficultyReplicable by building a nearby clusterAdvantage lives in models — cannot be copied by building factories
06

Both Tracks, Now

China already runs both tracks simultaneously — American Shenzhen-style density across major cities AND a closed intelligence loop connecting manufacturing to deployment to data to research. The US must match this architecture, not just the volume. The tracks are mutually reinforcing: clusters seed the intelligence layer, the intelligence layer makes clusters compound.

Track 1 · Immediate
American Shenzhen Clusters
Close the immediate critical gaps. Leverage VC and tech elite momentum because it's real and deployable now. Accept concentration risk as a transitional cost. Use clusters as training nodes for the intelligence layer.
  • Closes capability gaps in 5–7 years with concentrated capital
  • Leverages existing co-location infrastructure and workforce
  • Defense demand pull provides revenue legitimacy and urgency
  • Generates the production data that feeds the MDS intelligence layer
  • Proof-of-concept nodes where the metabolic architecture is trained
Track 2 · Strategic
Metabolic Rate Infrastructure
Build the model-defined, distributed, high-learning-rate system that China cannot replicate by copying factories. The intelligence layer is the strategic asset. Takes longer but compounds harder.
  • Model-Defined Systems (MDS): process knowledge in models, not heads
  • Zonal Governance: control plane above the data plane — the missing layer
  • PhysicsML substrate: each node makes all others smarter at machine speed
  • Distributed nodes: resilient, antifragile under adversarial pressure
  • Propagates cluster learning to the distributed network automatically
Integration Layer · MDS + Zonal Governance Architecture
Physical density→ Slodov's co-located HMLV manufacturing clusters (Track 1)
Intelligence architecture→ Model-defined factories with PhysicsML substrate (Track 2)
Zonal governance→ Control plane: timing, trust, liability, safety across distributed nodes
Integration effect→ Cluster learning → MDS → distributed nodes at machine propagation speed
Strategic outcome→ What China cannot replicate by building a factory nearby
07

The Strategic Blueprint

Seven layers, ordered from foundation to outcome. Not sequential — parallel and mutually reinforcing. The cultural shift makes institutional reform politically viable. Institutional reform creates capital structures. Capital structures fund infrastructure. Infrastructure enables manufacturing architecture. Manufacturing architecture runs on the human developmental pipeline.

Layer 01 · Foundation
Cultural Rewiring
Shift from consumption to creation. But cultural rewiring is harder than narrative alone. Howard Yu's diagnosis: thirty years of platform extractivism — Amazon taking 46¢ of every seller dollar until businesses filed for bankruptcy — pre-loads every "we're here to help" with suspicion. People resist not because they're anti-industrial but because they've been trained to assume new initiatives extract rather than form. The shift requires demonstrating repeatedly that capability formation lifts participants. Proto-Town and Re:Build are the frontier evidence. The broader culture requires a new track rectality. Without this, all downstream interventions fight cultural headwinds.
  • Forward-facing construction narrative at the highest leadership levels — not nostalgia, an aspirational frame
  • Status re-signaling: compensation and visibility for strategic manufacturing and engineering roles
  • Entrepreneurship reframe: physical world moats are deeper and more defensible than software moats
  • Maker culture elevated from hobbyist track to prestigious educational and professional pathway
Layer 02
Institutional Architecture
Create entities capable of holding long-horizon commitments and integrating across fragmented mandates. The time horizon mismatch and mandate fragmentation root causes require institutional solutions — new institutional forms with the right incentive structures built in, not individuals behaving against their incentives.
  • Industrial Resilience Authority: 25-year mandate, capital-deploying, Fed-style insulation from electoral cycles
  • Domain-based integration replacing function-based fragmentation — single entity owns full stack per strategic domain
  • OTA expansion + milestone-based contracts rewarding capability demonstration over specification compliance
  • Regulatory architecture reform: domain-specific, forward-looking, technology-neutral where possible
  • Caldwell County/Proto-Town as regulatory sandbox proof: minimal county rules produced nuclear reactor in 10 months vs. Vogtle's 16 years — scale this model to federal manufacturing zone designation
Layer 03
Capital Structures
Restructure what capital responds to rather than asking capital to behave against its incentives out of patriotism. The market isn't failing — it's responding correctly to broken incentive structures. Fix the structure, not the actor. Stop arguing markets vs. government; map the incentive structures required and design institutions that produce them.
  • Long-term offtake guarantees: eliminate demand uncertainty over payback horizons (highest-leverage single intervention)
  • Strategic stockpiling of GOES, GaN, rare earth magnets, distribution transformers — signals durable commitment to private capital
  • Patient capital vehicles: infrastructure fund structure, income-generating, utilization-tied returns, 20–30 year duration
  • Blended public-private capital: defense demand pull anchors revenue, public capital covers the developmental infrastructure return
  • Pharmaceutical-specific: BARDA-style advanced market commitments for domestic API and biologic manufacturing — complex and biologic APIs nearshore on market economics; commodity API reshoring requires demand-side subsidy
Layer 04
21st Century Infrastructure
Not infrastructure upgrades — phase transitions. Each creates a generational platform that compounds for decades. Reindustrialization on 20th century infrastructure produces a more sophisticated version of the same fragility. The metabolic rate system is bottlenecked by the infrastructure it runs on.
  • Superconducting power transmission: liberates manufacturing location from grid geography entirely
  • Autonomous logistics corridors: underground freight networks decoupling port congestion from surface traffic
  • Quantum communications: physically guaranteed secure transmission of the intelligence layer — not computationally secure
  • Small modular reactors: distributed baseload power at the manufacturing node level, no transmission dependency
  • Ocean industrialization: persistent domain presence, cable monitoring, critical mineral access beyond territorial reach
  • Advanced mining: DLE, electrochemical rare earth separation, autonomous underground extraction changing materials economics
Layer 05
Manufacturing Architecture
Two parallel tracks, not a choice. The intelligence layer — MDS + Zonal Governance — connects them, turning cluster learning into distributed capability and cluster nodes into seeds for the metabolic rate network. Slodov's physical architecture plus the model-defined intelligence layer is the complete thesis.
  • Track 1: American Shenzhen HMLV clusters — close immediate critical gaps, seed the intelligence layer
  • Track 2: Model-Defined Systems — process knowledge in models that survive disruption and propagate at machine speed
  • Zonal Governance: the control plane above the data plane — Slodov's Common Manufacturing Protocol needs this layer
  • PhysicsML substrate: converts production data into transferable dynamics — each node makes all others smarter
  • Antifragility design: distributed nodes; adversarial stress improves the network rather than breaking it
  • Re:Build Manufacturing as natural Phase 2 pilot partner: $545M raised, 1,100+ employees, integrated concept-to-production, defense and aerospace relationships — add the MDS intelligence layer above their Slate platform to create the distributed compounding system
Layer 06
Human Development: The Core
If the problem is capability formation, this is not Layer 6 of seven — it is the core of the entire argument. The Education Convergence Architecture's five layers (embodied foundation → constraint navigation → iterative mastery → AI orchestration → co-evolution) describe how capability is actually formed. Every other layer creates conditions. This layer is the mechanism. Without it, factories get built and plateau. With it, they compound. The fifteen-year sequence cannot be compressed. It must start produces durable strategic advantage or deferred dependency.
  • Five developmental layers applied to manufacturing: Embodied Foundation → Constraint Navigation → Iterative Mastery → AI Orchestration → Co-Evolution
  • AI-enabled developmental assessment: the only technology that makes process-based assessment scalable across a workforce
  • Cognitive gym structure: protect struggle where struggle is the formation mechanism — physical intuition before abstraction
  • Decision logs + iteration stacks: assess orchestration quality and iteration quality, not output rate
  • Evolution ledgers: track capability frontier advancement — operator vs. architect developmental trajectory over time
  • Pharma manufacturing: 80% skills mismatch with evolving job requirements documented (ISPE 2024) — pharmaceutical manufacturing is a high-priority domain for developmental architecture deployment given the operator/architect gap and the national security stakes of the supply chain
Layer 07 · Outcome
Substrate Sovereignty
The strategic outcome all six layers are building the capability to achieve. Material sovereignty (GOES, GaN, rare earths, APIs — at the processing layer, not just mining). Domain sovereignty (ocean, Arctic, space — persistent presence, not episodic campaigns). Knowledge sovereignty (process intelligence in models that survive disruption and propagate). Architect sovereignty: a manufacturing workforce whose capability compounds while competitors plateau at operator level. This is what cannot be purchased. It must be factories have learned and how that learning is governed.
  • Material sovereignty: GOES, GaN, rare earths, SiC — domestic processing capacity, not just mining rights
  • Domain sovereignty: persistent presence in ocean, Arctic, and space — not episodic expeditions
  • Knowledge sovereignty: process intelligence in models that survive workforce disruption and propagate across nodes
  • Architect sovereignty: a manufacturing workforce whose capability compounds while competitors plateau at operator level
08

The Human Layer: The Core

If the problem is capability formation, this is not one section among ten — it is the core mechanism. The five developmental layers describe how capability is actually built. Every other section creates conditions. This section is the answer to the question: how, exactly, does capability form?

01
Embodied Foundation
Education: Physical intuition, sensory integration, embodied awareness — the pre-cognitive substrate everything else runs on.
Manufacturing: Physical intuition with materials and machines. The machinist who hears tool failure coming. The welder who reads heat color to judge penetration. Irreducibly embodied — cannot be learned from a screen. Assessment: narrative trajectory documented by master craftspeople, not test scores.
02
Primal Intelligence
Education: Robustness under uncertainty. Functional when the path is unclear. Fletcher's "no answer key" problems as formation tool.
Manufacturing: Constraint navigation. Solve production problems with no established solution. Work around unplanned material substitutions. Redesign processes on the floor. Assessment: Challenge Logs documenting approach, constraint shift, adaptation, what held up vs. broke.
03
Adversity Architecture
Education: Reach capability through cycles of failure, feedback, and adjustment. The iteration stack is the artifact — not the final output.
Manufacturing: Structured production failure as formation. Document the full chain of attempts and adjustments. Assessment: Iteration Stacks — quality of adjustment between iterations, evidence of learning transfer to adjacent production challenges.
04
Partnership Capacity
Education: Direct the boundary between own thinking and AI contribution. Orchestration quality is the assessment — not output quality.
Manufacturing: Know when to trust the model and when to trust the floor. Detect when AI output has drifted from production reality. Assessment: Decision Logs tracking every human-AI interaction — a correct override of an AI recommendation is the most valuable developmental signal.
05
Co-Evolution
Education: Self-directed development. Assessment authority shifts to the individual. External evaluation becomes calibration, not control.
Manufacturing: The architect. Designs systems, not just operates them. Assessment: Evolution Ledger tracking capability frontier advancement — what can this person design now that they couldn't 12 months ago? Operator vs. architect trajectory over time.

China's manufacturing advantage is largely operator-class at scale.

From Cognitive Gym to National Power: Cognitive training builds developmental capacity, expressed through industrial manufacturing, enabling execution capability, infrastructure buildout, and ultimately national strategic strength. Feedback loops show how each layer compounds on the others.
Capability Formation Pipeline From Cognitive Gym to National Power: Cognitive training builds developmental capacity, expressed through industrial manufacturing, enabling execution capability, infrastructure buildout, and ultimately national strategic strength. Feedback loops show how each layer compounds on the others.
The architect-class advantage is winnable — but the frame has sharpened. China is training AI systems at scale through mass physical deployment: 80%+ of global humanoid robots, 470 robots per 10,000 workers. US manufacturing architects must build and govern AI systems competitive with systems trained on that deployment data. Embodied physical intuition — which aggregate data cannot yet replace — is the moat.

— The 15-Year Moat
09

Four Pilot Programs

Before the architecture scales, it must be tested against reality. Pilots serve three functions: test the architecture, generate proof-of-concept data, and reveal which scaling tradeoffs are theoretical versus actual blockers.

Phase 1 · Year 1–2
Assessment Infrastructure
AI-Enabled Developmental Assessment System
Build and validate the AI assessment infrastructure in partnership with 3–5 existing manufacturing training programs — community college advanced manufacturing, defense prime apprenticeship programs, vocational-technical schools. Deploy iteration stack tracking, orchestration quality monitoring, and developmental phase transition detection. This is the prerequisite that makes all subsequent pilots legible to capital — without it, you cannot demonstrate the developmental return that justifies patient investment. Funding: DARPA DSO + NSF Future of Work. Timeline: 18-month validated prototype. The critical design constraint: must not be experienced as surveillance. Developmental partnership, not performance monitoring.
Phase 2 · Year 2–4
Integrated Node
Single Integrated Manufacturing-Education Facility
One co-located facility operating simultaneously as an HMLV manufacturing cluster and a developmental manufacturing education environment. Real defense contracts. Real components — transformer laminations, GaN device packaging, precision machined defense components. Full five-layer developmental architecture. MDS intelligence layer from day one. 50–100 workers. Primary test: does developmental discipline survive production pressure? Candidate locations: Youngstown OH (America Makes), Cleveland OH (Cuyahoga CC + Cleveland-Cliffs proximity), Charlotte NC (Duke Energy demand pull). BMW Spartanburg dual-apprenticeship model as reference. Funding: DoD IBP + EDA Tech Hub. Decision gate at Phase 2 end: if developmental discipline failed under production pressure, refine architecture before propagating.
Phase 3 · Year 3–6
Distributed Network
Three-Node Distributed Metabolic Rate Network
Seed two additional facilities from Phase 2, using accumulated MDS knowledge. Site A: materials processing. Site B: component manufacturing. Site C: assembly and integration. Deliberately not co-located — that's the test. Can model-defined knowledge propagation substitute for geographic proximity while adding resilience? Primary measurement: time-to-competence for a novel challenge at Site B when the solution was developed at Site A vs. Site B developing independently. Antifragility stress test: deliberately disrupt one node, measure whether others absorb load using shared MDS. If second and third nodes reach competence faster than the first, the metabolic rate compounding claim is validated.
Phase 4 · Year 5–8
Ocean–Manufacturing Bridge
Cross-Domain Developmental Formation + Ocean Infrastructure
Manufacturing engineers in ocean infrastructure context — operating Ulysses-class AUV fleets monitoring undersea cable infrastructure. The ocean is the ultimate Layer 2 environment: no established procedures, constant constraint evolution, AI that cannot be trusted without physical intuition to calibrate it. Hypothesis: engineers who navigate real production failures in the ocean environment develop constraint navigation capacity that transfers to manufacturing contexts faster than domain-specific training alone. Dual-use value: developmental formation for the human pipeline and operational maritime domain awareness for national security. Acquisition pathway: SOCOM flexible authority for speed.
10

The Coalition Map

Phase 1 requires anchors across five categories simultaneously. A strong anchor in one category without corresponding anchors in the others creates an initiative that's well-funded but ungovernable, or well-governed but unfunded. Build the minimum viable coalition across all five before launch.

Federal Policy & Funding
  • DoD Office of Industrial Base Policy — IBAS program anchor
  • DARPA Defense Sciences Office — assessment infrastructure
  • EDA Tech Hubs program — regional node funding vehicle
  • NSF Future of Work — academic legitimacy + evidence base
  • DOE Manufacturing & Energy Supply Chains office
Defense Demand Pull
  • Army Materiel Command — organic industrial base surge urgency
  • NAVSEA — shipbuilding workforce crisis, acute pain point
  • Defense Logistics Agency — strategic materials offtake commitments
  • SOCOM — flexible acquisition authority for ocean pilot speed
Education Institutions
  • Cuyahoga Community College — Cleveland, OH (Cleveland-Cliffs proximity)
  • BMW Spartanburg apprenticeship — developmental culture model partner
  • Oak Ridge National Lab — Manufacturing Demonstration Facility
  • America Makes (Youngstown) — additive manufacturing MRC anchor
  • LIFT (Detroit) — lightweight metals, workforce mandate embedded
Technology & Capital
  • a16z American Dynamism — Katherine Boyle, primary VC anchor
  • Schmidt Futures — patient capital, national security focus
  • Palantir — MDS integration, DoD system connectivity pathway
  • Anthropic — developmental AI assessment research partnership
  • Breakthrough Energy Ventures — grid infrastructure node funding
Policy Validators
  • SCSP (Schmidt) — first engagement, validator that unlocks all others
  • Dan Wang / CFR — intellectual anchor for the cultural diagnosis
  • Sen. Warner + Sen. Young — bipartisan congressional champions
  • Atlantic Council GeoEconomics Center — allied dimension
  • ITIF — congressional audience credibility anchor
Industry Validators
  • Cleveland-Cliffs — sole US GOES producer, direct relevance signal
  • Toyota Manufacturing Kentucky — developmental culture benchmark
  • Ulysses / The Ocean Company — ocean pilot partnership anchor
  • Army Materiel Command organic base — DoD authenticity signal

Begin with SCSP. Their validator role makes every subsequent conversation easier. Then congressional champions. Then agency engagement. Then institutions and capital. Technology and capital partners follow the signal — when federal funding, defense demand pull, and institutional anchors are in place, the private capital conversation is fundamentally different. Build the coalition in sequence, not simultaneously.

— The Sequencing Imperative

A Capability Formation
Problem Has a Solution.

Created by
LaSalle Browne
Founder · AlphaSage · lbsage@alphasage.io

America has a capability formation problem. China is already building its closed loop — research, engineering, manufacturing, deployment, data, capital — across a dozen sectors simultaneously. The US must build a competing one. The fifteen-year developmental sequence, the physical infrastructure buildout, and the coalition must all start now. The window is open. It will not stay open indefinitely.

Engage Directly Read the Essay