Economics & Scarcity

What will be scarce in an AI-abundant world? Sources: Alex Imas's Substack essay on post-commodity economy, Gary's Economics on compounding and debt. See [[financial-markets]] for complementary material.

---

The Post-Commodity Economy (Alex Imas)

Source: Clippings/What will be scarce?.md

Opening example: Starbucks reversed its automation rollout — barista "craft" is the product, not the coffee.

The Commodity Form (Marx)

Products become commodities when they are:

• Standardized
• Interchangeable
• Priced only by inputs + margin

AI accelerates commoditization of knowledge work — writing, coding, analysis become nearly free at the margin.

Why This Changes Everything

Autor/Thompson framework: Two categories of tasks:

• Expert tasks (require judgment, context, creativity) — historically hard to automate
• Inexpert tasks (routine, rule-based) — historically automated first

AI breaks this: it automates expert tasks (legal analysis, medical diagnosis, financial modeling) while inexpert relational tasks remain human (hospitality, care, craft).

The Structural Change Theory (Comin, Lashkari, Mestieri)

Why do sectors shift over time?

The traditional answer: productivity differences. But empirical work finds only ~25% of sectoral shifts are explained by productivity / price effects.

75% is explained by income effects.

As income rises, people don't just consume more of the same things — they consume different things. They shift toward:

• Restaurants (not just food)
• Experiences (not just transport)
• Care relationships (not just health outcomes)
• Art and craft (not just functional objects)

Mimetic Preferences (the missing piece)

Standard economics assumes preferences are stable and independent. They're not.

Mimetic preferences (Girard meets economics):

• Comparative: value derived from social position, not absolute utility
• Never satiated: once you achieve the comparison point, a new reference emerges
• Income-elastic: as income rises, more spending shifts to mimetic goods

This explains luxury markets, status goods, and why the relational sector resists commoditization.

A Hermès bag loses value if everyone has one — its scarcity IS its product. A skilled therapist's relationship cannot be averaged away. A remarkable chef's dish cannot be mass-produced without destroying the proposition.

The Relational Sector

Jobs that structurally resist AI substitution:

• Teachers (not EdTech — actual mentor relationships)
• Therapists and counselors
• Nurses and care workers
• Hospitality and service excellence
• Craft producers (watchmakers, bespoke tailors, artisan food)
• Artists and performers (presence matters)

Key finding: AI involvement undermines the exclusivity premium.

Experiment data (cited in essay):

• 44% premium for "human-made" vs 21% premium for "AI-assisted"
• Knowing AI was involved in a luxury good reduced willingness to pay

Baumol's Cost Disease as Feature

Baumol's cost disease: Service sector productivity doesn't grow as fast as manufacturing, but wages must keep pace with the economy. Services therefore get relatively more expensive over time.

Traditional economics treats this as a problem (healthcare, education unaffordable).

Imas's reframe: In a world of commodity abundance, cost disease is the feature. The things that can't be efficiently produced are exactly what becomes scarce and valuable.

Summary: Post-commodity scarcity = human presence, authentic relationships, irreproducible craft, social status through genuine differentiation.

---

Connections to Other Work

Mimetic preferences ↔ [[philosophy#René Girard]]: Imas essentially operationalizes Girard's mimetic theory economically. Comparative desire is never satisfied because the model's desire constantly shifts the target.

Relational sector ↔ [[watches]]: Independent watchmakers (Hajime Asaoka, Rexhep Rexhepi, Laurent Ferrier) represent exactly this — their work is valuable because it is irreproducible, relational, and anti-commodity. Scale would destroy the proposition.

Mimetic marketing ↔ [[philosophy#René Girard]]: Luxury advertising sells access to the model's desire, not the product's utility. Girard explains why this works structurally.

AI automation of expert work ↔ [[cybersecurity-thesis]]: Hakyun's thesis is the sector-specific version of Imas's general argument — cybersecurity's complexity increases with AI because the attack surface grows faster than the automation of defense.

---

Summary Table

Concept	Key Insight
Commodity form	AI makes knowledge work free at the margin
Structural change	75% of sectoral shifts = income effects, not productivity
Mimetic preferences	Comparative, never satiated — income-elastic demand for status
Relational sector	Human presence, care, craft resist AI substitution
Baumol's cost disease	What can't be efficiently produced becomes the scarce luxury
AI involvement effect	Undermines exclusivity premium (44% → 21% WTP)

---

Related Pages

[[financial-markets]] | [[philosophy]] | [[watches]] | [[cybersecurity-thesis]] | [[singapore-finance]]