Claim
As AI commoditises cognitive work, the scarce resources shift from raw intelligence to taste, judgment, relationships, and the ability to identify what is worth doing. These are the human-comparative-advantage areas that AI cannot easily replicate, and they are where the value of human contribution will concentrate over the coming decade. Career and company strategies should orient around building these scarce capabilities, not around competing in capabilities AI is rapidly absorbing.
Mechanism
Markets reprice resources continuously based on supply. When intelligence is scarce (the historical norm), people who could perform cognitive work were rewarded for that capability — analysts, engineers, copywriters, junior doctors. As AI absorbs commodity cognitive work, the supply of "intelligence" becomes effectively infinite at near-zero cost. The remaining scarce dimensions — what is worth doing (judgment), how it should look and feel (taste), who you can mobilise (relationships), and the ability to recognise the right problem (identification) — become the dominant value layer. These cannot be trained at scale; they are products of long human experience, social embedding, and aesthetic sensibility. Career investments in these dimensions will compound; investments in commodity-cognitive skills will deflate.
Conditions
Holds when:
- The category genuinely sees AI commoditisation of cognitive work (most knowledge work over the next decade).
- The scarce dimensions remain genuinely scarce — judgment, taste, relationships are not themselves AI-commoditised.
- The operator can credibly invest in the scarce dimensions (relationships and judgment require time and lived experience).
Fails when:
- AI develops taste / judgment / relationship capabilities comparable to humans (low-probability but not zero).
- Scarcity shifts to other resources first (energy, compute, regulatory access) before the cognitive layer fully commoditises.
- The operator over-invests in "scarce" dimensions while ignoring still-valuable commodity cognitive skills that haven't yet collapsed.
Evidence
"the scarce resources become taste, judgment, relationships, and the ability to identify what is worth doing"
— see raw/expert-content/experts/sam-altman.md line 17.
Signals
- Career investments shift toward judgment-rich roles (founder, executive, strategist) and away from commodity-cognitive roles vulnerable to AI substitution.
- Hiring rubrics weight taste, judgment, and network depth more heavily over time.
- Compensation patterns track this shift: roles that combine human-comparative-advantage with AI augmentation pay more than pure-cognitive roles.
Counter-evidence
The "scarce resources" thesis is a forecast, not a fact. There are scenarios where AI develops competing capabilities in taste (image generation, music generation already producing recognisable taste), judgment (long-context reasoning models), and even relationships (synthetic personas). Naval's If you can be replaced by training, you will be — specific knowledge is what survives commoditisation makes a similar but more conservative claim — specific knowledge is what survives, but the half-life is shrinking faster than the original framing assumed.
Cross-references
- The cost of intelligence is converging toward the cost of electricity — durable advantage isn't using AI, it's parlaying AI — the Altman thesis on what's commoditising.
- The singularity is one smooth curve — vertical looking forward, flat looking backward, never the disruptive shock people expect — the trajectory under which the commoditisation unfolds.
- If you can be replaced by training, you will be — specific knowledge is what survives commoditisation — Naval's adjacent claim with more conservative scope.