Question
What happens when intelligence becomes abundant?
If cognitive work becomes cheap and plentiful, which bottlenecks remain—and which civilizational assumptions break.
Why it matters
Abundant intelligence could unlock science and coordination—or intensify deception, concentration of power, and meaning crises. Pretending we already know which is wishful.
Current understanding
Intelligence is becoming cheaper at the margin for many tasks, but embodiment, verification, trust, energy, and political control remain binding in different places. 'Abundance' is uneven and contested.
Known evidence
- Model capability on many cognitive benchmarks has risen rapidly this decade.
- Deployment lags show that institutional absorption, liability, and reliability constrain economic impact.
- Attention, judgment, and authenticity already face new spoofing costs online.
Competing explanations
- Explosive takeoff: abundant AI collapses most knowledge bottlenecks quickly.
- Implementation slog: organizations, regulation, and physical world absorb gains slowly.
- Power concentration: abundance exists technically but access remains oligopolistic.
- Verification bottleneck: generated answers outpace society's ability to check them.
Figuora hypothesis
Hypothesis: if intelligence becomes abundant, scarcity migrates to verification, trust, energy, embodiment, and collective choice—unless institutions reinvent those layers at the same speed.
Labeled as a hypothesis—not established conclusion. It should be pressure-tested against counterarguments and unknowns below.
Counterarguments
- AI may also automate verification, collapsing the bottleneck it creates.
- 'Abundance' may be marketing for partial automation of mid-skill cognitive tasks.
Unknowns
- Whether synthetic media forces a lasting trust reset or a messy equilibrium.
- How much physical-world friction remains after strong digital intelligence.
Sources
- 1. Situational Awareness — Leopold Aschenbrenner (2024) · Aggressive AGI timing claims; treat as scenario analysis.
- 2. The Second Machine Age — Erik Brynjolfsson and Andrew McAfee (2014) · Digitization and bounty/spread framing.
- 3. Average Is Over — Tyler Cowen (2013) · Complementarity to machines as labor sorting story.