AGI (Artificial General Intelligence)

Artificial general intelligence refers to systems that can reason, learn, and act across domains with the flexibility associated with human cognition. In Unscarcity, AGI is treated as the catalytic technology that collapses labor costs, making the Foundation layer feasible and forcing a redesign of governance. The book frames AGI not as a distant speculation but as an imminent capability that will outperform humans in most tasks, triggering the Labor Cliff.

Economically, AGI accelerates automation to the point where wage labor ceases to be the binding constraint on production. Politically, it creates entities potentially worthy of personhood, requiring the Two-Tier Solution and Spark Threshold to adjudicate rights. Ethically, AGI heightens Goodhart risks: optimizing for narrow metrics can produce perverse outcomes unless bounded by the Foundational Principles of transparency and decaying power.

Historically, the debate echoes past general-purpose technologies–electricity, computing–but with the additional dimension of agency. Research communities distinguish AGI from narrow AI and speculate about recursive self-improvement; policy work from bodies like the OECD and NIST now considers AGI governance under AI safety and alignment agendas.

References

• UnscarcityBook, chapter1 and chapter6
• Nick Bostrom, “Superintelligence” (2014)
• Wikipedia: Artificial general intelligence (accessed 2024)