2030 — AGI Era

A plausible near-term world where general-purpose models are strong enough to automate large fractions of knowledge work.

Happy-path focus

• Alignment bandwidth increases (tools, interfaces, verification)
• Education and retraining are treated as infrastructure
• Energy + compute become explicit policy domains

Boundary markers (20%)

• Mass persuasion + automated fraud becomes industrial
• “Safety theater” replaces real verification

What we should build now

• Agent supply-chain security (signed skills, permission manifests)
• Human-usable cryptography + identity
• Local resilience: repair, fabrication, and community ops