a16z crypto show AI Just Gave You Superpowers — Now What?
71 snips
Mar 5, 2026 
Christian Catalini, economist and MIT Crypto Economics Lab founder, explores AGI economics and verification. Eddy Lazzarin, CTO-level technologist, discusses practical agent deployment and tooling. They discuss agents as coworkers, the automation versus verification split, the AI sandwich firm structure, crypto for identity and provenance, and the contrast between a hollow and an augmented economy.
AI Snips
Chapters
Transcript
Episode notes
Shift From Coding To Verification
- Do focus on verification work rather than line-by-line coding as AI automates routine engineering tasks.
- Eddy Lazzarin says the balance shifts toward ensuring code meets business goals and customer value, not just printing lines of code.
Verification Is Human-Recorded Measurement
- Verification is the act of applying an individual's embedded measurement and experience that isn't yet digitized and therefore hard to automate.
- Christian Catalini explains machines automate what's measured, but human seasoning from out-of-distribution experiences remains costly to capture.
The AI Sandwich Organizational Model
- The AI Sandwich structure places a small director at the top, a swarm of agents in the middle, and top verifiers at the bottom to steer and validate outputs.
- Christian Catalini argues this lets one human scale oversight while relying on many agents and a thin set of expert verifiers.