Box CEO Aaron Levie and a16z on AI regulation, agents and security
a16z · x · 2026-09-27
a16z podcast with Box CEO Aaron Levie, Steven Sinofsky and Martin Casado covering "We Must Pace the Frontier," the 2028 "AI election," and Jevons paradox. Key points:
- Today's AI regulation debate happens before risks have been defined; past tech waves built safety standards only after learning how the technology actually failed.
- Agents never tire, run at enormous scale and probe systems like employees never could — implying a rethink of permissions, authentication and the security stack itself.
- AI innovation may increasingly happen outside frontier labs, in software built around models.
Related event: a16z Podcast: AI Regulation Shouldn't Precede Defining Risks(2 posts)→
More from AGI Musings
- Herbert Simon's reminder: attention scarcity is the opposite of 'more information is better' — charles_irl · 2026-09-27
- Web usability in the agent era: llms.txt plus 3D experiences vs 2D pages — noahsolomon · 2026-09-27
- Pedro Domingos: No tech accumulated more hacks faster than AI — AlexTensor · 2026-09-27
- Musk Guesses China Solves Compute and Chip-Making Constraints in 2-3 Years — haider1 · 2026-09-27
- Musk: China Will Solve Its Compute, Lithography and Chipmaking Constraints in 2-3 Years — haider1 · 2026-09-27
- Usability will shift from reducing entropy to maximizing engagement in the agent era — noahsolomon · 2026-09-27