Databricks CEO Ali Ghodsi: stop scaring people about AI — context, not capability, blocks enterprise adoption
a16z · youtube · 2026-09-19
Databricks co-founder and CEO Ali Ghodsi joins a16z GPs Martin Casado and Sarah Wang to discuss AI risk and enterprise adoption.
Key points:
- Today's models are already capable of automating far more work than companies use them for; the real bottleneck is context—models lack institutional knowledge, meeting history, and understanding of how decisions get made
- He advocates building an organizational "ontology" to close this gap, sharing Databricks' internal learnings
- Distinguishes speculative superintelligence risk from the immediate challenge of AI-powered cyberattacks
- Discusses the "doom talk vs IPO allocations" dynamic, industry self-policing vs federal involvement, and whether recursive self-improvement is actually happening
- Enterprise-side: managing exploding AI usage and costs, the shift to multiple models and harnesses, and how agents are reshaping infrastructure itself
- Includes Fortune 500 use cases that surprised even Ghodsi
Related event: Databricks CEO: Existential AI Risk Near Zero, Real Threat Is Cyberattacks(5 posts)→
More from AGI Musings
- Polymarket: majority of online articles published today are 'AI slop' for the first time — Polymarket · 2026-09-19
- Pedro Domingos: Opposing Data Centers Means Keeping Your Country Stupid — pmddomingos · 2026-09-19
- Tyler Cowen: AI Is a Top-10 Event in Human History — TheZvi · 2026-09-19
- Sentdex: Perception, Not Decision-Making, Is the Hard Part of Self-Driving — Sentdex · 2026-09-19
- Turing Post explains what actually makes recursive self-improvement recursive, and where current attempts fall short — TheTuringPost · 2026-09-19
- Gary Marcus pushes back on AI doom: humans decide what AI can do, not the technology — GaryMarcus · 2026-09-19