Databricks CEO Ali Ghodsi: runaway AI takeoff unlikely, but agentic cyber risk is the real threat
matei_zaharia · x · 2026-09-19
Databricks CEO Ali Ghodsi told a16z he isn't losing sleep over existential AI risk: a runaway takeoff would require several conditions to hold simultaneously, and opposite conditions prevail—frontier training runs demand ever more power, GPUs and engineers, with costly failures. The self-improving loop isn't imminent.
What he watches closely is cyber. Data/AI and security used to be separate worlds, but internal agents now interact with external agents, generating logs and fingerprints at scales orders of magnitude beyond a year or two ago. The window between a CVE being published and weaponized has collapsed since 2018-19, and most organizations aren't ready for agentic capabilities.
More from AGI Musings
- Investor: in the LLM era 'did it work' replaces diplomas and test scores — StewartalsopIII · 2026-11-04
- Should AGI labs have a reverse Conway's Law — design the org from the artifact? — akbirthko · 2026-11-04
- Verification Is Not Knowledge: Science needs more than AI answers — ersinesen · 2026-11-04
- Andreessen: Agents Are Entering Enterprises as Chaos, Just Like PCs and Spreadsheets Did — Dan_Jeffries1 · 2026-11-04
- 1000+ mathematicians sign pledge to shun AI companies and models — Shesawthat · 2026-11-04
- Von Neumann meme: "You don't have to be responsible for the world that you're in" — akbirthko · 2026-11-04