Databricks CEO Ali Ghodsi: Ontology graphs with millions of nodes give AI maximum org context
matei_zaharia · x · 2026-09-20
Databricks founder Ali Ghodsi explained the company's Ontology strategy in an a16z interview: digitizing the relationships between every goal, department, person, and project into a graph — with the largest customer deployment reaching millions of nodes.
- He analogized to org trees: information flows up and down the hierarchy, decisions escalate upward and percolate down; with an Ontology, much of that context-gathering and decision dissemination can now be done by AI.
- On existential risk, Ghodsi isn't losing sleep: frontier runs demand enormous power, GPUs and engineers with expensive failed attempts, so the self-improving takeoff loop's conditions aren't in place.
- What he watches most closely is cyber: most orgs aren't equipped for agentic capabilities, and the window between vulnerability disclosure and exploitation is shrinking fast.
More from AGI Musings
- HF researcher: 'Loss-of-control evals are sexy' while bias and social impact work gets ignored — evijit · 2026-09-20
- Peter Diamandis: 100,000 genius-level employees reporting to you is this decade's real offer — PeterDiamandis · 2026-09-20
- Demis, Jensen and AI leaders gathered by King Charles urge 'responsible optimism' on AI — Reza_Zadeh · 2026-09-20
- Are AI 'rogue agent' safety stories real capability demos or self-serving narratives? — North-Ad6031 · 2026-09-20
- User slams SSI's closed access to frontier model as dangerous, hopes it still goes public — iruletheworldmo · 2026-09-20
- Michael Levin: Intelligence doesn't need a brain — patterns themselves may be alive — burny_tech · 2026-09-20