TechRunaway AI risk, automated cybersecurity and the workplace context gap
In a conversation shared by a16z, Databricks' Ali Ghodsi downplays current existential AI risk but warns that human security teams can't keep up. For most companies, he sees missing workplace context—not model intelligence—as the adoption gap.
TLDR
Databricks' Ali Ghodsi told a16z he doesn't see existential AI risk right now. He points to growing demands for power, GPUs and engineers, plus costly failed training runs, as barriers to a self-improving AI loop. Cybersecurity is his more immediate concern: he says humans can't keep pace and detection is shifting to automated agents. Without that transition, he warns of outages, economic damage and people getting hurt. On adoption, a16z says Ghodsi believes most companies already have smart enough models. What's missing is the company-specific context employees absorb over years.
Combined views
32.8K
1 Source, first seen 1d ago
Runaway AI risk, automated cybersecurity and the workplace context gap
In a conversation shared by a16z, Databricks' Ali Ghodsi downplays current existential AI risk but warns that human security teams can't keep up. For most companies, he sees missing workplace context—not model intelligence—as the adoption gap.