Ben Horowitz: “The most dangerous thing in AI is Anthropic and Bill Gates”
Ben Horowitz calls Anthropic and Bill Gates the most dangerous voices in AI for amplifying theoretical risks politically
“we're actually talking about a risk that is purely theoretical at this point, and we're not getting anywhere near it”
Ben Horowitz on the All-In Podcast argues that Anthropic and Bill Gates are uniquely dangerous because their political influence amplifies AI existential risk fears without empirical basis. He invokes Databricks CEO Ali Godsi's four prerequisites for recursive self-improvement—smaller compute, faster training, higher accuracy, and repeatability—asserting none are being met. The segment represents a prominent VC counter-narrative to AI safety alarmism, with direct implications for AI governance and regulation debates.