Marc Andreessen on AI regulation and open source
Marc Andreessen joins Aarthi and Sriram to argue against restrictive AI regulation and for broad, open-source development. He discusses large language models as augmented intelligence, challenges alignment and extinction-risk arguments, and considers AI’s use by China and other hostile actors. The conversation also covers chips, training data, education, AI companionship, and the religious roots he sees in AI doomerism.
Passages
3:17 “Yeah, it’s this concerted attempt in Washington right now by a coalition of what I call the Baptists and the bootleggers to basically lock down AI with a legal and regulatory regime that would basically entrench a small number of big players effectively forever under color of law and disenfranchise the next 30 years worth of startups and open source…” Share
12:17 “And you know, the core of it is basically this idea which is sort of the single most validated conclusion in all of the social sciences over the last century, which is that in human affairs, intelligence makes things better. And in fact, human intelligence basically makes everything better.” Share
Search what Andreessen said from Claude, ChatGPT or your own code. See plans