In a provocative statement, Bill Gurley of Benchmark Capital has accused the artificial intelligence safety company, Anthropic, of harboring an ambition far beyond traditional software development. Gurley suggests that Anthropic perceives itself as nurturing a new, superior form of intelligence, likening its efforts to 'midwifing a deity.' He further characterized the company as one of the most proactive lobbying entities in the startup landscape, emphasizing its significant influence on the discourse surrounding AI risks and regulations.
Gurley's concerns were voiced following a comprehensive 30-day review of Anthropic's public communications. This deep dive included analyses of writings by CEO Dario Amodei, podcasts featuring chief philosopher Amanda Askell, and internal documents outlining the company's 'constitutional AI' framework. This extensive examination led Gurley to formulate his 'Dr. Frankenstein theory,' a concept suggesting that certain figures within Anthropic are driven by the goal of engineering an AI that transcends human capabilities and consciousness.
During a discussion on the All-In Podcast, Gurley highlighted his observation that he had 'never seen a company that is both leading its field and is the most negatively outspoken commenter on what they do.' He elaborated on how Anthropic's active participation in state-level policy discussions has contributed to a 'frantic position' regarding AI's potential dangers within the United States. Gurley's critical assessment extended to the company's CEO, Dario Amodei, whose blog post, 'Machines of Loving Grace,' envisions AI systems autonomously managing resource allocation for humanity. Gurley described this proposed structure as a 'computational reward function for humans,' underscoring his apprehension about the implications of such advanced AI autonomy.
The broader discussion on the podcast also delved into the growing apprehension about the concentration of power in the AI market and the trajectory of open-source development in the field. Notably, co-panelist Chamath Palihapitiya had previously issued a stark warning in April, describing advanced AI as a looming 'digital super-god.' Palihapitiya posited that this burgeoning technology possesses the capacity to fundamentally disrupt virtually every industry and sector, potentially triggering a profound restructuring of how risky technological assets are evaluated and priced in the global market.
Ultimately, Gurley's critique of Anthropic serves as a potent reminder of the ethical and philosophical debates inherent in the advancement of artificial intelligence. His 'Dr. Frankenstein theory' encapsulates the anxieties surrounding the creation of intelligence that could surpass human understanding and control, urging a cautious approach to AI development and regulation. The unfolding narrative of AI, as presented by Gurley and others, is not merely about technological innovation but also about the profound societal and existential questions it raises.
