#467 — EA, AI, and the End of Work
Sam Harris and William MacAskill discuss the current state of effective altruism after recent setbacks, covering core EA cause areas like global health, animal welfare, pandemic preparedness, and AI safety. They explore tensions between quantifiable interventions and harder-to-measure but potentially high-impact problems like political dysfunction and social media's effects on democracy.
Summary
This conversation examines effective altruism's resilience and growth following major setbacks from cryptocurrency scandals and other controversies. MacAskill reports that despite public relations challenges, EA continues growing substantially, with funding for effective nonprofits increasing 50% to nearly $2 billion annually and membership in Giving What We Can growing 20-30%. The discussion covers EA's core cause areas: global health and development (where $5,000 can save a life versus $50,000 for one year of life in the US), animal welfare (particularly factory farming affecting 90 billion animals annually), pandemic preparedness, and AI safety. They address criticisms from tech leaders who dismiss international aid as ineffective, with Harris noting that dismantling USAID could cost 14 million lives according to Lancet projections. The conversation explores tensions around expanding EA's scope to include concern for less obvious beings like shrimp or digital minds, with Harris worrying this could alienate supporters while MacAskill defending intellectual exploration of moral boundaries. They discuss the historical precedent of moral 'weirdos' like early Quakers who championed abolition and women's rights. The dialogue shifts to considering positive goods versus just preventing suffering, and Harris raises concerns about EA's focus on quantifiable interventions while potentially missing high-impact but hard-to-measure problems like political dysfunction, social media's democratic effects, and the influence of podcasters on political outcomes.
Key Insights
- MacAskill argues that EA has shown remarkable resilience after major setbacks, with funding growing 50% to nearly $2 billion annually despite public controversies and cryptocurrency scandals
- Harris contends that the cost-effectiveness argument for global health interventions faces significant political resistance from tech leaders who view international aid as fundamentally ineffective
- MacAskill defends intellectual exploration of seemingly extreme moral positions (like concern for shrimp welfare) by comparing modern moral 'weirdos' to historical figures like Quakers who pioneered abolition
- Harris worries that EA's focus on quantifiable interventions may cause it to miss potentially higher-impact but harder-to-measure problems like political dysfunction and social media's effects on democratic cooperation
- MacAskill argues that AI progress has followed a remarkably stable exponential trend that has consistently surprised both machine learning experts and forecasters with its speed, particularly in reasoning tasks
Topics
Full transcript available for MurmurCast members
Sign Up to Access