#36 William MacAskill: The Science of Doing Good

Episode Summary

Executive Summary: William MacAskill explains effective altruism as a rigorous, evidence-based approach to doing good that prioritizes the highest-impact uses of money, time, and careers. He argues that charity is often inefficient, that future generations matter morally, and that institutions should be designed to reward measurable impact rather than intention or overhead ratios.

Main Topics: Effective altruism as evidence-based philanthropy (Priority: 5/5): MacAskill frames effective altruism as importing scientific rigor into charity by comparing interventions on expected impact rather than relying on sentiment, intentions, or overhead ratios. The moral case for giving more strategically (Priority: 5/5): He uses Peter Singer’s drowning child analogy to argue that wealthy people have strong moral reasons to donate substantial portions of their income to save lives and reduce suffering. Career choice and earning to give (Priority: 4/5): MacAskill argues that many people can do more good by pursuing high-impact careers in business, finance, policy, research, or nonprofits, with personal fit being crucial. Limits of charity evaluators and nonprofit incentives (Priority: 5/5): He criticizes simplistic metrics like overhead ratios and explains how nonprofits lack the market feedback that weeds out inefficient for-profit firms, making independent evaluation essential. Longtermism, future generations, and existential risk (Priority: 5/5): A major focus is on the moral importance of future people and the need to reduce risks from AI, biotechnology, and global catastrophe that could cut short human civilization. Public policy, taxes, and government’s role (Priority: 4/5): He distinguishes taxes from charity, arguing governments provide public goods that make income possible, while philanthropy should fill gaps where markets and democracy fail. Practical wisdom, uncertainty, and honesty (Priority: 4/5): He emphasizes calibration, humility, and avoiding overconfidence in moral and practical judgments, while preferring honesty without endorsing unfiltered 'radical honesty.'

Key Arguments: Good intentions are not enough; philanthropy should be judged by measurable impact and evidence. The difference between mediocre and best charities can be enormous, sometimes hundreds of thousands of times. People in rich countries can often do far more good by donating a large share of income than by simply working in nonprofits. Personal fit matters: choose work where you can become highly competent, not just what you feel passionate about now. Markets usually allocate resources better than philanthropy; charity is most needed where markets and governments fail. Future generations have moral value even though they cannot vote or participate in markets, so long-term risks deserve attention. Overconfidence is a major cognitive error; people should calibrate probabilities more carefully and seek disconfirming views. Honesty is important, but radical honesty can ignore the symbolic and social meaning of language and harm relationships.

Data Points: Suggested giving pledge: 10% of income - Giving What We Can’s recommended donation commitment MacAskill’s annual donation share: about 25% of income this year - He donates everything above his personal spending baseline once per year Lifetime donation estimate: about half of lifetime earnings (original estimate), now likely more - His early calculation of what he would give away after setting a living baseline Number of hours typically worked in a career: 80,000 hours - Name of the career advice organization he co-founded Income gap cited: around 100x - Typical rich-country income versus the poorest 700 million people in the world Share of income explained by birth and parental status: 80% - He cites this as evidence that global income is largely luck-driven Harvard/Stanford endowment example: $35 billion - Used to argue against perpetual foundations spending down too slowly Life expectancy increase from global health aid: more than 50% - He cites this as evidence that global health aid has been highly effective Scared Straight harm estimate: $230 of social and penitentiary costs per $1 spent - Example of an intervention that was not just ineffective but harmful Typical rich-country income compared with poorest global population: 100x - Illustrates the extreme scale of global inequality Risk estimate style example: 1 in a million happens about 10% of the time; 1 in a thousand happens about 30% of the time - He cites this to show people are poorly calibrated about probabilities

Pivotal Quotes: "The approach we take is called effective altruism. It's, as you say, about trying to make the attempt to do good more scientifically rigorous, more evidence-based, more based on argument." — William MacAskill: Defining effective altruism and its contrast with emotional giving "The meaning of life is to contribute to the flourishing of humanity in the long run." — William MacAskill: His closing view on purpose, success, and moral direction "There's just a huge amount of spending goes on, and we don't even know whether that spending... is actually having more of an impact than simply giving poor people cash." — William MacAskill: Critique of government and philanthropic spending without rigorous evaluation

Implications: The conversation pushes listeners to rethink charity, careers, and policy around measurable impact, long-term risk, and moral impartiality. It suggests philanthropy should be more evidence-driven, and that avoiding catastrophic future harms may be one of society’s most important tasks.