Making Sense with Sam Harris · Sam Harris

#467 — EA, AI, and the End of Work

March 30, 2026·30 min·2 clips

Sam Harris cites Lancet study: dismantling USAID could kill 14 million people, including 4.5 million kids under five

Will McCaskill returns. Harris treats him as a familiar guide to ethical questions, especially the ones that gather under effective altruism. The occasion is simple enough: Doing Good Better has a tenth anniversary edition, with revised statistics and a new foreword that answers objections and looks back at the movement's growth. Then things get stranger. Harris worries about the point where moral attention drifts so far from ordinary sympathy that people can barely follow it, even if the question still matters. Shrimp are the first test. Digital minds are harder, because possible machine suffering asks listeners to picture moral status without the usual emotional signals. Harris stays careful. He is not rejecting the argument upfront. He is asking what happens when an idea might be true, but almost no one can feel it. McCaskill gives ground on part of it. He says he is not personally sold on the shrimp case, but still wants to defend the habit of taking odd ethical possibilities seriously. Public reaction matters, just not for everyone in the same way. Some people need to think about coalition building and persuasion. Others need to ask what moral reality might be hiding. His historical example is abolition. Early Quakers looked bizarre in their own time, and their views later fed into the British abolitionist movement. Weirdness does not prove anything, but ridicule alone does not settle the question either. Harris then turns to triage. How should charitable resources be used when the worst harms may not look like the measurable interventions on GiveWell's website? The problem becomes political. If American values, American politics, and social media are weakening the ability to cooperate, cost effectiveness math may miss the conditions that decide whether any problem can be solved. The excerpt stops behind the subscriber feed, with the hard part still hanging there: moral reasoning has to look at strange possibilities, but it also has to survive human attention, political breakdown, and the limits of compassion.

As heard by us

A sharp EA conversation about weird moral frontiers, AI anxiety, and the risk of dismissing tomorrow's obvious ethics.

Sam Harris brings Will McCaskill back to Making Sense for a conversation about effective altruism after a turbulent decade, framed by the 10th anniversary edition of Doing Good Better.

Read the full review in PlayNext →

Why you'd press play

You want the conversation that moves from effective altruism after FTX to the edge case of whether future digital minds could matter morally.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Making Sense with Sam Harris