Interview
'Future bias' and a possible solution to moral fanaticism | Christian Tarsney (2021)
- Christian Tarsny anticipates that improved answers to global prioritization will directly influence individual resource allocation, noting that significant "low-hanging fruit" remains because few have explicitly analyzed these decisions.
- He expects the field of global priorities research to grow, with a strong pipeline of PhD students in philosophy and economics entering the area and the discipline eventually branching into political science, history, and psychology.
- Tarsny plans to focus on epistemic issues regarding predicting and influencing the far future, citing a current lack of empirical data on human ability to affect events on centennial or millennial scales.
- He hypothesizes that time perception biases, specifically the view of future experiences as "ahead" and past experiences as "wiped off," drive altruistic preferences and risk aversion, suggesting a higher burden of justification for prioritizing the past.
- Tarsny notes that debates on the ontology of time (presentism, growing block, eternalism) remain unsettled, and while "shrinking block theory" could explain future bias, it is not currently defended by philosophers.
- He predicts that time travel might weaken the asymmetry between past and future concern but expects the bias to persist, while noting that evidential decision theory can link past care to choices like Newcomb's problem without time travel.
- Tarsny argues that standard decision theory does not necessitate "fanaticism" because utility functions can be bounded or concave, and background uncertainty about universal value can create a "sphere of permissibility" where extreme maximization is not required.
- He expects that for small marginal differences, such as reducing extinction risk by one in a billion, stochastic dominance combined with background uncertainty permits varying approaches, including a "sure thing" strategy.
- Tarsny predicts long-termism will remain robust if there is even a one in a billion credence assigned to extreme persistence scenarios, though he warns that assuming a one percent annual existential risk could render risk reduction efforts trivial and represents unreasonable overconfidence.
- He suggests a natural connection exists between improving future generations' capacity to respond to challenges and improving current lives, allowing for "subtle long-termism" that does not demand extreme sacrifice.
- Tarsny expects most long-termist efforts, aside from existential risk reduction, will focus on equipping upcoming generations with better decision-making capacities through improved social trust and institutions.
- He predicts that the often-cited one trillion dollars for long-termist causes may be justifiable from a short-term perspective, as risks like nuclear and biological threats are already recognized priorities.
- Tarsny warns against using historical trend lines like declining violence to predict the far future if the expectation is a shift out of the current regime, and he notes a paradox that high intelligence does not inherently imply a tendency to pursue the good over the bad.
- He proposes a moderate externalism view where empirical and moral uncertainties matter for decision-making, but rejects extreme externalism as defective and extreme internalism for leading to an infinite regress of higher-order principles.
- Tarsny expects the most likely reason for changing philanthropic priorities in the next 10 years will be discovering new ways to usefully spend money beyond existential risk reduction, particularly by identifying leverage points in government where altruistic agents can gain traction.
- He expresses cautious optimism for short-term progress on epistemic questions like reasoning under unawareness and optimism for long-term philosophical progress regarding consciousness in simple AIs and insects.
- He anticipates the Global Priorities Institute will maintain a focus on coordination to prevent "nerd sniping," fostering actual research collaborations more common there than in typical philosophy departments.
- Tarsny notes that while competitive debate teaches critical thinking, success in the arena does not guarantee arguments are truth-seeking, and he considers Barry's Paradox a self-referential problem that challenges linguistic intuition.
- He observes that experimental results indicate people change judgments when they can affect the past, which his prior philosophical intuitions about the irrelevance of the past did not fully account for, and suggests "multiple explanations" exist for future bias.
- He argues that if a modal scenario assumes humanity survives but remains dysfunctional, the expected value of reducing extinction risk could be lower than anticipated due to the possibility of a mediocre future.
- Tarsny identifies routes for non-philosophy students and high schoolers to contribute through quantitative historical research or by building institutions that better serve the public good.