I strongly agree. Even if bednets being used for fishing were a much more significant fiasco than it was, the fact that actions can have unintended consequences is a fact of life, not an argument against effective altruism. Presumably, many thoughts about what would happen if she did were part of Townsend’s motivation for writing her essay; e.g., the expectation that she would enjoy writing it and that it would promote ideas she thinks are worthwhile. But if it turned out that she unexpectedly did not enjoy writing it and that it met with little uptake, that need not indicate any foolishness on her part. Sometimes things don’t work out. What would be necessary here would be some argument that EA ~needs~ perfect certainty about the future in order for its practical and normative recommendations to make sense, and that this differs from the fallible future-thinking required by ordinary prudence. But no such argument is forthcoming.

I also found the consistent hostility to the use of numbers unfortunate. It is possible to be misled by numbers, yes. But it is also not possible to decide among a large number of highly complex and heterogenous competing priorities without using numbers.

The piece notes that QALYs/DALYs originate in health economics and health system priority-setting. It would benefit from considering why those fields require quantification, and why, despite their extremely well-known conceptual and methodological problems, QALYs and DALYs are still universally used. Answer: because there are over 85,000 ICD codes, aka medical diagnostic codes, which can be further discriminated by severity (not to mention social context!), and similarly many interventions, each with not only a range of costs and effects but also a range of evidence qualities supporting their use. And health systems must decide, using their limited budgets, what they will pay for. The idea that priority-setting at this scale of complexity could be done, without numbers, by some morally sensitive sage who eschews abstractions and grapples with REAL “muddy, hairy, smelly, messy people” instead of “figures in a ledger” is completely absurd. Any scheme designed in that way would almost certainly be dominated by a competing, quantitatively-informed approach. And of course, what’s true of health priority setting is just as true of philanthropic priority-setting more generally.

It is also worth noting, for those who (wrongly) do not care much about efficiency, that numbers aren’t just indispensable to efficient allocation. They are also necessary for just allocation. It is quantification that enables us to characterize inequalities such as, e.g., the much lower historical funding for Sickle Cell research as against its disease burden when compared to diseases which are not concentrated in Black populations. Without being able to abstract from specific individuals and their symptoms to a measure of generic health burden, that comparison would not be possible.

For what it’s worth, though I would quibble with some details, I found prof Townsend’s piece somewhat insightful. I certainly think trying to saddle her with the thesis that it’s bad to try to save lives is more uncharitable than any of her depictions of effective altruism. The difficulty I have with the Adelstein response is that it seems to be defending a kind of bare-logical/in-principle reading of effective altruism, whereas Townsend’s piece struck me as more focused on the attractions of the movement as a social phenomenon.

I’m surprised that her brief dig at malaria nets loomed larger for you than the association with the Sam Bankman Fried grift and the point about longtermism allowing the substitution of concerns with far-future AI conjectures to justify ignoring the concerns of people here and now. Neither of these are refutations of effective altruism as a set of ethical or philosophical theses, but they are clearly a part of how the movement has manifested itself in the world.