I weak downvoted this comment because of the following sentence:
āI hope youāre right, otherwise you might be doing a lot of harm.ā
I agree that technically, a person whoās very successful at advocating for a low impact intervention might do harm. But I think we should assume that most contributions to public discourse about cost-effectiveness are beneficial even if they are false.
Accusing people of doing harm makes it more difficult to discuss the issue in a distanced and calm way. It also makes it more difficult to get people contribute to those discussions.
Epistemic adjectives such as āunfoundedā, āweakā, āfalseā, āmisleadingā are still available, and these are less likely to provoke guilt and stifle discussion. Unless there are egregious violations of honesty, I donāt think people should be accused of doing harm for defending their beliefs on cause-prioritisation.
Separately, I disagree voted because it seems wrong about the facts. HLIās analysis relies on many different studies with a sample size >30K, as Michael explains below.
I weak downvoted this comment because of the following sentence:
āI hope youāre right, otherwise you might be doing a lot of harm.ā
I agree that technically, a person whoās very successful at advocating for a low impact intervention might do harm. But I think we should assume that most contributions to public discourse about cost-effectiveness are beneficial even if they are false.
Accusing people of doing harm makes it more difficult to discuss the issue in a distanced and calm way. It also makes it more difficult to get people contribute to those discussions.
Epistemic adjectives such as āunfoundedā, āweakā, āfalseā, āmisleadingā are still available, and these are less likely to provoke guilt and stifle discussion. Unless there are egregious violations of honesty, I donāt think people should be accused of doing harm for defending their beliefs on cause-prioritisation.
Separately, I disagree voted because it seems wrong about the facts. HLIās analysis relies on many different studies with a sample size >30K, as Michael explains below.