I like Campaign Zero's data-driven prioritization of solutions, but it's not clear to me how they'd use marginal funds. I suspect this gap explains its absence from CEAs and Open Phil recommendation lists.
One element of this I'd like to understand better is the trade-off between donating your own compute (i.e. donating money you're spending on electricity) compared to donating money directly to a protein folding effort that utilizes cloud computing. In general, how do the hardware savings of using an existing device compare to the greater efficiency of cloud computing? Is the key innovation of Folding@Home that they're able to obtain in-kind electricity donations from people who wouldn't otherwise donate money to fund the effort?
Very interesting analysis! The study on passionate love and cognition might be more about early stage love right? If so, this would be a small cost of the intervention. As an observational data point suggesting positive economic impacts, married men earn up to 70% more than single men (https://www.marketwatch.com/story/married-men-earn-more-than-single-or-married-women-and-single-men-2018-09-19).
How did Facebook's $7 million match in 2019 compare with the probabilities modeled?
I see two reasons to adjust the benefit of this practice downward:
1) Some of Facebook's potential match would otherwise have nonzero benefit, and some (small) fraction would go to EA causes.
2) (More importantly) You can't donate stock through the Facebook Fundraiser platform. Donating stock (or index funds) allows you to not incur capital gains taxes, which for most people would be about 15% LTCG tax * 10% return = 1.5% per year.
If you expect your income to increase over time (all else equal this is reasonable), that's an extra reason to bunch. However, it throws a wrench in an alternative to front-load bunching, i.e. give 50% in year 1 and then only 10% of your raise since year 1 in the subsequent four years.
Would you still offset if society enacted a carbon fee-and-dividend, e.g. with the price equal to the social cost of carbon? Such a policy would also internalize the externality without compensating for the specific harm.
In part this may come down to whether you see climate change as a threat separate from other societal problems. I see it as a mechanism that takes its toll on broadly-comparable outcomes like DALYs and economic growth. From that perspective, the harms (and therefore the offsets) are comparable to the harms of, e.g., not providing bednets.
I like half of offsets: It's important to internalize externalities, and if the state has failed to do this with carbon pricing, it makes sense to self-assess a carbon tax to incentivize yourself to reduce emissions.
But once you've raised a certain amount from that self-assessed carbon tax, why limit to climate causes? Conversely, if Coalition for Rainforest Nations is the most cost-effective charity out there (8 metric tons of CO2 averted per dollar), then why only donate to it from your carbon tax bucket and not, say, your Giving What We Can bucket?
I admit I haven't done this yet, but if I were to approach the offsets issue, I'd self-assess a carbon tax and treat the proceeds as fungible contributions to my overall EA budget. Even with Coalition for Rainforest Nations' powerful effect in the climate sphere, it would probably not rank above other top EA charities, so I would not give to them.
Someone asked how Hillary Clinton would have fared. Might be a lot of work but she and Obama would make interesting benchmarks. https://twitter.com/JohnCarltonKing/status/1123325554378313728
Would be good to have Bill Weld in a subsequent revision.
Donate to Delaney here: https://go.johndelaney.com/page/content/this-is-about-america/
He'll also give $2 to a nonprofit for each donation he gets (none of the options are EA charities).