Linkpost for July
Effective Altruism
Global Poverty
Elon Musk is likely responsible for the deaths of millions of people through cutting USAID [The Argument].
Animal Advocacy
Adult wild terrestrial vertebrates from large species seem to be most likely to die from human action, such as hunting or vehicle collisions [Wild Animal Initiative]. Juvenile terrestrial vertebrates and terrestrial vertebrates from small species seem to be most likely to die of predation. Insects seem to be most likely to die of parasitism and predation. Fish are too understudied to say anything one way or the other. These findings are relevant to whether wild animals have lives worth living-- often, people who believe that wild animal lives are net-negative assume that wild animals die in slow and painful ways, but the evidence for that claim is slight. Predation seems to me to be a quick death, at least sometimes.
Wild animal lives aren’t necessarily net-negative [Biology & Philosophy]. We don’t know how much suffering an average dying animal experiences, how bad various negative experiences are, or how good and common various positive experiences are. In species that produce a large number of offspring the vast majority of which die young, many of the young may be nonsentient, and if the surviving offspring live long enough their happiness might outweigh the suffering of the animals which die young. In conclusion, we can’t know whether there is net suffering in nature.
We are often clueless about the most important effects of an intervention-- for example, whether a particular animal welfare intervention makes the animal advocacy movement more unpopular and thus decreases the likelihood of abolishing animal agriculture entirely. Principles for figuring out what interventions are good when we’re clueless [Effective Altruism Forum]: accounts for potential, sufficiently certain negative side effects; doesn’t rely entirely on made-up numbers, tiny probabilities, or which particular reference class you choose; and doesn’t have a lot of unknown unknowns, because the intervention is testable, you can stop doing it if it turns out to be bad, and isn’t dependent on a long trend of causal reasoning.
Existential Risk and AI
In the past, scams had to be cheap, because most people don’t have that much money [In Pursuit of Laziness]. Large language models make it much cheaper to run scams, which means that we should expect to see more elaborate scams targeting ordinary people in the future. Our heuristics about how to avoid being scammed will need to update.
People often call for AI to be democratically governed. But we should distinguish between pluralist democracy (compromise between people with different values and lifestyles so they all have control over their lives with limited ability to control others) and homogenizing democracy (the faith that the majority knows what the right values and lifestyles are and should be left free to enforce them on others) [Andy Masley]. Often, calls for AI to be more democratic end up erasing large minority groups such as “people who actually want to use AI.” Also, a lot of people are wrong or have bad values, so we probably don’t want to set AI values by committee.
American Democracy
Tulsi Gabbard seems to have spent much of her political career getting advice from her guru, the leader of a breakaway Hare Krishna group [Washington Post].
A deep dive into whether Trump saved undocumented immigrant children from sex trafficking (no) [The Watch]. In fact, Trump’s policies probably increased the rate of trafficking-- for example, he slashed funding for the government’s primary anti-trafficking office and made it harder for victims of trafficking who report the crime to get a visa. Many so-called “lost” children were found simply by checking the address they were supposed to be and discovering they were there.
Particularly Good: the post-virtue era of politics in which, instead of saying “my opponent did something morally wrong, she shouldn’t do that,” we say “my opponent did something morally wrong, which justifies anything morally wrong that I do” [Derek Thompson].
