Zach Freitas-Groff

In Human Compatible , Stuart Russell makes an argument that I have heard him make repeatedly (I believe on the 80,000 Hours podcast and the Future of Life Institute conversation with Steven Pinker). He suggests a pretty bold and surprising claim: [C]onsider how content-selection algorithms function on social media... Typically, such algorithms are designed to maximize click-through , that is, the probability that the user clicks on presented items. The solution is simply to present items that the user likes to click on, right? Wrong. The solution is to change the user's preferences so that they become more predictable. A more predictable user can be fed items that they are likely to click on, thereby generating more revenue. People with more extreme political views tend to be more predictable in which items they will click on... Like any rational entity, the algorithm learns how to modify the state of its environment—in this case, the user's mind—in order to maximize its own r

  Learn something new every day. Today's piece of knowledge is this for me (from wikipedia): Eyestalk ablation is the removal of one (unilateral) or both (bilateral) eyestalks from a crustacean. It is routinely practiced on female shrimps (or prawns) in almost every marine shrimp maturation or reproduction facility in the world, both research and commercial. The aim of ablation under these circumstances is to stimulate the female shrimp to develop mature ovaries and spawn. Most captive conditions for shrimp cause inhibitions in females that prevent them from developing mature ovaries. Even in conditions where a given species will develop ovaries and spawn in captivity, use of eyestalk ablation increases total egg production and increases the percentage of females in a given population that will participate in reproduction. I read briefly about this in an update from the Aquatic Life Institute , which is engaged in the critical and radically neglected project of improving the lives

I took a much-needed break from research in December and spent much of that time seeing classic movies I've always wanted to see as well as some new ones at the nearby drive-in. My reviews: New Releases: Promising Young Woman was amazing, and shocking. I recommend it far and wide to anyone with emotional fortitude.  The screenplay is a masterpiece.  The movie concerns a woman out to avenge a wrong. Played by Carey Mulligan, she's a brilliant and tough protagonist. It's a performance that could easily have been monotonous, but Mulligan has us guessing her every move. She's the star, but there's not a weak performance in the movie. The bright color pallet pops off the screen, and the sumptuous set decoration establishes each scene perfectly. The snappy pop soundtrack seals the deal. But please, please, make sure you are ready for something disturbing before you see it. News of the World , on the other hand, was as boring as could be. It's like watching someone el

Might there be more unity? There are some more fundamental reasons than Biden or Trump why there could be more unity in the U.S. in the next few years: COVID-19 and the Capitol insurrection. Anecdotally, we all have stories of people coming together in times of crisis: families shifting roles to make ends meet in wartime or even just the surprisingly rapid change in social habits that happened last March as people in many parts of the country isolated nearly overnight. Some research backs up the anecdotes. Here I'll lump together work on both inequality and polarization since both involve a sort of unity in the population. First, a body of surveys and experiments finds that cooperation within local communities increases cooperation during wartime. Relatedly, the relationship between the rate of return to wealth and growth, which Piketty controversially tied to inequality, tends to flip during war time . 1   More recently, while we can all think of the COVID-19 crisis as being shar

It's been a hell of a week since the despicable white supremacist riot and attempted insurrection at the Capitol. The event should have been unsurprising (not that this specific thing would happen, but that something like it would happen). It's not even clear to me that it's the worst we've seen the past few years (unclear if it's worse than concentration camps for immigrant children or banning people from the U.S. based on their religion, probably somewhat worse than declaring a fake emergency). Nevertheless, here we are. I unsurprisingly agree with the many Democrats, Never - Trumpers , and Until-Yesterday-Trumpers that we need consequences. BUT all consequences are not equal, and I worry there are going to be some serious mistakes with consequences. With that in mind—here's a rough, informal list of things that I think are good, bad, or iffy. Some are sure to be off, so please tell me why I'm wrong. Impeachment-ish things Impeachment : iffy/good (good if

Does democracy help people who don't get to vote? Democracy has a strong track record of peace and prosperity, but it's not obvious that it would help those who, as a class, do not participate in the political process, such as future generations and nonhuman animals. Two nice economics experiments suggest that democracy may help those who don't vote. (Specifically, elections help relative to everyone deciding for themselves.) The first is Hauser et al. (2014), "Cooperating with the Future" (I'm working with the first author on a new project): What mechanisms can maintain cooperation with the future? To answer this question, we devise a new experimental paradigm, the ‘Intergenerational Goods Game’. A line-up of successive groups (generations) can each either extract a resource to exhaustion or leave something for the next group. Exhausting the resource maximizes the payoff for the present generation, but leaves all future generations empty-handed. Here we show