This week's links (4)

Collections of papers and articles that I’ve spotted the last two weeks this week that seem interesting. Comments on some of them.

Economics

• Are the rich more selfish than the poor, or do they just have more money? A natural field experiment.
  • Turns out that controlling for the marginal value of money, they behave equally
  • That is, people respond similarly to similar incentives here after that control.
  • Other papers about this:
    • More likely to make a charitable donation and contribute a larger percentage of money to charity, volunteer, help, and trust (Note the large sample size)
    • More unethical decision making according to other studies, but they test different things and their sample size is smaller.
  • The political implications of american concerns about economic inequality
    • Concern with inequality does not lead to anti-inequality policy?
• A composite perspective on British living standards during the Industrial Revolution
  • There are two perspectives, both agree that the IR improved the standards of living, but disagree about how fast they did. One camp claims that wages initially stagnated, the other camp claims they rose straight from the beginning.
  • The paper claims that if a broader welfare index is considered, including health, increases in working time, and inequality, standards of living improved from the beginning.

Physics

• No, we probably don't live in a computer simulation
  • Sabine 'Bee' Hossenfelder against Bostrom's popular idea, arguing that physics makes it really hard to do such simulations
  • It is my contention that regardless of how solid the argument is, it is as popular because it sounds contrarian and edgy.
• NASA proposes a magnetic shield to protect Mars' atmosphere

Psychology

• Conducting the Milgran experiment in Poland, psychologists show people still obey
• Global increases in individualism
  • The world is getting more individualist on average. Some countries have regressed, others have increased slowly.
  • They tested for some explanations (socioeconomic conditions, climate, pathogen prevalence, disaster frequency and year). Improvement of socioeconomic conditions is by far the best predictor, the effect pathogen prevalence is fairly small.
• How emotions are made: The secret life of the brain
  • Apparently, among other things, emotions do not have a specific place in the brain, e.g. fear does not need the amygdala.
• Sociodemographic correlates of sexlessness among american adults and associations with self-reported happiness levels.
  • Lack of sex does not correlate with unhapiness
• The future of secularism
  • Religious people have more kids
  • Having more kids and being religious is heritable. Hence, natural selection will push the world towards more religion, counteracting the effect of intelligence and science on religion.
  • We'll see.
• A behavioral economic assessment of individualizing versus binding moral foundations
  • Conservatives as measured by  a Haidtian moral framework cooperate less.
• Deliberation erodes cooperative behaviour, even towards competitive outgroups, even when using a control condition, even when controlling for sample bias.
  • That is, thinking more makes you more selfish.
  • How much more? About 20% more or so.
  • Don't worry, you won't turn into Ayn Rand by overthinking.
• Does everyone have a price? On the role of payoff magnitude for ethical decisionmaking.
  • Incentives affect different people differently

Philosophy

• Consistent vegetarianism and the suffering of wild animals
  • Wild Animal Suffering goes mainstream, and so does Brian Tomasik, being cited in the Journal of Practical Ethics.
  • WAS suffers from the issue that it is a cause that has received little attention, and most of the references of WAS can be traced back to Tomasik himself.
  • I think, on the other hand, on net it is better for life broadly to exist rather than not. Some arguments here.
• How to spot a fake science news story
  • Among other things, if it is about evolutionary psychology, heh.
• What is a good sample size?
  • My heuristic is N>100 to take something seriously
  • And N>250 seems even better

Artificial Intelligence

• Stopping GAN Violence: Generative Unadversarial Networks.
  • Best Machine Learning paper of the year.
• Feudal networks for hierarchical reinforcement learning
  • DeepMind goes medieval
  • Nice to see seaborn charts in a paper
  • This is a neural net (Manager) with net-serfs (Workers) working on subgoals
• Large-scale evolution of image classifiers
  • Evolved neural nets perform almost as well as intelligently designed ones.