Nat Friedman and Alexandr Wang discuss AI
Monday assorted links
1. “Before this attack, I had thought that I could take on a bear easily…”
2. China is publishing fewer economic indicators.
3.”A brief Googling suggests Rishi Sunak, Emmanuel Macron, Olof Scholz, Vladimir Putin and Volodymyr Zelensky are all 5’7, it’s an unprecedented consolidation of European power in the hands of short kings” Link here.
Russia sentence of the day
Some local girls married one VDV [part of the military] guy after another, because they’re dying quickly enough that you can get several compensations in a few months.
That is Kamil Galeev, interview with Jordan Schneider, excellent and interesting throughout. The close is with a Tatar folk song.
Pollution and Macro Inequality
Combining 36 years of satellite derived PM2.5 concentrations with individual-level administrative data provided by the U.S. Census Bureau and Internal Revenue Service (IRS), we provide new evidence on the important role that disparities in air pollution exposure play in shaping broader patterns of economic opportunity and inequality in the United States. We first document that early-life exposure to particulate matter is one of the top five predictors of upward mobility in the United States. Second, we exploit regulation-induced reductions in pollution exposure from the 1990 Clean Air Act Amendments to produce new age-specific estimates of pollution-earnings relationship. Combined with individual-level measures of pollution exposure during early childhood, we calculate that disparities in air pollution can account for 17-26 percent of the Black-White earnings gap, 5-27 percent of the Hispanic-White earnings gap, and 6-20 percent of the average neighborhood-earnings effect (Chetty and Hendren, 2018; Chetty, Hendren, and Katz, 2016). Collectively, our findings indicate that environmental inequality is an important contributor to observed patterns of racial economic disparities, income inequality and economic opportunity in the United States.
That’s Colmer, Voorheis and Williams summarizing Air Pollution and Economic Opportunity in the United States. The authors also estimate that
…a 1 μg/m3reduction in prenatal PM2.5 exposure is associated with a $1,105 increase in later-life W-2 earnings and…a 1 μg/m3 reduction in prenatal PM2.5 exposure is associated with a 1.29 percentile rank point increase in upward mobility…We estimate pollution-earnings relationships for each age of exposure from birth to age 12 and show that the relationship between pollution exposure and earnings is stable up to age 4 and then diminishes quickly. We do not estimate a meaningful relationship between particulate matter exposure and later-life earnings from age 8 onward.
These estimate are big but given the substantial number of micro-estimates of the effect of pollution on IQ and cognition that Tyler and I have discussed before (see also this video) substantial effects at the macro level are almost inevitable.
Are unions the saviors of workers?
Every now and then I write a deliberately obvious column, in part because the relevant points are not obvious to everybody. Here is one excerpt:
The brief for unions starts with the possibility of higher wages: Union workers enjoy a wage premium of 10% to 20% (although many of these estimates are dated, and globalization may have made the premium much lower, in some cases close to zero). Unions may also give workers a more effective medium for airing their workplace grievances, and may help them coordinate better services with their employers.
The more skeptical view of unions accepts the logic of these arguments, but questions the size of the associated benefits. The union wage premium is good for the workers that receive it, of course, but it also leads to higher prices. Other workers in turn pay those higher prices. So the net return to all workers is smaller than the wage premium would suggest.
It’s not that labor costs and retail prices move exactly in tandem. If labor costs rise by 12% but labor is a small part of the marginal cost of the product, prices may rise by much less than 12%. But if labor is only a small part of the cost, then unions are less likely to be involved. The more important the role of workers, the more likely that union wage hikes will be translated into retail price hikes.
Workers as a whole still may benefit from unions if the burden of the higher prices falls largely on the wealthy. That may be the case for fancy diamonds or expensive restaurants, but historically unions have been more common in industries that provide goods and services for the middle class, for instance automobiles…
Nonetheless it is plausible that higher union wages do harm some non-union workers, by lowering demand in the long run for labor in unionized sectors. It’s also worth noting that management tends to be hostile to unions, and part of the reason may be that unionization will lead to lower profits. And lower profits will eventually mean lower investment…
Even taking all these factors into account, it is still possible that unions benefit workers overall. But only by a modest degree. It would not be a reason to make unionization a top priority, or to put unions at the center of a theory of what makes workers better off.
Recommended for some, not all.
Sunday assorted links
2. Those new Brazilian service sector jobs (in Portuguese).
3. “After Denmark’s Queen Margrethe stripped the royal titles from four of her grandkids, news has surfaced that Norway’s Princess Märtha Louise may suffer the same fate.” Link here.
4. Knausgaard talk on the novel, recommended to me I have not heard it yet. But for this installment of The Norwegian Century you need to ff to about 29:00.
5. AI writes a thread on productivity hacks.
6. The real Stable Diffusion art.
7. Regulating the homeless (Ezra Klein, NYT).
The Age of Unpredictability
It is much harder to predict foreign policy outcomes, especially in times of turmoil when there is no “nothing happens” default path, than it is to predict the results of economic policies. There is no coherent model, no causal identification in the data, and the data are not very good or well-partitioned to begin with.
Currently it is very hard to predict an endgame for the Russia-Ukraine war. It is also very hard to predict the longer-run consequences of a Putin escalation, including but not restricted to the use of tactical nuclear weapons.
It is very hard to predict the outcome of China-Taiwan, and possible American responses, and this will hang over our heads for some while even if an invasion comes very soon.
With the recent Biden chip sanctions, I don’t think it is ridiculous to assert that the United States is very close to being informally at war with the two other most destructive nations in the world.
Iran-Israel is a neglected wild card in all of this. In many scenarios a superpower war spreads to parts of the Middle East.
You don’t have to fear “Skynet goes live” to recognize that the total net effects of forthcoming AI will be very hard to predict.
Whatever you think of UAPs, the reports are not going away and the observers are not being discredited. The sensor readings are not “mere flocks of geese.” The most likely non-alien drone probe scenario is a U.S. military innovation, but that too would be big news and a major increase in global uncertainty. Imagine the shock when the U.S. first deployed the atomic bomb.
By the way, it seems we are getting pretty effective vaccines for Covid, malaria, dengue, and some forms of cancer? That too was hard to predict until recently.
So we are entering a new time, The Age of Unpredictability. I am not sure what it means to be “ready” for this. But again, cheap hobbies are always underrated.
Emergent Ventures Africa and African diaspora, first cohort
Saron Berhane, Australian-Ethiopian living in Ghana, for research in synthetic biology (bio-leather produced using microbes). Previously she was the co-founder of an agriculture technology startup that specializes in real-time airborne disease detection.
Emmanuel Nnadi, Nigerian microbiologist and lecturer at Plateau State University in Bokkos, for research on phage therapies and the development of a phage bank in Nigeria.
Dithapelo Medupe, Botswana medical doctor and Anthropology PhD candidate at UPenn, to support her research on statistical approaches to multilinear evolution in African development and for general career development.
Mercy Muwanguzi, Ugandan high school senior, to support her research on robotics design and development for medical purposes in Kampala. She was jointly awarded the President’s Innovation Award for Science and Innovation.
Nseabasi Akpan, a professional photographer from Ibadan, for promoting photography education to young people in Nigeria.
Colin Clarke, an Astrophysics undergrad in Ireland, for travel to Nairobi to assist in providing astronomy education to rural schools in Kenya with a non-profit called the Travelling Telescope.
Geraud Neema, Congolese policy analyst living in Mauritius, for in-depth research on the domestic policy impact of Africans educated in China and for general career development.
I thank Rasheed Griffith for leading the selection process behind our new branch of Emergent Ventures Africa and African diaspora.
Saturday assorted links
Moon in June is these days a poltroon
Since the early 1990s, pop song lyrics have become significantly more negative (ht @mattyglesias)https://t.co/tcmOeEWBll
"We conclude that the proliferation of negative song lyrics may be explained partly by content bias … negative lyrics do better in the charts" pic.twitter.com/FTt7ZAIKQp
— Derek Thompson (@DKThomp) October 22, 2022
China, Coppola
Democracy and the Democrats
It is not sufficiently remarked upon that many Democrats have an increasingly difficult time believing in democracy (you can level criticisms at Republicans here too, but those already get far more play, and please anyone who invokes the concept of equivalence here is just a mood-affiliating, internet knee-jerk dummy). To put it bluntly, many Democrats have arrived at the position that democracy works satisfactorily only when it delivers sufficiently low gas prices!
See my previous post, and also the new column by Krugman.”Will Gas Prices Doom Democracy?”. All rhetorical contortions notwithstanding, I don’t see how many current Democrats avoid an implicit or maybe even explicit super-charged skepticism about, yes, democracy.
And it is increasingly hard to blame either “Big Tech” (largely run by Democrats), or campaign finance reform, which now on net is the friend of the Democrats, not the Republicans, at least at the national level.
Matters are easier for the classical liberal. Classical liberals are used to the idea of very bad politicians being elected. If some of those politicians become worse yet, that is highly unfortunate but it does not shake the underlying worldview to its core. Classical liberals also view government as pretty inefficient, often craven, inconsistent, and in many key matters unmanageable. That isn’t always so great! But it does provide some layers of protection against the very worst actors and their possible intentions.
I believe that in October 2022 the classical liberal case for democracy rests on firmer foundations than does the Democratic case for democracy.
Is it worthwhile to teach self-regulation?
Children’s self-regulation abilities are key predictors of educational success and other life outcomes such as income and health. However, self-regulation is not a school subject, and knowledge about how to generate lasting improvements in self-regulation and academic achievements with easily scalable, low-cost interventions is still limited. Here we report the results of a randomized controlled field study that integrates a short self-regulation teaching unit based on the concept of mental contrasting with implementation intentions into the school curriculum of first graders. We demonstrate that the treatment increases children’s skills in terms of impulse control and self-regulation while also generating lasting improvements in academic skills such as reading and monitoring careless mistakes. Moreover, it has a substantial effect on children’s long-term school career by increasing the likelihood of enrolling in an advanced secondary school track three years later. Thus, self-regulation teaching can be integrated into the regular school curriculum at low cost, is easily scalable, and can substantially improve important abilities and children’s educational career path.
That is from a new Nature Human Behavior paper by Daniel Schunk, Eva M. Berger, Henning Hermes, Kirsten Winkel, and Ernst Fehr. There should be more good papers on this topic…
On the five-year breakeven rate (from the comments)
Friday assorted links
1. Why did wind power take so long?, part III.
2. The importance of permitting reform for the environment.
4. Redux of my 2016 post on what is neo-reaction.
5. There has been catch-up at the bottom end of the U.S. wage distribution.
6. Negativity bias in the media is increasing over time, most of all it is pronounced for the Right.
7. Elad on how well will the new AI companies capture value.
That is from an Alex in the comments.