Mo Reads: Issue 12
Nuclear, animal welfare reform, operations, modeling, category theory, biodiversity,
Past issue, archive. First issue in about 2 years; only 7 reads instead of the usual 10.
Reads:
Nuclear energy - Our World in Data by Hannah Ritchie et al (5,000 words, 20 mins)
A Cost-Effectiveness Analysis of Historical Farmed Animal Welfare Ballot Initiatives by Laura Duffy (3,500 words, 14 mins)
Things I Learned by Spending Five Thousand Hours In Non-EA Charities by Jenn (2,400 words, 10 mins)
How Long Should you Take to Decide on a Career? by Applied Divinity Studies (2,700 words, 11 mins)
How an invariant from category theory solves a problem in mathematical ecology by Tom Leinster (118 slides, 20 mins)
Shopping for happiness by Jacob Falkovich (3,000 words, 12 mins)
Recursive Middle Manager Hell by Raemon (3,400 words, 14 mins)
Nuclear energy - Our World in Data by Hannah Ritchie et al (5,000 words, 20 mins) — some figures that jumped out at me:
Global generation of nuclear power grew from 25.54 TWh in 1965 to 2,721 TWh at its peak in 2006, for a >100x increase or +12% YoY growth sustained over 41 years vs the 2.7x growth in overall world energy consumption. In the 16 years since, it’s basically plateaued at 2,610 TWh (2022), while world energy consumption continued to increase by another +23%
The EU has a population of ~450 million and consumes ~6,400 kWh per capita, for a total annual electricity consumption of about 3,000 TWh. If the EU were powered completely by coal power plants (which in Europe have much better pollution controls than they do elsewhere), about 74,000 people would die prematurely every year, mostly from air pollution. Switch to oil and that drops to 55,000 premature deaths a year; switch to gas and that drops further to 8,500 premature deaths a year; switch to nuclear and that drops all the way down to less than 100 premature deaths a year on average — OWID calculates this figure by using a death toll of 2,314 for Fukushima (>95% of them due to evacuation stress and disruption to healthcare facilities) and 433 for Chernobyl (upper estimate; <100 actually confirmed, not all from acute radiation syndrome). Solar, wind and hydropower are all in the same ballpark as nuclear, at <150 premature deaths
Same picture from a greenhouse gas emissions perspective: if the EU were powered completely by coal plants it would emit ~2.5 billion tons of CO2e a year; switch to oil and that annual figure drops a bit to 2.2 billion tons of CO2e; switch to natural gas and that drops to 1.5 billion tons; switch to nuclear and that just falls off a cliff down to <10 million tons. Solar and wind are in the same ballpark; hydro is ~10x that, albeit still <95% lower than coal / oil / gas. For perspective, Speed & Scale notes that global annual CO2e emissions were 59 billion tons in 2020, or ~25x the EU-coal figure above, and that the global budget for limiting warming to under +1.5°C is 400 billion tons (which we’re therefore on track to hit by 2027 or so)
A Cost-Effectiveness Analysis of Historical Farmed Animal Welfare Ballot Initiatives by Laura Duffy (3,500 words, 14 mins): summarizes her 132-page document for Rethink Priorities that looks at 4 US ballot initiatives which “either restricted the use of common animal confinement methods (including extremely confining stalls and tethering for veal calves, extremely confining gestation crates for breeding sows, and conventional cages for egg-laying hens), set minimum per-animal space requirements, and/or mandated cage-free systems for egg-laying hens”. One key takeaway is that helping egg-laying hens is by far the most cost-effective farmed animal welfare reform, whether you look at animal-years improved or years of suffering avoided per dollar spent: “~99% of the reductions in farmed-animal suffering by these four initiatives can be attributed to bans on battery cages and/or cage-free requirements”. This matches the (sloppy but directionally informative) intuition that “chicken is more unethical per kilogram than beef because eat chicken for dinner and you've killed half a chicken; eat beef for dinner and you've killed less than 1% of a cow”. One way to reject this conclusion in a principled quantitative manner is to morally weigh species not equally, but as linear in brain mass or max synapses, if you think that’s a better rough proxy for “capability for suffering” (all proxies are extremely rough); another way is to argue that indirect suffering caused by beef consumption is substantially greater than that caused by chicken.
Things I Learned by Spending Five Thousand Hours In Non-EA Charities by Jenn (2,400 words, 10 mins) has some observations and recommendations gleaned from experience in operations roles:
Long-term reputation is priceless and unlocks a huge amount of value that’s hard to immediately buy with money (e.g. amenity rentals; the mayor’s endorsement; business services; pro-bono and monetary donations), so it’s worth thinking more about what actions and trade-offs orgs should take so they’re beloved institutions in 25 years’ time
Nonprofits shouldn’t be islands, so to take advantage of specialization and system-wide synergy, evaluation at the single-charity level should be de-emphasized slightly in favor of group-level evaluation, in particular [nonprofit + government]; another is to ask (in grant applications) not just “if we gave you 10x the money you requested, what would you do with it?” but also “what is your fantasy partner / complement organization?”
Organizational slack is powerful, because different parts of multifaceted orgs go through crunch time at different times and slack lets you regularly temporarily reallocate people to smooth out the spikiness, so it’s worth thinking through what sorts of resources can be constrained besides money (especially staffing), e.g. having lawyers/payroll/HR people on retainer on hand to consult with fledgling nonprofits who aren’t big enough to hire them full-time
Hospitality matters, because people won’t use your service if it seems impersonal and cold, even if their livelihoods depend on it (frequent feedback in this vein: “I’m here because Samaritans are the only ones that will actually listen to my problems”), so it’s worth having public-facing nonprofits and infra orgs care more about customer service
How Long Should you Take to Decide on a Career? by Applied Divinity Studies (2,700 words, 11 mins): ADS numerically models in Colab both the original secretary problem (given N options, and subject to certain constraints, you should evaluate 37% of them before committing) and a modified version (call it the MSP) that relaxes (in fact parametrizes) SP’s most problematic assumptions: (1) binary payoff (i.e. “2nd best is just as bad as the worst”), (2) no opportunity cost (i.e. no cost to making a selection after 100 seeing candidates rather than after 10 — relaxing this introduces a kind of explore-exploit tradeoff). The takeaway is thematically resonant with the classic essay Reality has a surprising amount of detail: while the MSP is arguably more realistic than the original SP, its solutions are dependent on plausible choices of parametrization to the extent that the results are not necessarily more applicable to real-world decisions like career choice; yet it can still useful against the sometimes-limitless badness of intuition (whether due to them being maladapted to the current moment or to your circumstances, or due to neurosis-driven decision-making)
How an invariant from category theory solves a problem in mathematical ecology by Tom Leinster (118 slides, 20 mins): how should biodiversity be quantified, and how should it be maximized? To first order, disagreements about how to quantify it map to a spectrum of viewpoints, one end of which is “what matters is conserving species” and the other end “… communities”. This “viewpoint spectrum” can be parametrized; call it q. Tom shows how, for a given list of species (and pre-specified similarity matrix), there is always a frequency distribution that maximizes diversity for all viewpoints (so everyone on the ‘species-to-communities’ spectrum is happy), given by a simple expression that uses a definition of ‘magnitude’ from category theory originally conceptualized to illuminate the connections between a few different ‘canonical notions of size’ in pure math. (Aside: I like how Tom does the storyboarding for this deck; what other presenters would compress into 20-30 slides he stretches out over 118 in a way that works wonderfully in-presentation.)
Shopping for happiness by Jacob Falkovich (3,000 words, 12 mins): summarizes the book Happy Money’s five principles for how to spend money to be happy (1/ buy experiences 2/ make it a treat 3/ buy time 4/ pay now, consume later 5/ invest in others) into a single equation: cost of happiness = cost / (duration x intensity). Some concrete memorable takeaways: (1) buy great soap, undies, toilet paper, socks (and more generally, buy excellent products in cheap categories) (2) eat popcorn with your non-dominant hand, and interrupt pleasant experiences like TV shows instead of binge-watching (more generally, use scarcity and novelty to keep experiences fresh and enjoyable) (3) follow these tips to buy time, focusing on buying out the worst hours of your life first (e.g. laundry) (4) buy story-able memories e.g. a trip to space (5) spend voluntarily for others in a way that creates personal connection (but separate this from effective giving)
Recursive Middle Manager Hell by Raemon (3,400 words, 14 mins): adding layers of middle management makes an org “more deceptive, less connected to its original goals, more focused on office politics, less able to communicate clearly within itself, and selected for more for sociopathy in upper management” in a way that’s “deeper and more insidious than you're probably realizing, with much higher costs than you're acknowledging — if you're scaling your org, this should be one of your primary worries”. Robert Jackall calls this phenomenon moral mazes, “where middle managers spend most of their time and energy on internal status competitions rather than improving the company's products”; Zvi has some suggestions for protecting against mazedom in large projects. A commenter who used to be a middle manager at Google noted that mazedom manifested there in 2 ways:
1/ If you try to make your organization productive by focusing your time on intensively coaching the people under you to be better at their jobs, this will make your org productive but will not result in your career advancement. This is because nobody at the level above you will be able to tell that the productivity increase is due to your efforts — your reports' testimony to this effect will not provide appropriate social proof because they are by definition less senior than you. To advance your career you must instead give priority to activities which call you to the attention of those who can provide that social proof. This is called "managing up and across."
2/ In order to ensure that the organization works in consistent, fair, legal, ethical, and legible ways, corporate policy in a multilayer organization tends to put more guardrails around the behavior of middle managers than on those either above or below them. This strips those middle managers of the feeling of agency and autonomy which might otherwise provide a non-ladder-climbing intrinsic motivation to do the work. Thus it strengthens the selection pressure for those whose main motivation is ladder-climbing.
One way to mitigate this, by Jeff Bezos (quote by a recipient of his infamous ‘?’ emails):
About a month after I started at Amazon I got an email from my boss that was a forward of an email Jeff sent him. The email that Jeff had sent read as follows:
“?”
That was it.
Attached below the “?” was an email from a customer to Jeff telling him he (the customer) takes a long time to find a certain type of screws on Amazon despite Amazon carrying the product.
A “question mark email” from Jeff is a known phenomenon inside Amazon & there’s even an internal wiki on how to handle it but that’s a story for another time. In a nutshell, Jeff's email is public and customers send him emails with suggestions, complaints, and praise all the time. While all emails Jeff receives get a response, he does not personally forward all of them to execs with a “?”. It means he thinks this is very important.
It was astonishing to me that Jeff picked that one seemingly trivial issue and a very small category of products (screws) to personally zoom in on. ...
Aside: it’s somewhat hard to ballpark what fraction of people work in moral mazes.