(1)

I met Max Chiswick in June of 2024 because I wrote something wrong on the internet. He tracked me down at Manifest to tell me so—I had written about the shortcomings of poker as a teaching tool, but I had made several wrong assumptions about how one would play poker differently with the goal of learning instead of recreation.

First, you'd play with two players instead of eight. We nearly always play with eight-ish players because we want a relaxing game, where large gaps in the action are considered a feature. With two players, though, not only is it your turn four times as often, but more of your decisions are "live" because it's correct to play more (initial) hands and there are fewer that you should fold on sight.

Second, you'd play a game with fewer chips—meaning that you'd make bets in coarser increments and with a smaller ratio of largest-possible bet to smallest-possible bet. This makes the decision tree shallower and more amenable to explicit case-by-case reasoning, and it also has the nice effect of making the highest-stakes decisions less terrifying (while keeping the stakes of bread-and-butter decisions suitably meaningful).

But most importantly, if you're playing poker as an exercise for training your mind, you should focus less on becoming a theoretically-optimal robot. You can read books and blogs about "game theory optimal" play (e.g., Nash equilibrium) and drill yourself to get arbitrarily accurate at executing it, but that's simply not the useful part.

The part of poker that makes you smarter—Max explained to me—is observing the non-optimal patterns in your opponents' play, and

This post describes my thoughts, at the end of 2024, about using money to make the universe a better place. I remain committed to using at least 10% of what income I earn to do so, and am excited to do more than that when I have the opportunity.

This year marks the tenth anniversary of my first $4,000 donation to GiveWell's top charities! (That donation was 10% of my summer internship salary, plus some other campus jobs.)

A lot has changed since my last post in 2021, only some of which I'm able to recap here. (I have tried to publish these posts annually, but missed 2022 and 2023 for idiosyncratic reasons.) I'll break this post into (1) general discussion and personal outlook, [incomplete], and logistics, (2) donations by cause area for 2024, 2023, and 2022, and summary lists, (3) [incomplete: events of 2022 and 2023], and personal-policy updates, and (4) other people's writeups that I have found interesting.

(1a)

Shortly after my last donations post, I left a career in quantitative trading where I had been

New drugs being developed can be "easy" drugs or "difficult" drugs.

In order to know whether your drug candidate is safe and effective, you're going to test it in a series of clinical trials. In each trial, you'll recruit some number of patients, give each patient the treatment, placebo, or a comparator drug, wait some time, and test them for pre-specified endpoints.

Within that framework, however the trials for different drugs will differ greatly. (And the different phases of a single drug's trials may differ by even more!) Typically, the greatest axes of variation will be:

• Who are your patients?
  • How common is the indication that you're treating? How often do people go to your trial site to get treatment for it? How many of them want to be in a trial?
  • Is your trial taking anyone with the disease? / Is it only for people who are not responding to some other standard treatment?
  • Are your patients otherwise healthy? / Do they have elevated risks for other complications?
• What condition does the drug affect?
  • Are you trying to change something that patients already have? / Are you trying to stop them from developing something else?
  • If you're preventing something, what fraction of your patients will develop it without treatment?
  • What change in the condition are you trying to measure? Is it yes-or-no or on a scale?
  • If the drug "works", what fraction of cases will it change enough for you to measure?
• Where does the trial take place?
  • Are you treating patients in a hospital?

Some trading firms are driven by good decisions made by humans. (Some aren't, but we can set those aside. This post is about the ones that are.) Humans don't make better-than-average-quality decisions by default, so the better class of intellectually-driven quantitative trading firm realizes that they are in the business of training humans to make better decisions. (The second-best class of firm contents themselves with merely selecting talent.) Some firms, famously, use poker to teach traders about decision making under uncertainty.

First, the case for poker-as-educational-tool: You have to make decisions. (Goodbye, Candy Land.) You have to make them under uncertainty. (Goodbye, chess.) If you want to win

“Why can’t a robot be built without the First Law? What’s so sacred about it?”

Dr. Gerrigel looked startled, then tittered, “Oh, Mr. Baley.”

“Well, what’s the answer?”

“Surely, Mr. Baley, if you even know a little about robotics, you must know the gigantic task involved, both mathematically and electronically, in building a positronic brain.”

“I have an idea,” said Baley. He remembered well his visit to a robot factory once in the way of business. He had seen their library of book-films, long ones, each of which contained the mathematical analysis of a single type of positronic brain. [...] Oh, it was a job, all right. Baley wouldn’t deny that.

Dr. Gerrigel said, “Well, then, you must understand that a design for a new type of positronic brain, even one where only minor innovations are involved, is not the matter of a night’s work. It usually involves the entire research staff of a moderately sized factory and takes anywhere up to a year of time. Even this large expenditure of work would not be nearly enough if it were not that the basic theory of such circuits has already been standardized and may be used as a foundation for further elaboration. The standard basic theory involves the Three Laws of Robotics: the First Law, which you’ve quoted; the Second Law which states, ‘A robot must obey the orders given by human beings except where such orders would conflict with the First Law,’ and the Third Law, which states, ‘A robot must protect its own existence as long as such protection does not conflict with the

and some notes on Silicon Valley Bank

A "liquidity" crisis

There's one kind of "bank run" where the story, in stylized terms, starts like this:

• A bank opens up and offers 4%/ann interest on customer deposits.
• 100 people each deposit $75 to the bank.
• The bank uses $7,500 to buy government debt that will pay back $10,000 in five years. Let's call this "$10,000-par of Treasury notes", and call that a 5%/ann interest rate for simplicity. (Normally, government debt pays off a bit every month and then a large amount at the end, but that's just the same thing as having a portfolio of single-payout (or "zero coupon") notes with different sizes and maturity dates, and the single-payout notes are easier to think about, so I'm going to use them here.) We're going to assume for this entire post that government debt never defaults and everyone knows that and assumes it never defaults.
• The thing you hope will happen is for every depositor to leave their money for five years, at which point you'll repay them $95 each and keep $500—which is needed to run the bank.
• Instead, the next week, one customer withdraws their deposit; the bank sells $100-par of T-notes for $75, and gives them $75 back. No problem.
• A second customer withdraws their deposit; oops, the best price the bank can get for $100-par of T-notes, right now after it just sold a bit, is $74. Problem.
• But next week, let's say, it would

Junk fees are in the news from the 2023 State of the Union address, get picked up by Matt Yglesias, and Zvi responds in Junk Fees, Bundling and Unbundling. Matt and my former colleague propose an economic framing of "bundling versus unbundling", and Zvi identifies four win-win advantages of bundling, and four 'advantages' of unbundling (two win-wins, one (company win)-(customer lose), and one mixed win-lose).

I think Zvi is broadly right on the points he makes, but he and Matt both skip over the basic, conventional econ-101 analysis of bundling goods on prices and customer welfare. I think that, for a broader audience, it's worth covering the "naïve microeconomics" perspective as background for the customer-behavioral story (which, admittedly, is more juicy and fun). Rather than responding to the whole conversation, this post will restrict its focus to the econ-101 microeconomics story of bundling, and ignore the behavioral / political / moral dimensions that Zvi, Matt, and others are discussing.

I'm going to assume no formal economic background, and build up an explanation why a world of perfect economic agents should have sellers preferring to sell bundles of goods instead of individually-priced items.

(1) Selling sandwiches

Let's say that you're a sandwich-seller. Every day, 110 people come by your sandwich stall, look at the price you put on the menu, and buy a sandwich if they value it more than the price. (No bargaining or haggling; you don't have time for such things.)

Sandwiches cost you $1 to make, and people value sandwiches somewhere between $0 and $11, fairly uniformly:

[ed: The first-best version of this post includes interactive charts with sliders that let you explore