or, Scope-sensitive snafus in summing speculations

(1)

Should I expect monkeypox to be a big deal for the world? Well, fortunately, Metaculus has a pair of questions that ask users to predict how many infections and deaths there will be in 2022:

203 users(!) made 817 predictions of infections, and Metaculus helpfully aggregates those into a "community prediction" of ~248k infections. 77 users made 180 predictions of deaths, with a community prediction of 541.

The y-axis is on a log scale (as are the predictors' distributions). This is a good choice! Whatever you expect the most-likely case to be, there's definitely a chance with things like this that one a misestimation or shift in one factor can make it bigger or smaller by a multiple, not just an additive amount.

What's not a good choice is to report the median outcome of the aggregate position as the "community prediction". This causes a headline reported value that is way too low. Like, four to seven times too low (at least for my intended purposes).

Because the predictors gave (and are scored on) probability distributions, Metaculus will happily give you an aggregate distribution, of which the 248k "community prediction" is the median scenario (the middle of the three dashed lines):

However, on the same plot, the aggregate distribution predicts a 10% chance of at least 4,950k infections. If it's 10% likely to be 5 million infections, then that's already lot more concerning than the 250k in the community prediction! And when I say I'm interested in how much monkeypox to expect, I

Yesterday, the US FDA responded to a submission for Emergency Use Authorization (that was submitted 146 days earlier) for use of fluvoxamine "for the outpatient treatment of adults 24 years and older...to prevent progression to severe COVID-19 and/or hospitalization" (p. 2).

The FDA's conclusion is:

Due to limitations in the available clinical study results for fluvoxamine in the proposed patient population, lack of compelling in vitro and in vivo data to support the proposed [mechanism of action] of fluvoxamine for the treatment of mild COVID-19 disease, and context of increasingly available therapies with well-characterized [mechanisms of action] and consistent efficacy results in nonhospitalized patients, the FDA cannot reasonably conclude that fluvoxamine may be effective for the treatment of COVID-19. As such, FDA has determined that the criteria for issuance of an EUA are not met at this time.

While the FDA has concluded that the existing clinical data are insufficient to support the issuance of an EUA, these data suggest that further clinical investigation may be warranted. (p. 24, emphasis mine)

Personally, I'm disappointed. I've supported the fluvoxamine trials with my personal dollars, and directed more funding from the FTX Foundation to the same trialists for other drug trials (which I remain very, very excited about). But the most productive thing to do with that disappointment is to dig in to the FDA's reasoning in their decision, and decide what's to be done next regarding fluvoxamine.

Important note: This post isn't speaking for the FTX Foundation, the TOGETHER Trial team, the fluvoxamine EUA application team, or anyone but myself in my personal capacity. The FTX Foundation has funded the TOGETHER Trial and committed to additional

To recap:

The Series I savings bond is a US government bond offered to US citizens, with purchases limited to $10k per person per year. It pays interest set by a formula based on the official inflation rate, with a built-in lag. (...)

In particular, the interest rate formula for bonds bought before May 1, 2022 will be:

• 3.56% until October 1
• X% for the next 6 months (where X% is the inflation of CPI-U from Sept 2021 to March 2022)
• Y% for the following 6 months (where Y% is the inflation of CPI-U from March 2022 to Sept 2022)

...and so on. You'll forfeit the last 3 months of interest if you redeem anytime before 5 years (which I'm assuming will be correct to do).

When we checked in December, we'd seen 2 months of the Sept'21->Mar'22 period and I had conservatively estimated that the 6-month inflation rate would come in at 2.17%, implying a 4.57%/ann rate for redeeming after 12 months (and abandoning the last 3 months of interest). But now there are three more months of CPI-U out, and we can make a more-informed estimate.

Inflation has slightly accelerated since then -- Sept->Nov saw 2 months of 8.23%/ann inflation, and Nov->Feb saw 3 months of 8.56%/ann inflation. If we assume (very conservatively, I think) that Feb->Mar inflation will be halfway between 8.43%/ann and historical 2.62%/ann, we get 3.90% for the Sept->Mar interval, and a 6.03% annualized return for holding for the first 15

an incomplete reading list, seeking suggestions

Since announcing my career change to working on biosecurity (Twitter thread), I've had a few people ask me for resources for learning more about (careers in) pandemic preparedness and biosecurity. Here are some resources that I've found helpful in my reading and exploration, in the hopes that they'll be helpful for you, too.

(living list, updated as I find more things)

Other reading Lists

• Chris Bakerlee's biosecurity reading list
• Greg Lewis's biosecurity reading list

Article-to-paper sized content

• 80,000 Hours's article on Reducing global catastrophic biological risks (and more 80k content)
• Open Philanthropy's 2014 cause report on biosecurity
• Andrew Snyder-Beattie's talk Biotechnology and Existential Risk at EA Global London 2017
• Andrew Synder-Beattie and Ethan Alley's Concrete Biosecurity Projects list
• relevant sections of the Future Fund's project ideas list
• proposals for early warning systems from the Council on Strategic Risks, David Ecker, and the Nucleic Acid Observatory
• Erik Minikel's How pharmaceutical industry financial modelers think about your rare disease
• Kevin Esvelt's "Inoculating science against potential pandemics and information hazards", in PLoS Pathogens
• The Apollo Program for Biodefense final report

Blogs and substacks

• De Novo, a substack by Metacelsus
• Construction Physics, a substack by Brian Potter
• ASeq, a substack by Nava Whiteford
• Jeff Kaufman's posts on sequencing, from working at the Nucleic Acid Observatory

Books

• The Precipice, by Toby Ord
• Deadliest Enemy, by Michael Osterholm and Mark Olshaker
• Biosecurity Dilemmas, by Christian Enemark
• The Dead Hand, by David Hoffman
• (for basic biology) Molecular Biology of the Cell, by Bruce Alberts et al.
• (for basic biology) Cell Biology by the Numbers, by Ron Milo and Rob Phillips

This January, I left my job as a trader at Jane Street (most recently in Hong Kong) to take a role leading a project at the FTX Foundation. I have positive things to say about Jane Street as an excellent place for many of my friends to work, and my open offer of thoughts and advice to anyone hoping to work there remains, well, open.

Looking forward, I'm thrilled to be working with the Foundation because I want to help build a better future. I've been interested in thoughtful, rigorous, altruistic attempts to improve the world since 2013, and pledged in 2016 to donate at least a tenth of my income to the most effective organizations I can find. I am beyond excited to find an opportunity where I estimate I'll do even more good than I could by earning and donating a quant trader's salary (which I still think was really, really good for the world, on any objective scale).

It's still early days here, and we don't yet know all that the things the Foundation's work will cover. To start, colleagues of mine just launched the Future Fund, a grantmaking initiative with the goal of improving the lives of future generations in highly levered ways. The launch blog post, Areas of Interest, and Project Ideas pages give a taste of where the Future Fund has set their sights, and I'm extremely excited about the possible ways their work can go from here.

For my own work, I'm not working on grantmaking or the Future Fund directly. Rather, I'm currently heads-down on an in-house skunkworks moonshot to make the world safer against catastrophic pandemics. (More on that later.

At just under nine years since I started blogging, I'm changing this blog to a new name. There's no huge story here, just a vague but growing dissatisfaction with My Faults My Own as my brand for personal writing.

(1)

I took the name My Faults My Own in August 2013, in homage to Vi Hart's video Doodling in Math Class: Connecting Dots, particularly the quote:

Here's the thing about connecting dots. You can have all the steps laid out for you, taking whatever next step is easiest and closest and be sure of what you're getting the whole time. This way is safe and comfortable.

Or, you can try new ways of connecting dots and not know what you're going to get. Maybe it will be something great, maybe it will fail. And when it fails it will be your fault. You can't blame anyone else, not mathematics or the system or the check-boxes.

But if I am to have faults, I would rather they be my own.

I continue to believe that {intellectual self-ownership, independence, creativity} are effectively necessities for strong analytical ability, which you very much do want to have on your side.

But I also have other intellectual principles, and elevating this particular one to pride of place in the title has stopped seeming right to me. My faults are my own, I would rather they be my own, and I expect that I'll believe that approximately forever. I'm well settled into that stance, and I'm not going back. But I don't blog under Water is Wet or Sky is Blue, and blogging under a declaration of independence borrowed from another...eh, it doesn't quite

This post describes my thoughts, at the end of 2021, about donating money to make the universe a better place. I remain committed to using at least 10% of what income I earn to do so.

This year is the first time since 2017 that I'm substantially changing my approach to donations. In 2017, I shifted my focus from mainly global poverty/health to a mix of global poverty, animal welfare, long-term future, and meta EA. This year, I'm shifting to a mix of global poverty, animal welfare, and explicitly 'exploratory' categories.

(1a)

Perhaps the most important long-term development in effective altruism in 2020 and 2021 has been the crypto boom, which has moved at least $30 billion of wealth to committed effective altruists. (The most widely visible of these is Sam Bankman-Fried, CEO of FTX.) I would roughly estimate that the top ~ten self-identifying EAs now have plans to move at least $45 billion of donations to EA causes in the next ~30 years -- even in the case that all of their various business investments mostly stagnate at current valuations.

That's a lot of money.

Given this landscape, my goals as an EA donor this year were (1) to move my personal focus to finding and supporting things that I expect existing large donors to have difficulty finding, evaluating, or supporting, and (2) to use my giving to 'set up' for having a larger impact later, by exploring things that I can