Turn EVERYTHING into math!
The 2nd newsletter in the neilsletter series
Hello everyone! Second time I sent updates about what I find important to my closest friends. You may read this, skip this, or tape this to the wall and pray to it every night. This week I’m just going to rant about a single concept: putting numbers on things is helpful.
Tl;dr: putting numbers on things lets you better achieve your goals. You don’t have to be a “math person” to do this, or be “obsessed with optimization”, or be a high-functioning autist. It’s common sense. Seriously, you should just try putting numbers on things.
Ω
Humans got powerful in the first place by turning things into math. Want more water? Run the numbers and you can build a structurally stable aqueduct. Want to defeat your enemies? Run the numbers and you’ll know how big each of your armies and supply chains should be. Want to raise taxes? Determine how much money people make and how much you can take without wrecking the economy.
Putting numbers on things is how average humans like you can make the insanely delicate and precise machine you’re reading this on. Math is unreasonably effective at getting things done.
Given the unreasonable effectiveness of math, I tried putting numbers on things. I’ve been sending a survey to myself every night using this tool to collect data on my life.
Life data:
I’ve been measuring
Happiness, on a scale from 0 to 1
Agency, from 0 to 1
Procrastination, from 0 to 1
Total helpfulness to my goals, [0;+∞]
Whether I managed to do at least two things on my to-do list
Bedtime, measured in minutes after midnight
I’m still attempting to figure out how to translate the raw data from 50 days of this into useful correlations and fancy graphs. I told myself I would send this email by the end of today and I’m out of time.
Before I continue, I’m going to go on an
Economics tangent!
The economy works by putting numbers on things. Money is a measure of value. Capitalism is the idea that if you let people make whatever choices they want, they will naturally pick what they most value, and supply will naturally respond to demand: the result will be an equilibrium where the price of something is just right, like the Goldilocks zone.
An unimaginable amount of variables are squeezed into the price of something. No-one independently decided a tissue box (2 euros) was worth a bag of corn (2 euros) or a 1-hour bus ride (2 euros). The markets just… decided that. An excellent, beautifully written essay called I am a Pencil is an ode to this “invisible hand of the market”.
There’s a saying: “economics is the deadliest science”. If you mess up the price of something, literally millions will die (cf. Soviet Union). Get it right! Free markets don’t always value things in the Goldilocks zone either. For instance, the future costs of climate change aren’t included in the cost of oil well projects.1 And the risk of human extinction is not included at all in the cost of AI projects.2 This doesn’t mean, if you’re a policymaker, you should rush to fix this. In both these cases, going all Soviet Union on things and rigging the price yourself can often backfire. These are fabricated options, where you might think you can rig the price of e.g. life-saving supplies during a crisis, but this just makes things worse.
1: To be clear, the Soviet Union and Communist China are responsible for horrible environmental damage as well. People somehow think criticizing capitalism corresponds to a defense of socialism: not so.
2: Breakdown: If you value a life at 28,000 euros, then human extinction costs 28 000 * 8 000 000 000 = 224 000 000 000 000 euros. 224 trillion is more than twice annual world GDP. The average probability AI researchers give that everyone gets slaughtered by AI in the next 9 months is somewhere in between 4% and 20%: we’ll say 10%. This corresponds to 22 trillion dollars of lost value this year from AI extinction, which is not taken into account by the markets!
I’ll write someday on how actually, it’s totally fine to slap a number onto human lives—not because human value isn’t infinite (I believe a human life is worth a lot more than 28,000 euros)—but because you should tap into the unreasonable effectiveness of math.
How much is your work worth?
Okay. The reason I went on an economics tangent is to point out how useful it is to put a dollar value on things. I explain in this article how putting a dollar value on lives gives you a visceral idea for just how much you care about things, and how big our problems are. I think you can also put an approximate dollar value on your work.
I just finished translating a paper by Nick Bostrom, an Oxford professor. Google says that lower-end translators cost about 10 cents per word, and there were 2,500 words in his essay. That means that I’ve given Nick Bostrom, for free, about 250 dollars of value. If I run the translation by a few people who actually know what they’re doing, perhaps I can inch closer to the results of an upper-end translator (30 cents per word), which would make my work worth 750 dollars.
I got 70 survey responses from Sciences Po students by walking up to them during portes ouvertes, and 30 Magendie students by putting QR codes on the walls. Through online services, getting 100 survey responses costs 200-500 dollars. You can also do this free on a survey-for-survey basis, if you wish to replace money with time. At an average of 5 minutes per survey that’s 100 * 5 = 500 minutes, or over 7 hours of your life. To put it another way, spending money to run this survey would have valued my time at 200 / 7 hours = 28 dollars per hour, which is two and a half times the SMIC. Unless I get a job that pays more than the SMIC, it’s probably worth it spending the 6 hours then.
Figure out how much your time is worth! Figure out how much you would like for your time to be worth! Try calculating the dollar/minute effectiveness of scrolling on TickTokk versus reading a book, or writing a blog post! Cost/benefit calculation for everything! Everywhere! All the time! MWAHAHAHA!
Conclusion:
I can’t identify as a “math person”, but I don’t want to identify as “not a math person” either. I’m not compelled by curiosity alone to search for exciting new theorems or proofs (at least I won’t prioritize that curiosity, given I’m curious about plenty of other things), but that doesn’t mean I can’t use math from time to time to figure some important stuff out. Given my friend group, you probably don’t identify as a math person either. So… keep this in mind!
Next up: Bayes' theorem! (Yeah that's AI generated, I had to find some way to justify the one-month subscription I got for GPT-4.)