If I Was An Eccentric Trillionaire

cross-posted from niplav.site

What I might do if I magically got hold of a very large amount of money, and couldn’t spend it maximally altruistically.

I sometimes like to engage in idle speculation. One of those speculations is: “If someone came up to me and told me that they would give me a lot of money, but only under the condition that I would spend most of it on unconventional and interesting projects, and I was forbidden to give it to Effective Altruist organizations narrowly defined, what would I do? Not disallowing the projects from having positive consequences accidentally, of course.”

The following is a result of this speculation. Many of the ideas might be of questionable morality; I hope it’s clear I would think a bit more about them if I were to actually put them into practice (which I won’t, since I don’t have that type of money, nor am I likely to get hold of it myself anytime soon).

Lots of these ideas aren’t mine, and I have tried to attribute them wherever I could find the source. If guess that if they were implemented (not sure whether that’s possible: legality & all that) I’d very likely become very unpopular in polite society. But the resulting discourse would absolutely be worth it.

Intervention Cost
Snowball fights ?
Buy a small island nation $5 bio.
Personal futarchy on steroids $100 mio.
Save dying languages $2 bio.
Raise native speakers of an engineered conlang $30.8 mio.
Philosophically solve language $10 mio.
SCP series $1 bio.
Antimemetics Division spinoff $200 mio.
Discontinuous & fast AI takeoff movie $500 mio.
Double Crux podcast $2 mio.
Fictional ethnography of native Antarcticans $100k
Studying foreveraloners $900 mio.
Really Out There Stuff Institute $60 mio.
Sum $ 9.81 bio.

Culture

  • Snowball fights: I remember having a lot of fun with snowball fights during breaks in school, but I haven’t seen them happen anywhere outside since leaving school. People sign up for participating in (moderated) snowball fights, and are notified on weekends with especially good snow. Large open grassy fields (or perhaps even stadiums) are rented, and separated into at least 2 parts, each parts gets a large pile of fresh snow (to minimize the chance of stones in snowballs). People are encouraged to bring protective gear to allow for rougher fighting. If one is to allow for physical encounters, one’d need to group by fitness/​size.

  • Buy a small island nation: Since other billionaires are now unable or unwilling to buy a small remote island for eccentric trillionaire purposes, this duty now falls on us. Owning an island has several advantages, such as being able to provide shelter during some catastrophic events and be a base for organizing other eccentric activities. Plus points if the island owned is a nation state, as one could pass ones own laws (within the bounds of international agreements, of course). Let’s say we e.g. try to buy Nauru[1]. Nauru’s GPD is ~$135 mio., which at a discount rate of ~5% has a net present value of , which we can round up to $3 bio. What if we instead try to buy Nauru from each Nauruan individually? The average GDP per capita of Nauru seems to be ~$12k, but to be conservative we can round that up to $15k. Then the net present value (again at 5%) of each Nauruans future income is . There are 11k Nauruans, which results in an expense of . In both cases, the Nauruans would be giving up a significant amount of their civil rights, and might want to find new citizenships, to support them with this one could allocate more money so that the sum nicely comes out to $5bio. So it would be financially feasible to do this, but would it be politically and legally feasible? I don’t know about that, and don’t know of any precedent either. Leopold II. privately owned the Congo Free State, but I don’t think he bought it, and instead convinced the other European states to be allowed to militarily seize it. (Our intentions are far more benevolent (and probably weirder) than Leopold’s). Nation states have bought vast swathes of territory from others (a prominent example being the Alaska Purchase for $140 mio. 2021 dollars (peanuts!), and there were attempts from the US side to also buy Greenland from Denmark—ultimately unsuccessful. More examples here), but to my best knowledge nobody has ever acquared a country. (Yet. Growth mindset.)

  • Personal futarchy on steroids: In You Can Do Futarchy Yourself, Tetraspace outlines a way of implementing a preliminary versions of Futarchy (using prediction markets to determine policies whose outcomes are predicted to be optimal alongside some pre-defined metric) by submitting conditional questions to forecasting platforms like Metaculus or Manifold Markets.

    • As an example, say one only cared about GDP. Then one could submit the questions “If a Democratic candidate wins the 2024 presidential election, what will the US GDP be in 2026?” and “If a Republican candidate wins the 2024 presidential election, what will the US GDP be in 2026?”, and, depending on the forecasts others make, decide to vote one way or the other. Online forecasting platforms are fairly reliable, especially for short-term political questions, so the information would be quite valuable.

    • More is possible: Another nice thing about this proposal is that it is very scalable: The minimum viable product can be made by a motivated individual in one evening, but more sophisticated versions are possible. One possible extension could look like a voting advice application, similar to the German Wahl-O-Mat or the international Vote Compass.

    • Create markets on elections: For a given national election, a team of political scientists would enumerate the lists of likely outcomes of the election (either determining which coalitions are likely, or (in single-winner elections) which parties could win). They would also collect common & clearly measurable desired outcomes the voters might have for the election, such as GPD, life expectancy, crime levels, the Human Development Index and others. (It is possible to define more specific criteria, such as whether certain laws would be passed, but this might slide into already specifying beforehand which party would come out ahead). For a specific election , candidate and indicator , a question would then be created on a prediction platform (or, even better, market): “If gets elected in , what will the value of be at the end of the legislative period of ?” (There are some problems here with the many markets that don’t resolve, which we will ignore for the time being, perhaps combinatorial prediction markets or latent variable prediction markets can help, but I don’t know enough about them).

    • Make a website: One would then create a website giving users sliders choose which indicators they care more or less about, and about the desired sign of those indicators[2]. Using the probability distributions over indicators from the prediction platforms, the website would then compute the expected value for each candidate and report the list of candidates to the user, sorted by desirability — giving them information about which candidates are most likely to actually succeed at giving them the outcomes they desire, and thereby influencing voting behavior. The whole website would be accompanied with a short video explaining the concept, and perhaps a longer explainer text going in detail. Voting platforms are fairly popular, sometimes drawing >10% of the population to use them.

    • Cost: The underlying software should not be very expensive (a website, an app and a backend that could support tens to potentially hundreds of thousands of visits a day), and could be re-used for each election: An initial expense of ~$5mio., with ~$500k per ear of maintenance seems on the conservative side for an estimate. Assuming we want to supply the elections of the 40 biggest (democratic) countries with information, there’d be (assuming an electoral cycle of ~4 years) ~10 elections a year — something a team of 5 political scientists (~$100k a year each) should be able to handle. Assuming that for each election, there are ~6 candidates/​coalitions that cover most of the election outcomes, and ~6 indicators we want to use to evaluate election outcomes, subsidizing each market with $10k comes out at . If the whole thing runs for 20 years, we then pay . This is surprisingly cheap, so I’ll fudge upwards to $100 mio, just in case. Still a steal, if you ask me. Just not sure whether I’d have to fear angry politicians.

Language

  • Save dying languages: There are ~6900 living languages in the world, but the number of speakers for languages is heavy-tailed: ~500 of those languages are nearly extinct, likely due to very few and elderly native speakers. Extinct languages are a loss of the cultural heritage of humankind, so just as we want to save endangered species or artworks from their destruction, it would be cool to save nearly extinct languages from their demise. This is not just hypothetical: From 2013 to 2023, 19 languages have died. To do this, we employ linguists to learn each of the the endangered languages and practice it. 10 linguists per language should be enough, so we employ 5000 of them at ~$50k a year, for 4 years, to spend learning the language full-time, which yields us expenses of ~$1 bio. Then say that we continue employing them part-time for 20 years to continue practicing until we’ve found a way to permanently store the languages, for example by training large language models (text and audio) on those languages; this would give us (at $10k salary a year) another $1 bio. in expenses, plus whatever cost the preservation procedure entails.

  • Raise native speakers of an engineered conlang: Spurred on by the Sapir-Whorf hypothesis, people have created several engineered constructed languages. Such languages attempt to introduce new paradigms to spoken languages and aim for extreme properties along one or more axes: emotional expressivity (Láadan), complete grammar and lack of syntactic ambiguity (Lojban), simplicity (Toki Pona), expressivity and semantic richness (Ithkuil)…
    We know that people can use a constructed language as their native tongue, as there are >1k native Esperanto speakers in the world. But I do not know of any examples of raising a child primarily on a language engineered to exceed the bounds of natural language, the closest being this video. So it would interesting to pay some new parents (ideally already both speakers of the engineered language) to raise a child with that language. The difficulty of achieving this depends on how difficult the target language is to learn, and how many speakers there are: Toki Pona should be easiest (allegedly has ~100 speakers), followed by Lojban (hard to learn, has ~15 speakers) and Láadan (perhaps easier to learn, but less developed and there are negligibly many speakers (and therefore likely none willing to raise a child)), Kēlen would be quite difficult (since there are probably no fluent speakers, and speakers would need to be trained) and Ithkuil is probably impossible, as even the creator can’t speak it fluently. I don’t know what price parents would put on raising one of their children in primarily the constructed language, which might be in the highest case several hundreds of thousands of dollars per year: If we have two children in different families per language, and pick Toki Pona, Láadan, Lojban and Kēlen, at $200k per parent and year, until the child is 18 years old, we pay . We know that children can be bilingual, so the danger of inability to communicate can basically be excluded—and since money is not a huge issue, one could offer a ~$10 mio. insurance against worst-case outcomes. If we assume that worst-case outcomes are possible but unlikely, we pay (in expectation) , for a total of $30.8 mio.

  • Philosophically solve language: Related to raising native speakers of engineered conlangs, I am not very impressed by the degree of effort that has gone into trying to philosophically solve language—all work has been done by hobbyists, without a neat concentration of force. We have John Wilkin’s unnamed language & Lojban & aUI & Láadan & Ithkuil, but there’s never been a concerted effort at exploring the space—we e.g. haven’t yet tried to create a language for Lullism, deleuzian post-structuralism or the insights from the rationality community. There are reasons to be skeptical about the allure of philosophically powerful languages, and it’s not quite clear what exactly would be looked for here, but that’s all part of the problem statement! If we employ 5 philosophers and 5 linguists at $100k/​year for 10 years, we spend $10 mio, and at worst we get some very interesting speculation.

Art

  • SCP series: The SCP Foundation Wiki is probably the most successful collaborative fiction writing project of the internet age (though probably not the most successful collaborative fiction project ever; any medium-large scale religious mythology has higher longevity, more detail & more consistency than SCP). It has inspired an animated series[3] and some short films, as well as several books. However, despite the clamorings of the community, the creative commons license used for the SCP wiki content (CC-BY-SA-3.0) has hindered the development of a professional adaption—technically using the material for commercial purposes is possible, but the share-alike property demands that the produced works be freely copyable and under a license no more restrictive than the original license. So the production of such a professional series would be, after all, a public good. But that doesn’t concern us, the eccentric trillionaire: We just want some cool SCP content to watch. The SCP universe lends itself to an anthology series like Love, Death & Robots or The Animatrix; e.g. following MTF Omega-7 (“Pandora’s Box”) with Able (showing the uneasy situation the Foundation finds itself in, caught between the Scylla of using SCPs but with potentially disastrous outcomes, and the charybdis of being crushed and consumed by the anomalous world around it), or the Reluctant Dimension Hopper on some of his unfortunate travels, a slice-of-life-with-a-twist episode with a member of one of the Gamers Against Weed, a Pythonesque sketch on someone (unsuccessfully) trying to explain SCP 426 to someone else, a short episode with the content of Revenants, a little exploration of the log of anomalous items in form of an introduction to a new foundation researcher… What budget one would need to be is not quite as clear, on the high end Game of Thrones commanded a budget of >$600 mio. for 73 episodes, and The Lord of the Rings: The Rings of Power has a committed budget of at least $1 bio. for 5 seasons. I haven’t been able to find information on the budget of Love, Death & Robots or The Animatrix. Spending $1 bio. on an SCP series is not unthinkable in this case.

    • Spinoff with Antimemetics Division: Of course I can’t neglect my favorite work of fiction: The novel There Is No Antimemetics Division would lend itself well to a one-season spinoff.

  • Discontinuous & fast AI takeoff movie: Most movies with artificial intelligence feel pretty provincial to me: Most of the AI systems in those movies are simply humans in funny suits, at best they are not bipedal or perhaps distributed over some amount of bipedal bodies, or perhaps wholly without a body but very human in behavior. I do enjoy movies with very non-human aliens—can we have the same for AIs? One way to achieve this would to have a movie about an AI system that crosses a threshold of capabilities extremely quickly and suddenly. Humans die in the first ~2 minutes of the movie (less than a couple of days in-movie time), followed by the AI systems constructing megastructures in space by deconstructing planets, sending probes to all reachable galaxies &c. Interactions with other highly advanced alien civilizations could be shown, including trade (though hard to convey that trade is taking place) and warfare (much easier to convey: nicer visual effects from black hole bombs, induced supernovae and near-c projectiles; but perhaps not as likely to happen in reality because warfare is negative-sum, and we should expect advanced agents to avoid negative sum behavior). Processes within the AI could also be fascinating to visualize: the creation, pursuit and extermination of optimization daemons and the difficulties of trying to align future versions (perhaps put on screen as blurry simulations of successor behavior) should be interesting to look at. If we’re willing to get really weird, one could try to depict acausal trade, multiverse-wide evidential cooperation and later ontokinetics shenanigans—but then one’d need to launch a research project on how these things could possibly play out. One difficulty of making this movie: The strategies of such a system would probably be extremely alien by human standards, and the resulting movie would be quite conceptual as a result—and by Vinge’s law the scriptwriters wouldn’t be able to make it even remotely realistic. But it might be pretty—similar to Koyaanisqatsi, maybe using the Brandenburg Concertos as music. Really expensive movies cost at most ~$400 mio., we could easily top that by spending half a billion on a frivolous nerd project.

  • Double Crux podcast: Double crux is a technique for resolving disagreements developed by the Center for Applied Rationality. In it, (usually) two interlocutors with differing opinions on a subject have a conversation in which they try to understand each others’ beliefs and models of the world, with the goal of finding a single statement (ideally factually checkable or at least eventually resolvable) for which it holds that believes and believes , and if it turned out that then would change their mind on (adopting s position), and vice versa for and . With motivated participants, this tends to produce debates which are more oriented towards finding the truth. Debates on podcasts are often unsatisfactory because participants usually have an adversarial stance towards each other and the time is limite. So a step towards a solution could be to start a new podcast using the double crux framework. The discussions might become exceedingly long (≫10h seem plausible, although that alone needn’t retract as much from viewership as expected), and ideally spiced up with moderation, fact checking, intermittent summarizing of positions after a while &c. Initial attempts at this format seem promising. This is a long shot, as it is unclear how much conversations can be improved, how good debate is at truth-finding, and how entertaining or interesting this would be to listeners. Examples such as the 2021 MIRI conversations have left some participants with a lower opinion of trying to hash out long-standing disagreements. Cost: Probably two full-time equivalents for producer/​interviewer and audio engineering, each at ~$50k per year, and maybe another moderator and a fact checker as a part-time (10h per week on average perhaps?) position, at a total cost of ~$150k per year, fudging upwards to $200k per year.

  • A fictional ethnography and anthropology of native Antarcticans: Antarctica wasn’t settled by humans before Fabian Gottlieb von Bellingshausen and Mikhail Lazarev sighted the Fimbul ice shelf. But what if it had?

    • Finding Antarctica: Humans managed to find and settle Hawaii in ~1250 C.E from the Marquesas islands (distance: 3530km) and the Society islands (distance: 3990km). Similarly, the polynesians that settled Hawaii (probably) also settled New Zealand. The southernmost point of New Zealand is 2466km from the nearest point at the Antarctic coast (the closest Antarctic point from Stewart Island) but if you’re willing to make some hops over the Snares Islands, Auckland Island[4], Macquarie Island and this unnamed icy spot in the Pacific the distance increases to 2770km, but with no hop being greater than 1295km (the one from Macquarie Island to the small island before Antarctica). So it is not inconceivable that a Māori society with more durable ships and warmer clothes could have sailed southwards repeatedly and finally put foot on those icy shores. (Indeed I find it pretty likely that some crazy guys might’ve done this, given the number of islands out in the Pacific that were just discovered (?) by Polynesians sailing around.)[5][6]

    • Settling Antarctica: Actually settling Antarctica is a whole other can of worms: The continent is by far the coldest and (as a polar desert) driest, it harbors nearly no vegetation (except lichen & moss), and away from the coasts there are very ~no large organisms. So we’d like to know, assuming it is possible for there to be native antarctican populations descended from Māori people, how such a population might manage to endure, and what their culture and daily life might look like. Ideas: instead of using wood (since there isn’t any), maybe the native Antarcticans could create a pykrete variant with lichen (“likrete”) or moss (“mokrete”); perhaps (with much effort?) it is possible to domesticate fur seals (leopard seals seem harder), and whaling might still be an option. Antarcticans would probably need to be hunter-gatherers, but the absence of wood makes it quite hard to hunt where the (mostly marine) animals are, similar with spears and bows—is it conceivable to use tools made from ice and stone instead? Heating is a big problem, as well as surviving through the winter, but Antarctica has some accessible coal Merrill 2016, some of it fairly close to the coast where Māori might land. Maybe our native population would need to overwinter in shelters (made from ice, stone or likrete?), surviving on seal meat and blubber. It would also be interesting to have a description of Antarctican culture: What kind of mythology would develop in such a ruthless place? (Inuit mythology is already fascinating, and the very different cultural lineage from polynesian societies would add an interesting twist).

    • Cost: With AI tools it might be surprisingly cheap to produce such a fictional ethnography, paying an anthropologist and a designer $50k/​year each, for a year. Output format: An illustrated book in the style of early 20th century ethnographies, with descriptions of daily life, technology, culture, language &c, similar to The Native Tribes of Central Australia or Report on the Work of the Horn Scientific Expedition.

    • See Also:

  • Prevent House of Leaves movie: House of Leaves is one of my favorite works of fiction. I’m also completely convinced that if anyone ever tried to make a movie out of it, they would completely butcher it and soil the book by sheer association. This scenario must be avoided. So I would buy the rights to the movie for House of Leaves and simply hold them, squatting on the ground others might poison, and in the meantime thinking about what to do when the books copyright runs out (perhaps intervening by subsidizing longevity treatments for Mark Danielewski as soon as they are available). Cost: Unclear. I guess that Danielewski would either be willing to sell the rights to me for ~$50mio., or unwilling to sell the rights to anyone, which would be just as good.

  • Build beautiful geometric objects

    • The platonic solids in physical form, similar to tungsten cubes, but of varying size & material (from a steel icosahedron with a few centimetres vertex length to a granite cube or obsidian dodecahedron twenty metres at each side)—the stuff that would make future people say “they considered themselves a powerful culture”

  • A History of Eve Online

  • Languishing drafts podcast

Science

  • Studying Foreveraloners: Some people are unable to find a partner for their entire life and perhaps never lose their virginity. They give themselves various names: “LoveShy”, “Forever alone” or “involuntary celibate”. A common point of disagreement in those communities is whether lack sexual activity by itself (independently from relationships) has negative effects on people: The incels wiki lists a (likely selective) collection of studies, however nearly all of them are correlational & the ones who aren’t don’t test sexual activity per se (merely physical contact). This offers a nice opportunity for a longitudinal randomized experiment: Find 300 men[7] above the age of 25 who have never had a partner of sexual intercourse (/​r/​ForeverAlone has ~180k subscribers, so the subjects would be 1/​600th of the number of subscribers: seems doable), and randomize them into two groups: For a ten-year period, men in the treatment group see a (paid for) prostitute once a week for ~2 hours, men in the control group don’t (they are, however, not prevented from finding partners for intercourse themselves, and are perhaps rewarded a cash price after ten years that is large enough that they’re still willing to participate in data collection without misreporting). Nobody is forced (by contract or otherwise) to neglect finding a partner, or to sleep with the prostitute, however participants in both groups will receive the cash price in the end if and only if they participate in the data collection. Data collected might be income, educational attainment, relationship status, other sexual activity, mood (via experience sampling), blood pressure and other various health indicators. Cost: $1k per prostitute visit at two visits per week for 150 men for 52 weeks a year for 10 years, which comes out at $156 mio., $200k cash price for participants in the treatment group & $1mio. cash price for members of the control group at (in total) $180mio.[8], ~$100mio. for the data collection (I don’t have a clear idea how much more or less this would need to be), which comes out at ~$436mio. Should one be able to find a similarly sized set of female participants, this doubles the cost to ~$872mio. Setting the experiment up to find participants who are actually celibate and not simply interested in the money or prostitution might pose some difficulty, but not be insurmountable.

  • Breeding superintelligent octopuses or parrots

  • Cloning Extinct Animals In order: Passenger Pigeons, Thylacines, Dodos, Wooly Mammoths, Neanderthals, John von Neumann, Srinivasa Ramanujan…

  • Masturbation research

  • Explorables for causal inference

  • Find and test amnestics: Find drugs that cause short-term memory loss with ideally no side effects (useful for blinding in scientific studies)

  • Run Newcomb’s problem in real life

  • Run the generalized wada test

  • Better wireheading: Achieving reversible (you can remove it and the rat feels/​behaves as before) & non-adaptative (no hedonic adaptation happening) & non-addictive (isn’t sought after when removed) & safe (doesn’t have any serious health consequences) & ability-preserving (cognitive function isn’t impaired, and ideally improved) & anti-slippery (when in the state the being doesn’t seek more of it) long-term pleasure centric wireheading via extensive experiments on animals, either by direct brain stimulation or novel compounds.

    • This one’s pretty ethically questionable

    • Replicate the work at the semalab that uses ultrasound brain stimulation for meditation enhancement.

Metascience

  • The distillation team: A group of people whose sole job it is to talk with smart people whose perfectionism blocks them from writing any of their ideas down and doing it for them

  • The meta-analysis team: A group of people who, whenever I have a question, do a deep dive into the topic & produce a large & in-depth report on the topic (ideally Gwern-style, but I would be willing to make compromises)

  • A roadmap to learn everything: A list of textbooks whose order is optimised for having all dependencies covered. Plus professionally done sets of flashcards for those textbooks.

Other

  • New modes of being: What activity-shaped holes are there in our society? How can we find them? Examples: At some point someone must have invented music. That was a pretty big deal! Maybe we’re missing a bunch of these kinds of things and don’t even notice. Same for meditation.

  • Really Out There Stuff Institute (ROTSI): Global priorities research is a branch of (mostly) philosophy focusing on crucial considerations: statements that, if they were true, would radically change which things are most relevant for global and long-term thinking. This research encompasses some of my favorite parts of philosophy: population axiology, moral uncertainty, problems in decision theory such as Pascal’s mugging (there caled fanaticism) and much more. Separately, a strand of amateur thinking about philosophical problems with a strongly computational bent has emerged from the website LessWrong. This view is very interested in formal epistemology using Solomonoff induction, has broadly mathematical or computational Platonist metaphysics (“all mathematical structures/​possible programs exist”, inspired by the mathematical universe hypothesis (MUH)). I find this strand of thinking fascinating, and would love to see a team of ~20 people or so working on it. The focus of that research would include topics such as exotic decision theories (TDT and its variants & UDT and possible successors), acausal trade, UDASSA, the possible implications of being in simulated universes (such as the option to escape to base-universes), implications of assuming the existence of universes allowing for hypercomputation in the MUH (and different results from assuming the existence of universes higher in the arithmetical and hyperarithmetical hierarchy), as well as problems with the framework (such as the arbitrariness of the Turing machine used) and more. Global priorities research on steroids (being potentially less rigorous or philosophically satisfying, but also willing to take larger steps). Currently, the only organisations doing similar work (that I’m aware of) are the Center for Long-Term Risk and the aforementioned Global Priorities Institute.

    • Cost: Hiring ~20 researchers (evenly split between analytical philosophers, physicists, theoretical computer scientists and economists, plus ~10 (?) supporting staff, each at $100k per year, for ~20 years, giving $60 mio. in total.

    • Output format: Mostly long PDFs with titles such as “Acausalism: A Primer” or “Chaitinudassicon”.


  1. ↩︎

    It’s not clear that Nauru is the best choice here. While it probably is the smallest nation state that can conceivably be bought (I don’t think there is any realistic (or unrealistic) amount of money for which Vatican City could be acquired), it is not very fertile, and has only limited freshwater reserves, relying mostly on rainwater. The highest point is only 71 metres above sea level, which means that a large part of the island might be at risk of going under water with rising sea levels.

  2. ↩︎

    “Yes, I want housing costs to be AS HIGH AS POSSIBLE! MWAHAHAHAHAH!”

  3. ↩︎
  4. ↩︎

    Indeed, there is some evidence that Auckland island was settled briefly by Polynesians 600-700 years ago.

  5. ↩︎

    Maybe I’m lacking in imagination, but this implies both that polynesians can survive for weeks on the open ocean, can reliably find their way back home if need be, and are adventurous enough to just sail out to the open ocean in the hopes of finding new islands. This seems extremely crazy to me.

  6. ↩︎

    Another method of finding and moving to Antarctica would be from the Tierra del Fuego to the Siffrey point, which is much closer (~1030km). I’m not sure whether this is more or less likely: the Yahgan people have lived in the Tierra del Fuego for ~8k years, which would give far more time for for extensive exploration, and the Prime Head is likely warmer and more hospitable than the rest of Antarctica, but I believe that the Polynesians were much better at spending long durations of time at sea, and at finding far away land from subtle cues.

  7. ↩︎

    Since the experiment would solely involve prostitution, my best guess is that it would be significantly more difficult to find a similar number of female participants.

  8. ↩︎

    I’d like to hear feedback on what people believe the right amounts of money for indifference between membership of the two groups+participation would be.