Posts
Comments
It's clearly bad because Altman is launching a coup attempt against the board, making clear that we cannot control dangerous AI even with a nonprofit structure as long as someone has career/profit incentives to build it. Altman should be put down like a rabid dog.
This post is stolen valor. We just ended a 20-year war against some of the most horrible people in the world in 2021. Anyone over age 20 who claims to long for righteous combat who didn't enlist during the Global War on Terrorism (or make a serious attempt and have it denied for reasons fully beyond their control) is a liar. As is almost anyone who makes this claim now and hasn't signed up to be a mercenary in Ukraine (special exception for someone who needs a valid US security clearance and would put it at risk by joining a foreign army). Policy should not cater to their fake preferences.
5% of world pop or GDP means China, India, and the US are the only countries used for the calculation in 12. Which seems questionable.
Also, 11-13 and maybe even 6 and 14 are lagging indicators that seem quite unhelpful in making before-the-fact predictions
The interesting measure would be absolute returns, not risk-adjusted.
Szilard got his key insight precisely by being pissed at Rutherford's hubris. https://blogs.scientificamerican.com/the-curious-wavefunction/leo-szilard-a-traffic-light-and-a-slice-of-nuclear-history/
Choosing the wrong reference class?
A version of this type of situation seems to cover a lot of career decisions made by sufficiently talented people. If you're a young Mark Zuckerberg, should you drop out of Harvard? Dropping out of college is a bad idea, on average. It's not quite as clear cut as OP suggested because you can't reliably replicate what Bill Gates did, but there may be strong indicators that startup founders, or some more specific subclass like startup founders experiencing x% monthly growth in recurring users, and not the average college dropout, is the reference class you should look out. And maybe an experienced startup or VC person could point one to an even better reference class that wouldn't occur to me.
Haha. I didn't really know what was a reasonable amount to save when I started because I had just gotten my first real job and really had no idea how expensive a lifestyle I might want in the future. But I knew I didn't want to be poor ever again. So I set a fairly arbitrary goal, spent a few more years living on the poverty-level income I had had before getting a good job so that I could save while it still had lots and lots of time to grow, and now it's done.
And from a stress/flexibility standpoint I think it was the right decision. I probably don't have to think about saving ever again, except for fun, so if I want to take a job that is funner but pays less, I have absolute freedom to do that.
And it turns out even having money I can live pretty cheap. There were only a few material things I hated about being poor. The constant stress over money was the real problem most of the time. I don't enjoy cooking and the food was boring when I couldn't afford restaurants, so I eat more takeout. And walking 5 miles bc the bus doesn't go where you want kinda sucks, so I take more cabs/ubers/lyfts.
I am far away from retirement so not at 30x yet. But assuming 7% real returns, my projected nest egg is about 108x my current living expenses around the age I want to retire. If something goes wrong before then I can always put more in. Compounding is fucking magic if you start it in your early 20s.
The concept of leverage is not complicated. How it affects volatility drag is, or at least seems so to me when I hear ppl explain it. There is a disconnect between how my bran conceptualizes the abstract percentages vs actually holding an asset.
So, the basic idea for an unleveraged investment is your geometric returns are lower than arithmetic returns because of volatility. E.g. if you have $100, gain 10% one period and lose 5% the next, the arithmetic average return is 2.5% per period, calculated as (10+(-5)/2 but you actually only have $104.5, a return of 2.25% per period, because you are losing 5% of a bigger number than you are gaining 10% on. Easy enough.
But let's say you leverage 2x. Assume no interest to keep it simple. Then this is 20% gain and 10% loss. You have $108. A bigger gain than in the above example, but not 2x as big. Or at least that's what I see articles online saying. But this doesn't make sense to me when I try to conceptualize it as actually holding an asset. Let's say I buy one share of the stock using my own money and one share using a loan. I hold exactly the two shares for the two periods regardless of what the price does, then sell them at the end and pay off the loan. My portfolio is 200, goes to 220 (10% gain), then goes to 209 (5% loss). Then I sell, pay off the loan, and I have $109, not $108. The problem comes if I am not allowed to have a loan too large compared to my assets and have to sell at a bad time. So if the 5% drop happens first, I have $190, of which 100 is borrowed. Have to sell $10 of stock to bring my loan to parity with my own investment. Then I have $180, of which 90 is borrowed, and can only make $18 when the market moves 10% up, instead of the 19 I'd have if I held on to everything. So then my return really is only 8% instead of 9%, because I was forced to maintain constant leverage ratio.
So among ETFs, investing on margin, and futures, which allows me to remain closest to the buy and hold strategy? Or do I face roughly the same constraint no matter what?
Increasing labor income vs investing is not 100% fungible but there are some tradeoffs, especially being self-employed. Any time I spend to learn or manage finance stuff is time I could have spent working. And at least in principle there should be opportunities to spend money to increase my income, but it's a lot more unpredictable--I could advertise, in a non-pandemic environment I could join associations or go to events where I might meet lucrative clients, I could hire lower paid staff and take on clients who it is not worthwhile for me personally to perform services for due to opportunity cost, I could perhaps trade current income for prestige in some aspects of work hoping it will raise my stature and bring more money later, etc.
My sense with leverage is it's more complicated than it looks. My naive intuition was you could match an underlying asset any time futures are available, by holding a total portfolio equal to the value of X shares of the asset, consisting of X futures on the asset and then the balance in cash. Which implies you could beat the asset without additional risk by investing that balance in treasuries.
But there are a few things I don't understand here.
1. I assume treasury futures are more volatile than treasuries, to a sufficient extent that you could not use them to implement this and juice the returns even further. Is that correct?
2. I don't know how futures pricing works. If people know it's possible to do what I am suggesting, will futures prices already be bid up compared to the underlying, erasing the potential gains?
3. Futures don't pay dividends. Am I correct in assuming this is reflected in the price somehow, or is this just a net loss in carrying futures compared to the underlying?
4. I get the basic idea that volatility decay exists and I think I understand it for unleveraged investments but don't really understand how it works with leverage.
Does it happen because people are rebalancing or is it inherent in the use of leverage? If you could let your leverage ratio float a bit instead of selling in response to margin calls, would you then have long run compound returns of (leverage ratio * compound returns of the underlying)? Why or why not? And does maintaining a balance in treasuries or treasury futures actually allow you to avoid margin calls in practice, or are there brokerage restrictions that would prevent it?
The one advantage I do have over the market is more risk tolerance. I don't assume I can beat it on a risk-adjusted basis, but since I disvalue risk less than normal ppl do, beating it in absolute expected value terms is fine, and even EMH says there should be opportunity to get higher returns that way. Will take a look at the alpha architect paper.
I'm interested in the possibility but it doesn't have enough of a history for me to be able to think about it usefully. What is the broader reference class I should be looking at, and what is the evidence on historical returns for that class?
Easily worth it for me. Margin is probably a better deal and should be taken advantage of first. But the idea of being so attached to a particular house that you'd give up a chance to be significantly wealthier just to avoid the risk of foreclosure sounds nuts to me.
The market may go down for 10 years but you're not gonna stay unemployed for 10 years. If you lose your job in a developed country for reasons that are correlated with the market (i.e. not quitting or misconduct) you get unemployment, which is more than enough to live on.
Given the difference between mortgage rates and average market returns, you probably should mortgage your house to invest in the stock market even today.
I think most or all of these rituals are still functional even if participants can't see it. If you want to see that up close, look at how you get the rules waived. If a dispute over a document goes to court, a judge can waive almost any formal defect. But they won't always do it, and someone else, like the clerk, can't. The formalities take the place of human judgment because only a few people are trusted to exercise the appropriate level of judgment and getting it in front of such a person is expensive and time-consuming.
EMH requires liquid markets where you can't cheat. Buying non-US stocks may be putting you on the wrong side of an inefficient market.
uuhhhh most of the middle class health care professionals I know are stoners
Does this study prove it? Maybe not. But if there are confounding effects, industrialization ain't it.
Industrialization clearly doesn't change everything here. British political reform cleanly precedes the industrial revolution; US and French revolutions are post-British IR and may have been affected by it some, but before those countries really industrialized.
I predicted 80% on this particular set of reforms being good but I would have put 60 on the opposite if he had disguised the identity of the revolutionaries. Democracy and freedom are good and the French revolution/Napoleon were generally better at those than the pre-existing institutions. But most revolutions (and wars, and natural disasters) don't budge it either way. The average is a short burst of bloodshed followed by more of what came before.
I highly doubt that wars or disasters are net good on average. Being defeated in WWII was probably good for the axis powers' economies because they replaced totalitarian states with democracies. But I doubt that's the default outcome. Losing a war with Spain, for example, seems to have been very bad for most countries' long-term prospects.
Is it actually uncorrelated? Seems like default would be significantly more likely at the same times that a stock market crash would be, because that's when ppl lose their jobs. Can you do P2P loans to ppl in other countries? That would reduce correlation somewhat.
The returns I've seen from ppl who've done this were in line with the stock market so on a risk-adjusted basis it seems like a bad idea.
IIRC studies have found IQ is more strongly correlated with job performance than most other things people use. BUT there is one exception--tests that mirror the actual work. Which of course is more legally defensible, in addition to being a genuinely better measure. If I were hiring people, I would probably do interviews just to not seem weird to applicants, but all the real selection would be based on performance testing.
A bettor who can make an infinite number of expected profitable bets is going to outperform one who can only make a finite number of bets.
(any number between 1 and 0 exclusive)^infinity=0, i.e. for an infinite series of bets, the probability of ruin with naive EV maximization is 1. So, expected value is actually -1x your bet size.
You're leaving out geometric growth of successive bets. Kelly maximizes expected geometric growth rate. Therefore over enough bets Kelly maximizes expected, i.e. mean, wealth, not merely median wealth.
How far back can I change the institutions? Once you change institutions enough you do change the technology and perhaps the psychology. Something like the Glorious Revolution happening in Rome or any sufficiently centralized ancient state and restoring the Republic on a more sustainable foundation probably results in an early Industrial Revolution. Technology takes off as soon as people see they can benefit from their investment instead of having it all confiscated by an autocratic state, or by bandits if the state is too weak.
Slavery never getting a foothold in the US probably does not change the technological frontier much, but results in the South being as rich and technologically advanced as the North throughout most of US history, in addition to obviously being freer. Also results in more democratic design of federal government. Probably results in earlier elimination of things like malaria and hookworm from the South, and higher GDP per capita and more secure voting rights there today. More tentatively, makes the US a more appealing model for post-colonial states to design their governments around, so maybe more of them develop separation of powers and end up democratic and prosperous, and fewer ever become communist.
I dunno about you but I am basically just a pants-wearing monke. I have an intuitive aversion to the experience machine but I think that is tied to my own experience of life and that my monke brain just does not fully understand the hypothetical being put forward and how radically different it is. In life thus far, and in any version of the scenario that seems plausible to me, there are real hedonic costs, or at least risks of costs, to being deceived. Maybe the machine company steals my organs or forgets to feed me or something while I am deluded and unable to care for myself in the real world. And then either I die, or experience much more pain than I would have. I think if I had perfect certainty that the machine would work as advertised I would take it, but my monke brain thinks it's foolish to have such certainty, and in any real-world situation it is probably right.
I used to think I held truth as a terminal value but now I think that is not even a meaningful thing to state. To quote a famous movie guy, "YOU CAN'T HANDLE THE TRUTH!!!" Or more fundamwntally, you can't even KNOW the truth. You can know a few limited truths, but without a brain enhancement so serious that it's questionable whether the resulting being is even you anymore, that knowledge is quite limited. We make fun of pigeons and goldfish for being dumb but we can't even see all the colors they see. Even your vision of the milk right in front of you is a simulation, and a pale one that leaves out details so basic they can be understood by a pigeon. Even limiting it to knowledge humans can actually know, the body of knowledge is too large for one individual to know it all, so you are limited to knowing just a few things that interest you. While in a simulation, there really is some combination of particles forming the thing you see, so it is "real" in the sense that it is a thing that exists in the universe, even if not the thing your mind thinks it is. There is a sense in which the world outside the machine is more real, but one human's understanding of that world is so limited anyway that I am not sure the warping done by the machine matters, except insofar as it has external consequences.