rossry
Karma: 57
NewTop
How many distinct possible ant!observer-moments are there? What is the entropy of their distribution in the status quo?
How many distinct possible human!observer-moments are there? What is the entropy of their distribution in the status quo?
(Confidence intervals okay; I just have no intuition about these quantities, and you seem to have considered them, so I’m curious what estimates you’re working with.)
My experience has been that in practice it almost always suffices to express second-order knowledge qualitatively rather than quantitatively. Granted, it requires some common context and social trust to be adequately calibrated on “50%, to make up a number” < “50%, just to say a number” < “let’s say 50%” < “something in the ballpark of 50%” < “plausibly 50%” < “probably 50%” < “roughly 50%” < “actually just 50%” < “precisely 50%” (to pick syntax that I’m used to using with people I work with), but you probably don’t actually have good (third-order!) calibration of your second-order knowledge, so why bother with the extra precision?
The only other thing I’ve seen work when you absolutely need to pin down levels of second-order knowledge is just talking about where your uncertainty is coming from, what the gears of your epistemic model are, or sometimes how much time of concerted effort it might take you to resolve X percentage points of uncertainty in expectation.
I’m confused; why don’t you just pick black with probability 100%? (Assuming that your utility is 1 if you make the on-further-reflection-correct choice, 0 else.)
This isn’t the same thing as knowing what you’re going to think at the end of your fretting—you still don’t know—but the correct response to uncertainty is not half speed, and just picking the 40%-to-be-right option all of the time is the expectancy-maximizing response to uncertainty.
Pretty confident they meant it that way:
I am not thinking about physics-time, I am thinking about logical-time. If something is in your past, but has no effect on what algorithm you are running on what observations you get, then it might as well be considered as space-like separated from you. If you compute how everything in the universe evaluates, the space-like separated things are the things that can be evaluated either before or after you, since their output does not change yours or vice-versa. If you partially observe a fact, then I want to say you can decompose that fact into the part that you observed and the part that you didn’t, and say that the part you observed is in your past, while the part you didn’t observe is space-like separated from you.
It seems pretty easy for such mechanisms to be adapted for maximizing reproduction in some ancestral excitement but maladapted for maximizing your preferences in the modern environment.
I think I agree that your point is generally under-considered, especially by the sort of people who compulsively tear down Chesterton’s fences.
RCTs (and p-values) don’t seem to be popular in physics or geology.
What evidence makes you say p-values aren’t popular in physics? My passing (and mainly secondhand) understanding of cosmology is that it uses ~no RCTs but is very wedded to p-values (generally around the 5 sigma level).
A single house in a single market levered up 5x isn’t the good kind of diversification
Hm, I can see your point. Are you saying because you’ve worked through a back-of-the-envelope calculation, or are you just guessing? (Me, I was just guessing.)
and will make your expected portfolio performance worse.
Do you mean “will make the expected dollar-returns of your portfolio worse”, or are you making a claim about the expected utility-adjusted returns?
There is no compelling reason to think such outperformance will continue into the future.
The market disagrees with you.
What market instruments express opinions on the future returns of US equity investments? Everything I can think of serves to express an opinion on present prices of equity investments or on things like the future risk-free rate of return, not the future return on, e.g., the S&P500 index.
What text from the post suggests accounting for the emotional disutility of indebtedness?
Here’s a quote that to me seems to argue against it:
The question becomes: what risk-free rate of return is equivalent to your best available investment? If you pay a higher interest on your debt than that, you should pay it off. If the rate you’re paying is lower, you should invest.
Following this algorithm suggests investing in a risk-free 3.01% return before paying off debts costing a net 3% in interest.
There’s a bit of section 5 that accepts emotional disutility of volatility, but that’s not the same thing.
As for ChristianKl, they comment that to them it doesn’t make sense to value indebtedness at zero, agreeing with you that the disutility should be accounted for.
I think you are in agreement with ChristianKl (i.e., you both think the OP overlooks the emotional disutility of knowing that you have debt), but your tone seems to me to indicate disagreement.
Don’t hold any debt at above 4%.
It depends. If debt is tax deductible then it can make sense to hold it. Or based on the size of the opportunity.
Note that this is in the section of the OP’s advice to themself, and presumably accounts for them already checking for tax-deductible opportunities and opportunity costs.
An anecdote for an anecdote: I made up and plugged in some numbers that roughly corresponded to buying my last (rented) apartment, and the calculator said it’d be cheaper than my rent by ~10%.
Tl;dr if buying vs renting was an obvious win the market would arbitrage that.
As I understand it, lots of the win comes from limit-one-per-person tax advantages, so “the market” can’t straightforwardly arbitrage it away.
In any case, along the lines of the OP’s section 4 above, the typical person should probably, ceteris paribus, prefer some investment in some real estate over additional investment in whatever market portfolio they already have, because not-perfectly-correlated assets add expectations but less-than-add volatilities. I think the calculator you linked is getting this wrong.
Also not sure if it’s a standard concept in path search, but searching from both ends seems fruitful, even in its most naive implementations:
if the graph is sufficiently random, I think it should take the expected fraction of the maze you have to walk from N/2 to sqrt(N)
you can apply it to either BFS or DFS, though I suspect that in practice DFS works better (and even better still if you DFS by switching your first decision (rather than the OP’s algorithm of switching your last)
naturally, there’s plenty of room for layering heuristics on top, like seeking towards the opposing cursor, or even the nearest point of the opposing explored-so-far surface…
Answering my own musing: it’s implementable as a 2-state cellular automaton, rather than requiring the ~5 states that distributed DFS does. So there’s that.
(Side question: there’s at least one more-human-intuitive way to apply chunking to mazes. Can you figure it out?)
What comes to my mind is a kind of algorithm for auto-chunking: repeatedly take a dead end and fill it in until the point where it branches off from a hallway or intersection. Eventually you’re left with a single hallway from entrance to exit.
I’m not certain how it’s an improvement over DFS, though.
Notes on a recent wave of spam
Oh, I certainly didn’t mean to imply that there weren’t cases of suitable replacements for slurs (or that it wouldn’t be valuable to find such); rather, I only meant to claim that there existed a case where it isn’t obvious how to find a suitable replacement (contra jimrandomh above).
Does anyone else have this experience of inversion? Is there already a name for it?
It sounds to me like something akin to the “figure-ground inversion” that Scott Alexander moots in a review of House of God:
House of God does a weird form of figure-ground inversion.
An example of what I mean, taken from politics: some people think of government as another name for the things we do together, like providing food to the hungry, or ensuring that old people have the health care they need. These people know that some politicians are corrupt, and sometimes the money actually goes to whoever’s best at demanding pork, and the regulations sometimes favor whichever giant corporation has the best lobbyists. But this is viewed as a weird disease of the body politic, something that can be abstracted away as noise in the system.
And then there are other people who think of government as a giant pork-distribution system, where obviously representatives and bureaucrats, incentivized in every way to support the forces that provide them with campaign funding and personal prestige, will take those incentives. Obviously they’ll use the government to crush their enemies. Sometimes this system also involves the hungry getting food and the elderly getting medical care, as an epiphenomenon of its pork-distribution role, but this isn’t particularly important and can be abstracted away as noise.
I think I can go back and forth between these two models when I need to, but it’s a weird switch of perspective, where the parts you view as noise in one model resolve into the essence of the other and vice versa.
1) My model of people who use slurs as a significant part of their expressive vocabulary is that at least some of them use the slur to mean “member of group [X] I don’t like”, as explicitly opposed to “member of group [X] I feel indifferent-to-positive about”. A neutral group signifier plus optional insult alone fails to encode this distinction, perhaps making it a less-than-suitable replacement.
2) I read:
I think people are within their rights to reject a proposed replacement for not meaning the right thing, sounding ugly, being one syllable longer, being hard to spell, not rhyming in a poem they’re trying to write, and vague gut feeling that you’re just trying to control them.
...to be pretty clear about whether a more-verbose construction that requires the speaker to separate their personal insult can fail to be a suitable replacement.
None of this means that the general principle can’t or shouldn’t have a carve-out for slurs; My only intended argument is that, as expressed above, it seems plausible to me that finding suitable replacements requires significant effort (and basically is never done in practice by people attempting to remove slurs from others’ excessive vocabulary).
Right, so if we’re using a uniform distribution over 2^30000, there should be exactly zero ants sharing observer-moments, so in order to argue that ants’ overlap in observer-moments should discount their total weight, we’re going to need to squeeze that space a lot harder than that.
I’ve also spent some time recently staring at ~randomly generated grids of color for an unrelated project, and I think there’s basically no way that the human visual system is getting so much as 5000 bits of entropy (i.e., 50x50 grid of four-color choices) out of the observer-experience of the visual field. So I think using 2^#receptors is just the wrong starting point. Similarly, assuming that neurons operate independently is going to give you a number in entirely the wrong realm of numbers entirely. (Wikipedia says an ant has ~250,000 neurons.)
I think that if you want to get to the belief that two ants might ever actually share an experience, you’re going to need to work in a significantly smaller domain, like your suggestion of output actions, though applying the domain of “typical reactions of a human to new objects” is going to grossly undercount the number of human possible observer-experiences, so now I’m back to being stuck wondering how to do that at all.