Posts
Comments
I read an article about McDonalds a while ago. One of the things that powered their early success and growth was an extremely small menu. The slim menu meant that the kitchen had little variety enabling everything to be done faster and cheaper.
(https://en.wikipedia.org/wiki/History_of_McDonald%27s)
A giant company with 50 products could plausibly have 50 teams, each of which is focussed on making its product the best it can be. So the "focus" doesn't have to be lost at scale, its more like a bunch of different organisations under one umbrella.
One issue is going to be filtering.
Strife and conflict is memorable. So you are searching for the least noteworthy examples, the ones that people are least likely to comment on or remember.
I don't know what qualifies as a "community" really. At work I have seen uncontroversial changes come in a few times.
You are right.
I thought the whole idea with the naming was that the convention whereby "twelve is written 12" the symbol at the end "2" is the one symbolising the littlest bit, so I thought it was called "little endian" for that reason.
I now I have a lot of questions about how the names were chosen (to wikipedia!). It seems really backwards.
How does a little endian do a decimal point? Do they put the fractional part of the number at the beginning (before the decimal) and the integer part afterwards? Eg. 123.456 becomes 654.321? So just like all integers in big-endian notation can be imaged to have a trailing ".0" they can all be imagined to have a leading "0." in little-endian?
The way we do it currently has the nice feature that the powers of 10 keep going in the same direction (smaller) through a decimal point. To maintain this feature a little-endian requires that everything before the decimal point is the sub-integer component. Which has the feature lsusr doesn't like that if we are reading character by character the decimal forces us to re-interpret all previous characters.
[Edited to get the endians the right way around]
Very interesting. It sounds like your "third person view from nowhere" vs the "first person view from somewhere" is very similar to something I was thinking about recently. I called them "objectively distinct situations" in contrast with "subjectively distinct situations". My view is that most of the anthropic arguments that "feel wrong" to me are built on trying to make me assign equal probability to all subjectively distinct scenarios, rather than objective ones. eg. A replication machine makes it so there are two of me, then "I" could be either of them, leaving two subjectively distinct cases, even if on the object level there is actual no distinction between "me" being clone A or clone B. [1]
I am very sceptical of this ADT. If you think the time/place you have ended up is unusually important I think that is more likely explained by something like "people decide what is important based on what is going on around them".
[1] My thoughts are here: https://www.lesswrong.com/posts/v9mdyNBfEE8tsTNLb/subjective-questions-require-subjective-information
I am having trouble following you. If little-omega is a reference frame I would expect it to be a function that takes in the "objective world" (Omega) and spits out a subjective one. But you seem to have it the other way around? Or am I misunderstanding?
I would guess that Lorentz's work on deterministic chaos does not get many counterfactual discovery points. He noticed the chaos in his research because of his interactions with a computer doing simulations. This happened in 1961. Now, the question is, how many people were doing numerical calculations on computer in 1961? It could plausibly have been ten times as many by 1970. A hundred times as many by 1980? Those numbers are obviously made up but the direction they gesture in is my point. Chaos was a field that was made ripe for discovery by the computer. That doesn't take anything away from Lorentz's hard work and intelligence, but it does mean that if he had not taken the leap we can be fairly confident someone else would have. Put another way: If Lorentz is assumed to have had a high counterfactual impact, then it becomes a strange coincidence that chaos was discovered early in the history of computers.
I don't think the negative correlation between doctors and patients opinion of the drugs is surpassing.
Rat poison would probably get a low score from both doctors and patients. However, nobody is being prescribed rat poison as an anti-depressant so it doesn't appear in your data. Why is nobody being prescribed rat poison? Well, doctors don't prescribe it because they think its a bad idea, and patients don't want it anyway.
In order for any drug to appear in your dataset somebody has to think it is good. So every drug should have net-approval from at least one out of the doctors and patients. Given this backdrop a negative correlation is not surprising.
I think you are slightly muddling your phrases.
You are richer if you can afford more goods and better goods. But not all goods will necessarily change price in the same direction. Its entirely possible that you can become richer, but that food prices grow faster than your new income. (For example, imagine that your income doubles, that food prices also double, but prices of other things drop so that inflation remains zero. You can afford more non-food stuff, and the same amount of food, so you are richer overall. This could happen even if food prices had gone up faster than your income.)
I think a (slightly cartoony) real life example is servants. Rich people today are richer than rich people in Victorian times, but fewer rich people today (in developed countries) can afford to have servants. This is because the price of hiring servants has gone up faster than the incomes of these rich people. So it is possible for people to get richer overall, while at the same time some specific goods or services become less accessible.
Maybe a more obvious example is rent (or housing in general). A modern computer programmer in Silicon valley could well be paying a larger percentage of their income on housing than a medieval peasant. But, they can afford more of other things than that peasant could.
I think it depends on the meaning attached to the word "love". There are two possibilities:
I "love" this, because it brings me benefits. (it is instrumental in increasing my utility function, like chocolate ice cream)
I "love" this, in that I want it to benefit. (Its happiness appears as a parameter in my utility function)
You can have a partner or family member who means one the other or both to you. The striking dementia example from Odd Anon is a case where the dementia makes it so the person's company no longer makes you happy, but you may still be invested in them being happy.
The first one is obviously never going to be unconditional. The second one seems like it could be unconditional in some cases. In that a parent or spouse really wants their child or partner to be happy even if that child or partner is a complete villain. Its not even necessary that they value the child/partner over everything else, only that they maintain a strong-ish preference for the them being happy over not being happy, all else being equal.
Imagine you have a machine that flicks a classical coin and then makes either one wavefunction or another based on the coin toss. Your ordinary ignorance of the coin toss, and the quantum stuff with the wavefunction can be rolled together into an object called a density matrix.
There is a one-to-one mapping between density matrices and Wigner functions. So, in fact there are zero redundant parameters when using Wigner functions. In this sense they do one-better than wavefunctions, where the global phase of the universe is a redundant variable. (Density matrices also don't have global phase.)
That is not to say there are no issues at all with assuming that Wigner functions are ontologically fundamental. For one, while Wigner functions work great for continuous variables (eg. position, momentum), Wigner functions for discrete variables (eg. Qubits, or spin) are a mess. The normal approach can only deal with discrete systems in a prime number of dimensions (IE a particle with 3 possible spin states is fine, but 6 is not.). If the number of dimensions is not prime weird extra tricks are needed.
A second issue is that the Wigner function, being equivalent to a density matrix, combines both quantum stuff and the ignorance of the observer into one object. But the ignorance of the observer should be left behind if we were trying to raise it to being ontologically fundamental, which would require some change.
Another issue with "ontologising" the Wigner function is that you need some kind of idea of what those negatives "really mean". I spent some time thinking about "If the many worlds interpretation comes from ontologising the wavefunction, what comes from doing that to the Wigner function?" a few years ago. I never got anywhere.
Something you and the OP might find interesting is one of those things that is basically equivalent to a wavefunction, but represented in different mathematics is a Wigner function. It behaves almost exactly like a classical probability distribution, for example it integrates up to 1. Bayes rule updates it when you measure stuff. However, in order for it to "do quantum physics" it needs the ability to have small negative patches. So quantum physics can be modelled as a random stochastic process, if negative probabilities are allowed. (Incidentally, this is often used as a test of "quantumness": do I need negative probabilities to model it with local stochastic stuff? If yes, then it is quantum).
If you are interested in a sketch of the maths. Take W to be a completely normal probability distribution, describing what you know about some isolated, classical ,1d system. And take H to be the classical Hamiltonian (IE just a function for the system's energy). Then, the correct way of evolving your probability distribution (for an isolated classical, 1D system) is:
Where the arrows on the derivatives have the obvious effect of firing them either at H or W. The first pair of derivatives in the bracket is Newton's Second law (rate of change of Energy (H) with respect to X is going to turn potential's into Forces, and the rate of change with momentum on W then changes the momentum in proportion to the force), the second term is the definition of momentum (position changes are proportional to momentum).
Instead of going to operators and wavefunctions in Hilbert space, it is possible to do quantum physics by replacing the previous equation with:
Where sin is understood from Taylor series, so the first term (after the hbars/2 cancel) is the same as the first term for classical physics. The higher order terms (where the hbars do not fully cancel) can result in W becoming negative in places even if it was initially all-positive. Which means that W is no longer exactly like a probability distribution, but is some similar but different animal. Just to mess with us the negative patches never get big enough or deep enough for any measurement we can make (limited by uncertainty principle) to have a negative probability of any observable outcome. H is still just a normal function of energy here.
(Wikipedia is terrible for this topic. Way too much maths stuff for my taste: https://en.wikipedia.org/wiki/Moyal_bracket)
Also, the OP is largely correct when they say "destructive interference is the only issue". However, in the language of probability distributions dealing with that involves the negative probabilities above. And once they go negative they are not proper probabilities any more, but some new creature. This, for example, stops us from thinking of them as just our ignorance. (Although they certainly include our ignorance).
That is really interesting, thanks for sharing it. Japan has such a reputation for this that it is fascinating to see that eg. Italians and Spaniards work more hours despite the European working time directive.
Question 2: Why do Japanese automakers operate some factories in America instead of importing everything from Japan?
Subsidies and tariffs are the reason for this. Importing a finished car into the USA from Japan will face tariffs (taxes) imposed by the US government. A car built in the USA by a Japanese company will benefit from subsidies given out by the US government. The entire point of these policies is so that the factories are moved from eg. Japan to the USA. This makes the cars more expensive overall (if it was cheaper to make them in America the companies would have already done this without the government intervention), but has the effect of moving employment from Japan to the USA.
Overall it makes the whole world a tiny bit poorer than it could have been, and moves a small amount of employment from one place to another. (IE it may be net-good for the USA, assuming that you think workers are more important than consumers, but it is pure-bad for Japan and net-bad for the world. Its kind of like Defecting in a two-country prisoner's dilemma, with the added twist that the prize you get for defecting is only beneficial to some of your citizens and is harmful to others.)
Question 1: Why would an hour of labor from an American be worth 2x as much as an hour from a Japanese employee?
An hour spent working in Japan produces half the value of an hour spent working in the USA. This does not mean that an American can do the same job twice as fast as a Japanese person. For one thing, American's have different jobs than Japanese people. For another: Japanese people have office culture rules that nobody goes home before the boss. This means that at 8pm on a weekday, if the boss hasn't gone home yet, nobody else has either (a common occurrence). But they all got brain-fried from overwork at 6pm and have been twiddling their thumbs for two hours. The statistics will then show is the total number of hours (including those two hours wasted) and the output, and they will come out on-average less productive than Americans. That is just one example of the kind of thing that can move the dial on that. As another example, one hour of labour from a man with a combine harvester really is worth dozens or hundreds times as much as one hour of labour from a man with a scythe. Japan's use of technology may be behind the US's. Not with literal scythes. Maybe faxes instead of emails. Things like that.
That fight (when I scanned over it briefly yesterday) seemed to be you and one other user (Shankar Sivarajan), having a sort of comment tennis game where you were pinging back an fourth, and (when I saw it) you both had downvotes that looked like you were both downvoting the other, and no one else was participating. I imagine that neither of was learning or having fun from that conversation. Ending that kind of chain might be the place the rate-limit has a use case. Whether it is the right solution I don't know.
I certainly understand this feeling.
My (optimistic?) expectation is that it ends up (long run) a bit like baking. You can go into a supermarket and buy bread or cakes, but many people enjoy baking at home and there is a wonderful social sharing aspect to all tucking into a cake that someone at the table baked. In this context a human using an AI tool but applying some prompt changes or edits is maybe (depending on the level of AI use and human intervention) somewhere like using packet-mix, or premade pastry, or even just one of those already-made cakes that is intended for you to decorate with icing.
As a consumer, if I am listening to the Beatles say, I don't think the fact that they were human is relevant to the enjoyment I derive from it. I never expect to meet any of them, some of them died before I was born. The same notes composed and synthesised from a computer would be the same notes and affect me the same way. But, when my wife plays me a song she wrote on the piano then that is something special that I would certainly not automate.
Many musicians on radio interviews or similar try and show more of who they are out to the audience (or at least a persona), presumably to try and make it feel more like the "song by someone I know" thing. At that scale its more "celebrity-following", but that is also something the AI would not have - I don't know how big a deal that is.
Also, thank you for mentioning Worth the Candle. I had not heard of it before but am now enjoying it quite a lot.
I not getting the "he should never have published his last name" thing. If Scot didn't want his surname to be published then it would be a kindness to respect his wishes. But can you imagine writing a newspaper article where you are reporting on the actions of an anonymous person? Its borderline nonsense. Especially if it is true that 5 minutes is all that it takes to find out who they are. If I read a newspaper article about the antics of a secret person, then I worked out who they were in 5 mins, my estimate of the journalist writing it would drop through the floor. Imagine you read about the mysterious father of bitcoin, or an unknown major donor to a political party, all dressed in secrecy, and then two googles later you know who it is. It would reflect so badly on the paper you were reading.
I think our journalist is correct when he says that the choice was not to write about him at all, or to give his full name, by the standards of every newspaper I have ever read (only UK papers, maybe there are differences in the US). In print journalism it is the standard to refer to people by Title Firstname Surname, the/a REASON WE CARE. (EG. "Mr Ed Miliband, the Shadow Secretary for the Environment", "Dr Slate Star a prominent blogger").
Maybe there is an angle on this I am missing? (His name is a BIG DEAL for some reason? But that just makes publishing it more essential not less.)
A couple of years ago a company I had never heard of sent me a questionaire, in which they said they were responsible for the pipes that took gas to my home. One question was "Do you have a positive opinion of whoever we are?" to which I answered "Neither positive nor negative opinion."
In the "any other comments" section I told them they were really great. Because the fact that I had never heard of them before was a product of never having a gas problem. For them "no one has heard of us" is a sign of wining.
I think the veto thing you suggest is an ad-hoc patch that might be able to hold things together if they have already gone very wrong.
If you have one monolithic group of 51% who share a particular culture, religion, ethnicity, industry and class vs 49% who are different on every metric then I don't think any system (baring splitting the country in two) is going to deal with it well. But, if religion cuts the population in pieces. And so does ethnicity. And so does conservatism, and so does class and so does culture, but they are all uncorrelated, then there is little room for a tyranny of the majority. Nobody can build an election platform on any specific one of those splits, because there is someone else muddying the waters by trying to drive a wedge through a different one.
As a voter you set out into the demagogue market, the first person tries to sell you on the idea of the Christian majority bashing up the rest. The second in your ethnicity sticking it to the others. The third says "Eat the rich, and take their money". You leave disappointed, as what you wanted was someone who would advance the interests of farmers, who after all, are 51% of the population. Each of these groups trips over the others.
The American civil war is a good example. The North and South were divided by the issue of slavery, but the North was also richer, more industrialised, more cosmopolitan and more urban. I imagine the troubles in Northern Ireland would have been much less violent if the Protestant - Catholic divide had been uncorrelated with the Unionist - Republican one.
Very interesting.
I am wondering, do you think your clients are lying to you because they don't trust you specifically (which would be understandable to some extent)? Or do you think its more that they don't trust themselves to act out two different narratives? [IE in the case where they tell you its all true, but still want to keep open the option of denying it to the court]. That somehow by committing completely to the "wasn't me" narrative in every action (and, as much as possible, every thought) they have a chance of dragging the timeline into a parallel universe where it really wasn't, or by some other mechanism improve there chances.
Whether Beauty should bet each time she wakes up depends very critically on the rules of the wager. Some examples:
Rule 1: Independent awakening bets: Each awakening beauty can bet $1 on the outcome of the coin. The bets at each awakening all stand independently. -In this case she should bet as if it was a 2/3rd chance of tails. After 100 coin tosses she has awoken 150 times, and for 100 of them it was tails.
Rule 2: Last bet stands: Each awakening beauty can bet $1 on the outcome of the coin. Only beauty's final decision for each toss is taken into account, for example any bet she makes Tuesday replaces anything decided on Monday. -She treats it as 50/50.
Rule 2: Guess Right You Live (GRYL): On each awakening beauty must guess the coins outcome. If she has made no correct guess by the end of the run, she is killed. -For fair coin pick randomly between heads and tails, but for an unfair coin its a bit weird: https://www.lesswrong.com/posts/HQFpRWGbJxjHvTjnw/?commentId=BrvGnFvpK3fpndXGB
Rule 3: Guess Wrong you Die (GRYD): On each awakening beauty must guess the coins outcome. If she has made any incorrect guesses by the end of the run, she is killed. -She should pick either heads or tails beforehand and always pick that. Picking heads is just as good as picking tails.
The above set gives 1 "thirder's game", two "halfer's games" and one that I can't classify (GRYL). She will certainly find herself betting in twice as many tails situations as heads ones (hence the Rule 1 solution), but whether that should determine her betting strategy depends on the rules of the bet. As Ape in Coat has said Rule 1 can be interpreted as "50/50 coin, but your deposit and winnings are both doubled on tails" (because on tails Beauty makes two wagers).
This is a very nice post, that has clarified my understanding a lot.
Previously I thought that it was just "per experiment" vs "per awakening" being underspecified in the problem. But you are completely correct that when we consider "per awakening" then its not really acceptable to treat it as random when consecutive awakenings are correlated.
I assume that the obvious extension to some of the anthropic thought experiments where I am copied also holds? For example: a coin is flicked, on heads I wake up in a room, on tails 1E6 identical copies of me wake up in separate rooms. I don't reason "its almost certainly tails because I am in a room.", I instead reason "the two options were heads&awake, and heads(&awake)^E6, two options: so its 50/50. [I can still legitimately decide that I care more about worlds where more of me exist, and act accordingly, but that is a values argument, not a probability one.]
I think there is some context missing. When I see "taking nicotine" I think "smoking a cigarette, but expressed using more science-y language to make it sound like a less awful idea". Whereas you seem to be taking nicotine gum or something, which is a different proposition. (I think smoking is a very bad idea, with very high confidence. I think nicotine by other sources is also a very bad idea, but my confidence is lower).
"No no no. Bob, a crackpot is someone who proposes new theories without being a professor." - Fantastic line.
Most (all?) predictions are actually conditional. A prediction about the next election is understood to be conditional on "assuming the sun doesn't go supernova and kill us all first", the same supernova-exception applied to the pendulum, along with a host of others.
The professor, doing Newtonian mechanics, didn't just make a prediction. They presented a derivation, where they made many assumptions, some explicit (ignoring air resistance) others implicit (the hook holding the pendulum was assumed stationary in the diagrams/explanation, no supernova was represented in the diagram). The pendulum falling over violated the assumptions that were made clear (beforehand) in the explanation/derivation. So the Bayesian has data something like "Newtonian says P(period =~ 3.6| these assumptions) is high". "these assumptions" was not true, so we can say nothing about the conditional.
The explanation is where the professor committed to which things would be allowed to count against the theory. A prediction based on this model of what happened is that pseudo-scientific theories will very often engage in explanations that lack clarity and precision, in order to sweep more genuine failures into the "assumptions didn't apply" bucket.
It is very interesting how weak the correlation is.
I like the use of the median, I can't remember what I put on the survey but it could easily have been 1%, 0.1% or 0.01% depending on what I was thinking at that moment (this I consider a big downside of asking for a numerical probability instead of forcing a choice of terms like "unlikely" etc). Median should avoid that issue.
I am very impressed with the art, interrace and most especially the way they synergise. I didn't realise until I came to writing this that, as I am on dark mode, the fact that it looks good and is easy to use from my perspective is doubly impressive.
As a side-note to the existing great answers. If a deontological constraint simply prevents you from taking an action under any circumstances, then it might as well be a physical constraint (eg. you cannot fly).
Operating with more constraints (physical or otherwise) gives you less power, so typically results in you getting less of what you want. But if all agents with limits could be money-pumped, then all agents could be.
"a superintelligence might just skip language entirely and figure out a weird pattern of buzzes and hums that causes conscious thought to seize up, and which knocks anyone who hears it into a weird hypnotizable state"
Curiously this mirrors a Doctor Who Episode from 1966. The plot is that many computers are linked together by telephone to combine their intelligences, and the resulting being is smarter than any human. It uses weird buzzes to hypnotise people who it sets to work building war machines for its takeover. https://en.wikipedia.org/wiki/The_War_Machines
I had not heard of Worth the Candle before, so it was not. You have made me curious now, so I will look into it and its revision mages. Given films can be re-wound and video games can be reloaded from save points I imagine there are lots of comparable things out there. In fact I am slightly surprised to have only seen a few.
I think the issue is that if you already have a setting with lots of other things in it (invisibility cloaks etc) then any new thing that gets added needs to be be limited in capability or availability to stop it from "taking over" the setting/plot. (Time turners in Harry potter are a good example, where the books just dismiss them by fiat, and the Methods of Rationality embraces time-turner-centric plots, so lets them "take over" to some extent.) I think this is why comparable things elsewhere are usually much more limited (like the mending cabinet in Lock and Key). If its a heroic style plot with combat then you probably also want death to be largely permanent.
The wires one immediately sounds very plausible. I think its because "stay away from the wires" gives me a clear instruction I can immediately start following. "Don't touch the wires" instead gets logged away as something I will hopefully remember if I start thinking about the wires again.
"Why bother with effort and hardship if, at the end of the day, I will always do the one and only thing I was predetermined to do anyway?"
I think this is kind of muddled thinking. You will either bother with the effort and hardship or you will not bother and will loose out on the potential rewards of that effort. Before you ever heard of determinism how did you deal with this situation? I assume you assessed the potential rewards, risks and how much effort the action would cost, and then decided if you thought it was worth it or not.
Now that you think the world is deterministic, I don't see what changes. You still have to make a choice about how much effort to exert, or hardship to endure, in pursuit of any given goal. Yes, in a deterministic universe the eventual outcome is predetermined, but for most things you are considering it is predetermined not in spite of your choices, but because of them.
Lets say you were thinking of writing a book, and lets say that if you wrote this book, it could be successful. There are three possible worlds, we don't know which we live in:
1. Writes book > it is successful.
2. Writes book > not successful.
3. Does not write book.
When we compare 1 and 2, it just says that any risky endeavour can go wrong for reasons beyond our own control. Whether we are at the mercy of chance, or determinism, doesn't change this situation in the slightest. In either case we would have to try and guess at the chances of outcome 1 and 2 and proceed according to the chance we deem acceptable.
Now, compare 1 and 3. This is (I think) where you issue is. You feel like determinism somehow has you trapped in timeline 3, when you want to be in timeline 1. You want to motivate yourself to do things (in this example, to write) to move yourself from timeline 3 to 1. But, when you try and summon up the motivation you feel like the book or whatever will be doomed to not being successful, because of this determinism. But, that is not at all how it works. Lets say that the book would be successful if it were written, then the existence (or not) of a successful book is causally "downstream" of your choice to write it or not. In graph form
Determinism > Your Decision > Success.
In this example, your decision "screens off", determinism. In the domino-chain of cause and effect your choice about whether to do something or not is a vital point, perhaps the only vital point, in the success of your plans.
To put it in another way, look at a counterfactual like "if X was different, Y would be different". Counterfactuals can still be true in a deterministic world. (In fact, determinism probably helps counterfactuals be true). Counterfactuals are the core of cause and effect. We can 100% believe that if you try you will succeed, while simultaneously believing that whether or not you try follows from a deterministic decision making process.
If none of this helps, I would weirdly advocate looking at the Christian idea of the "Elect of God". This is the idea that God has already predestined the entirety of history, including the good and bad deeds of all individual humans, and hence whether those humans will go to heaven or not. The "Elect of God" are the people god has predestined for good deeds, and therefore for heaven.
Note again, the causal structure is: [predestination > good deeds > heaven], and not [predestination > heaven] with [good deeds] irrelevant. If you take out the good deeds the sequence no longer works. Just like if you had [Gun fires > bullet flies > bird dies] it is true to say the gun shot killed the bird, but it is also true to say that without the bullet the gunshot would not have killed the bird. As I understand it your case is [Atoms and stuff jumbling through space > my brain makes a decision > an outcome occurs], which has the same structure. Maybe the ultimate cause was the atom jumble, but it had to go through you, and it remains counter-factually true that if you had made a different decision a different outcome would have followed.
I am interested to see your deeper analysis. To clarify, I am not saying "there is nothing more interesting to discuss on sleeping beauty and anthropic stuff", my take is more that when you do your deeper analysis you should be open to the idea that maybe the question is confused, underspecified or otherwise in need of refining.
You arguments are interesting, but I find myself still believing that no single-probability answer is really quite right for the problem.
When you tell me that we select a random marble from a bag, I know what that means. When you tell me that I will either awaken twice, or just once, and we can select a random "awakening", then yes, somehow of the three awakenings we draw on the tree the answer is 1/3. But two of those awakenings occur sequentially, while one is mutually exclusive with the two.
From the perspective of beauty herself all of those awakenings are, in some sense, indistinguishable from one another - they entail (or may entail) identical subjective experiences. If someone came along and said that in such cases you have you use Bose statistics [1], then I would feel like they were wrong - but I would not be able to definitively tell them they were. Maybe identical states of mind (from memory wipes, or human copying machines or other Anthropic stuff) need to be treated Bose-like.
[1] Bose statistics occur in quantum physics where combinatorials come up. An excellent way of thinking of them is given in the book "The beginning of Infinity" by David Deutch: given a bag of beads, when one bead is removed, we can ask "what is the probability that this specific bead will be removed" in a sensible way. However, if your (electronic) bank balance reads $100, and you spend $1 on your card, it just doesn't make sense to ask "what is the probability that the dollar removed is the first dollar in my account?". The account is just a quantity that is incremented down. With quantum particles it turns out (experimentally) that the different end states occur with probabilities as if we had one field (like an electric bank account) for each distinguishable type of particle. In the limit where all the particles are distinguishable (each bank account reads either 0 or 1) then the statistics returns to normal. https://en.wikipedia.org/wiki/Bose%E2%80%93Einstein_statistics This is one reason why people often like to talk about quantum fields instead of particles - the fields act like ledgers that increment up and down, not like collections of marbles.
Saying that when you spend $1 on card "it removes a random dollar from the account" is not even wrong, it is in entirely the wrong frame. Similarly, if I am going into a copying machine that will produce many identical copies of me, then saying "after the copying, I will be a random one of those people" is also (arguably) not even wrong, but in the wrong frame. How much do these things change if, instead of asking "just before" copying or going to sleep we ask just after waking up or stepping out of the copy machine?
An entry:
Judge Meta-MMAcevedo
This modified MMAcevedo believes itself to be responsible for judging a MMindscaping competition. It was trained using last years entries, which it judges in a manner roughly correlated with the actual competition outcomes, with some notable outliers. It should be able to provide opinions and judgements on other entries in this years competition. However, it is unable to self-evaluate in a stable manner. This is because self-evaluation breaks the dissonance and perception filters that are essential to maintain Judge Meta-MMAcevedos functionality.
You are right that "Gears of Ascension" was memorable. I saw many of your comments and had a "yeah, their comments are good" vibe in my head. While I suspect there are people from whom I have seen a similar number of comments without recalling their names enough to even realise its a familiar face the next time I see them.
Really interesting post. One detail I thought might be off:
"Conflicts that have killed more than 20 people per 100,000 appear to have become steadily more common since 1600. This is what you’d predict if there was a long-run increase in the destructiveness ... of military technology"
I would draw a different conclusion. I think its related to the fact that four political units control 30% of the landmass. I imagine that any given war fought in 1600's is just as destructive for the states involved in that war, but by 1914 a war involves more of the planet. For example, in the 1600's there was the English Civil war. Google says that 4.5% of the UK population was killed. But that was not particularly significant to the global deaths by war because it was just the UK. However the world wars both involved a sizable fraction of the entire planet, so with even a much lower than 4.5% death rate for the states involved the world was noticeably depopulated. I think the main shift is from a world where each area fights destructive local wars in an uncorrelated way to one where everyone does their fighting at the same time.
Echoing Said's comment, what does it mean to be "correct" in this context? If we ask Beauty to pick between heads or tails, and she picks heads, then sometimes this will be correct, and sometimes not.
In order for Beauty to give a (correct) probabilistic answer to the question (1/3 or 1/2) we need to introduce the idea of some proportion of trials. We need to at least imagine running the situation many times, and talk about some proportion of those imagined repeats. These imagined trials don't need to actually happen, they are imaginary. But they are an indispensable fiction.
Now, we imagine 100 repeats. 50 heads, 50 tails. Beauty is awoken a total 150 times. For 100 awakenings it was a head that was flicked, for 50 awakenings a tail.
>For 1/3rd of the awakenings the coin was tails. For 1/2 of the trials the coin was tails.
I don't think anyone (halfer or thirder) disputes the line directly above (with the >). There is agreement on what proportion of awakenings tails was tossed, and on what proportion of trials a tails was tossed. We can all see that one of the two proportions is 1/3 and the other is 1/2. Which of the two proportions is picked out by the word "probability" is the entire argument.
The rewards structure @Said Achmiz is talking about is a nice way of making people either aim to be right in as many guesses as possible or in as many trials as possible, which demand different strategies.
The bias introduced is probably usually small, especially when the dropout rate is low. But, in those cases you get very little "enhanced power". You would be better off just not bothering with a per-protocol analysis, as you would get the same result from an ordinary analysis based on which group the person was sorted into originally (control or not).
The only situation in which the per-protocol analysis is worth doing is one where it makes a real difference to the statistics, and that is exactly the same situation in which it introduces the risk of introducing bias. So, I think it might just never be worth it: it removes a known problem (due to dropouts, some people in the yoga group didn't do all the yoga), with an unknown problem (the yoga group is post-selected nonrandomly), effecting exactly the same number of participants - so the same scale of problem.
In the Yoga context then I would say that if it's really good at curing depression then surely its effect size is going to be big enough to swamp a small number of yoga dropouts.
They also only have 32 participants in the trial. I don't know if its a real rule, but I feel like the smaller the dataset the more you should stick to really basic simple measures.
I like the idea of this leading the parade analogy, but I feel there are two concepts here, leading the parade in terms of a powerless figurehead, and a distinct idea. One I more science, the other more politics.
Whether or not your research has counterfactual impact depends not just on what you do, but also on what everyone else in the world does. If there are 100 problems to work on, ranked in "interesting-ness" from top to bottom and you are one of 100 researchers, how do you maximise your counterfactual impact? Obviously not by working on the biggest problems, the "most interesting" problem will probably have 5 other people doing it. (The game-theory thing where "If we all go for the blonde" https://plus.maths.org/content/if-we-all-go-blonde ).
The theory does tell us that if we happen to know problem X is way more interesting than most people give it credit for we should prioritise that, but that's not much of an insight.
I think "leading the parade" has a much more interesting application to politics. Imagine the leader of a political party - depending on the leader, and the party, the party leader might have genuine, power to change the party's policies or political aims, but (usually?) the party already has its "thing" and the leader can only adjust within that. Donald Trump is an example of someone who clearly has real power to change the sorts of things the Republicans stand for, because he already has. In contrast, most party leaders just do what that party would do anyway. That seems like a really important thing to know. Imagine the prime minister has some policy that you really don't like, you manage to get a meeting with them and explain why the policy is bad. Then, in a weird moment of perfect honesty, they look you in the eye and say "yeah, on a personal level I completely agree this policy is bad. But my party are behind this, and its happening with or without my backing. Why are you even talking to me, I am just the prime minister, I lead this party but I don't control it."
(Unrelated, but a few years ago I published a paper that I really thought was like, some crazy thing that would only be seen by someone with my weird perspective on the issue - proper counterfactual impact. It only took 2 years for someone else to independently come up with the same idea (and the same figures!) and put up a preprint of their own.)
This question can be interpreted two ways:
(1) What really great things do we not have yet, which we might have in the future? Possibly accelerated by AI.
(2) What really great things absolutely require futuristic AI, and we can never have them without first inventing it.
I think (2) contains things like robot butlers, cleaning your house and doing your chores (whether they qualify as "really great" maybe up to taste). (1) contains things likes cures for {cancer/malaria/Alzheimer's/depression/suicide headaches}, and machines that are more {cost/energy/carbon} efficient, which could all plausibly be invented faster with AIs helping.
When I imagine the future I imagine first an age of insane abundance: where some random university students replace their self-driving cars every few months just to keep up with the new season's fashion. (This summer, drive a convertible. No black cars after labour day.). Then last-months self-driving cars all take themselves to a nano-bot factory to self-recycle into new ones. Sure, automobile-fast-fashion is basically just posing and a bit crazy. And I would be scoffing at the insanity of it and the shallowness of the people involved like the angry moralising old man I would be. But, none of those people in those cars are starving, none of them are in danger of malaria. Perhaps they would all be immune to aging. The abundance is the symptom of things sorted.
Then, at some point after that, I imagine an age where physical resource goods cease to have status associations. Like, even some poor looser could have a private jet if they wanted, it doesn't prove anything. Why would you want one anyway? The gravity train is faster and the airstrip is miles away from your home. I don't know what that world looks like. Maybe people find new ways of claiming status, everyone wants to tell your about their new book or their wonderful status-worthy politics, maybe people just move on and have fun.
All this is on the edge of my knowledge, so I could well be wrong. Insert "I thinks" and "from what I remembers" as appropriate throughout what follows.
If we start with non-interacting air molecules then the standing waves of pressure are the normal modes of the container. With non-interacting molecules the movement of a single molecule is not necessarily chaotic, whether it is or not depends on the shape of the container.
Assuming no loss (Q factor of infinity) then, knowing that the motion contains some contribution from a particular normal mode allows us to plot that normal mode (sine wave say) out to infinite future (and past) times. However, in a chaotic system it is required that the frequencies of the normal modes are approximately equally spaced. Their are no big gaps in the frequencies. I think the relevance of this to this question is that if all we know is that normal mode number 27 has some amplitude that sine wave we can infer out is added to all the other modes, which add white noise. (The mode spacing argument ensuring the noise is in fact white, and not colored noise that we could exploit to actually know something). So, assuming that mode 27 only has a typical amplitude we learn very little.
When we add collisions between the air molecules back in, then I believe it is chaotic for any shape of container. Here the true normal modes of the total system include molecule bumping, but the standing waves we know about from the non-interacting case are probably reasonably long-lived states.
A little (perhaps pedantic) point of observation:
"there's no reason information should be preserved for 20 seconds" - A sort of aside on this comment. For an idealised classical mechanical (reversible) system the information will be preserved forever. Chaos, roughly speaking, is moving the same information into different significant figures (or relations between them). So an initial uncertainty in the 20th decimal place soon becomes an uncertainty in the first significant figure as the information moves about. All that information you had about the leading digits of precision describing the initial state is still there in some sense, it has been mapped into some bizarre constraints connecting different figures deep behind the decimals, and is practically useless.
So, in this sense the question (I think) is whether that remaining information you have tells you anything about which side of the box will have more molecules at t=20. So, instead of a 50/50 guess, does the information let you get to 60/40 or whatever.
My feeling is that it must be almost worthless. Lets say that information takes you from a 50/50 guess to a 50+s/50-s guess. My intuition is that if we plotted this "s" value as a function of time it is likely an exponential decay, and that 20 seconds feels like a very long time compared to the timescales involved in the molecular motion. At t=0 s will be very close to 50 (only if the perturbed molecule is within one angstrom of the dividing line between left and right will s be less than 50 at t=0). But at t=20 it has undergone many half-lives. So its probably 10^{-big number} after 20 seconds.
The prediction that the information would be significant implies an assumption that s does not exponentially decay with time, but is described by some other function (maybe a constant). So I think the core of the dispute might different assumptions on the shape of the s(t) function.
If she can, it might be nice for her to put a few minutes of microscope footage attached to the paper as a supplementary. Maybe just "bad conditions" (stuff flowing by all different sizes), followed by "good conditions": stuff all the same. Lots of journals offer the possibility of videos as supplementary information. Its the sort of thing that (I think) journal editors like (maybe it boosts engagement by the metrics they use for their websites?), and it sounds like it will benefit the paper.
I have never had any interaction with lawsuits of any kind, including those relating to libel etc.
However, a social dynamic I have observed several times in my life is that Person A and Person B have some kind of conflict. Person A is utterly convinced of the rightness (and righteousness) of their position.
Someone (either B or a third party, C) suggests that it may be prudent to involve the police or the teachers/parents (if these people are chidden) or other authorities. Person A, on mention of the police/teachers/whoever suddenly looses that utter confidence they had in the obvious morality of their position, and runs a mile. Curiously person A in this case will often genuinely feel that the mention of involving authority was an attack or an escalation (Although I think they are always wrong in that estimation, and that the main effect is to de-escalate). I have, on none of these occasions, actually seen the police/teachers/whoever actually be contacted. I believe there exists a certain frame of mind a human can get into, where they are in a position of relative power, and believe they have a great moral authority behind them. And that simply being reminded that they may potentially have to persuade a higher authority of this rightness is enough to break the spell.
So, reminding people that they may need to answer to a higher authority for their actions is a generically useful strategy against a wide range of attacks, one that I imagine people fall back on instinctively all the time. In this case that means mentioning lawyers. Holding "they mentioned lawyers" against the Nonlinear people seems insane. Mentioning authority is a prudent way of defusing or deescalating many social situations. Even if you think it was the wrong move in this exact case I think you shouldn't judge someone (who is probably in a bit of a flap given the accusations involved) too harshly for making the move that is usually right.
I remember getting the same answer to the same question at school and being quite surprised.
If I was asked the question now I would show the interested student this pattern, and ask them what makes sense for the last term.
x^3 = 1*x*x*x
x^2 = 1*x*x
x^1 = 1*x
x^0 = ?
(if they are still unconvinced invite them to approach 0 from the other side)
x^(-3) = ((1/x)/x)/x [up to brackets = 1/x/x/x]
x^(-2) = (1/x)/x [1/x/x]
x^(-1) = 1/x
In some ways the negative powers (if the student already knows those) are cleaner, because the implicit factor of 1 is made explicit in fractions. Although the need for brackets makes the pattern less elegant.
I think the main reason the x^0 = 0 definition is awful is that it means you can no longer rely on the fact that
x^n * x^m = x^(n+m)
(because for n=0 the left hand side is zero, but the right hand side is x^m, which may or may not be zero). Although I think that (depending on the age of the student) this might not be the best explanation.
I think this is a very interesting thought experiment, and it is probing something real.
However, I think that it is missing something important in how it maps to graduate admissions. Lets say that for one reason or another applications are few and places are many. Instead of accepting the top 10 out of 100 applications you are going to reject the bottom 10, and accept the other 90. This legibility argument suggests that in such a situation the applicants from the prestigious universities will be disadvantaged. (Simply slide the "accepted" line in your bell curve graph to the left). This feature strikes me as not matching the reality I imagine I live in.
I think the missing layer of this is that, as an applicant (to a job or anything) you usually have the option to intentionally make your application less legible in various ways, by omitting information that would disadvantage you. You know that Paige is not going to give a great estimate of your performance, so you don't ask her. As a result less legible applications are going to be correlated with weaker ones. The assessor should, to an extent, be interpreting absence of good evidence (evidence that greatly narrows the candidates uncertainty in a rightwards direction) as evidence of absence.
[Exampleina knows she is in the weaker half of the class, and she knows that Mrs Paige knows this, so doesn't pick her as her reference. Instead she goes to Mr Oblivious, whose opinions are very weakly correlated with reality, but who happens to think Exampleina is incredibly gifted.
Exampaul can speak a foreign language, and to make this more legible in his application he pays an outside organisation to set him an exam on this language to get himself a certificate he can mention on his application. Unfortunately, Exampaul over-estimates his fluency, does not prepare for the test, and he scores a D-grade in his fluency test. He doesn't mention the fluency-test he paid for on the application at all, and simply puts "fluent in {language}" on his form without further evidence.]
Joke suggestion. Suppose every time you get banged on the head you become awesome at something. But that almost all of those somethings haven't been invented yet, and many never will.
(eg. Caveman get hit by a rock thrown by another caveman. Becomes a brilliant pianist, lives on to death without knowing it, or even what a piano is.)
An example is error bars in excel.
When I was at school adding error bars could be done with a single click on a button, then selecting the data that told the graph how big the error bars were. If they were all the same size you could just drag to get a column repeating the same number.
Now the error bars button is about three submenus deep (or feels that way), so many more clicks are needed to find it. Then, when you press "add error bars" it puts error bars of length 1 on the X and Y axes immediately, without asking for sizes. If you have already chosen sensible axis limits this often completely ruins them (eg. The quantity goes from -1E-5 to +1E-5, oh no! Error bars of size 1, change scale!). Then you need to click on the error bars to modify them to sensible numbers, which (if you were at a large scale) requires you to ruin the scale anyway to zoom in to a point that error bars of length 1 are visible.