## LESSWRONGLW

Domain Theory and the Prisoner's Dilemma: FairBot

We can model events (such as states of a program) with posets, related with each other by monotone maps. By beliefs I mean such posets or their elements (belief states). A state of an event can be enacted by an agent if the agent can bring it into that state. So if the agent decides on an action, that action can be enacted. But if instead the agent only decides on things like beliefs about actions (powersets of sets of possible actions), these can't be enacted, for example it can't ensure that the action is {C, D} or ⊥, that doesn't make sense. But for the... (read more)

2Gurkenglas4dYeah, enacting comes in at a higher level of interpretation than is yet considered here. The increasing levels of interpretation here are: Set theory or other math foundations; we consider sets of queries and beliefs and players with functions between them; we add porder and monotonicity; we specify proof engines and their properties like consistency; we define utilities, decision theories, and what makes some players better than others. (Category theory is good at keeping these separate.) I'd start talking about "enacting" when we define a decision theory like "Make the decision such that I can prove the best lower bound on utility.". What do you mean by deciding on a belief state? "Decision" is defined before I establish any causation from decisions to beliefs. Oh, I thought you meant you didn't see why any two beliefs had an upper bound. My choice to make players monotonic comes from intuition that that's how the math is supposed to look. I'd define Query=P(Decision) as Decision->2 as well but that plainly makes no sense so I'm looking for the true posetty definition of Query, and "logical formulas" looks good so far. Switching back and forth sounds more like you want to do multiple decisions, one after the other. There's also a more grounded case to be made that your policy should become more certain as your knowledge does, do you see it?
Domain Theory and the Prisoner's Dilemma: FairBot

Decisions should be taken about beliefs (as posets), not actions (as discrete sets). With the environment modeled by monotone maps, solutions (as fixpoints) can only be enacted for beliefs that are modeled as themselves, not for things that are modeled as something other than themselves (powersets for events with small finite sets of possible states, etc.). Also, only things that shall be enacted should be modeled as themselves, or else solutions won't be enacted into correctness.

This way, actions may depend on beliefs in an arbitrary way, the same as even... (read more)

2Gurkenglas5dThe belief state diagram is upward closed because I included the inconsistent belief states. We could say that a one-query player "decides to defect" if his query is proven false. Then he will only decide on both decisions when his beliefs are inconsistent. Alternatively we could have a query for each primitive decision, inducing a monotone map from P({C,D}) to queries; or we could identify players with these monotone maps. I didn't follow the bit about being modeled as oneself. Every definition of the belief space gives us a player space, yes? And once we specify some beliefs we have a tournament to examine, an interesting one if we happen to pipe the player's outputs into their inputs through some proof engines. Define enacted.
Small and Vulnerable

There is never native ultimate bedrock with human minds that has any clarity to it. Concepts for how people think are mostly about cognitive technology that someone might happen to implement in their thinking, they become more reliably descriptive only at that point. All sorts of preferences and especially personal pursuits are possible, without a clear/principled reason they develop. The abstract arguments I'm gesturing at amplify/focus a vague attitude of "suffering is bad", which is not rare and doesn't require any particular circumstances to form, into actionable recommendations.

Decontextualizing Morality

There are many ways of framing the situation, looking for models of what's going on that have radically different shapes. It's crucial to establish some sort of clarity about what kind of model we are looking for, what kind of questions or judgements we are trying to develop. You seem to be conflating a lot of this, so I gave examples of importantly different framings. Some of these might fit what you are looking for, or help with noticing specific cases where they are getting mixed up.

1ACrackedPot8dI feel like I was reasonably clear that the major concern was about how utilitarianism interacts with being human, as much of the focus is on moral luck. Insofar as an intelligent boat can be made miserable by failing to live up to an impossible moral system, well, I don't know, maybe don't design it that way.
Decontextualizing Morality

When a highly intelligent self-driving boat on the bank of a lake doesn't try to save a drowning child, what is the nature of the problem? Perhaps the boat is morally repugnant and the world will be a better place if it experiences a rapid planned disassembly. Or the boat is a person and disassembling or punishing them would in itself be wrong, apart from any instrumental value gained in the other consequences of such an action. Or the fact that they are a person yet do nothing qualifies them as evil and deserving of disassembly, which would not be the cas... (read more)

1ACrackedPot8dWould you mind unpacking this?
Josh Smith-Brennan's Shortform

At this level of technical discussion it's hopeless to attempt to understand anything. Maybe try going for depth first, learning some things at least to a level where passing hypothetical exams on those topics would be likely, to get a sense of what a usable level of technical understanding is. Taking a wild guess, perhaps something like Sipser's "Introduction to the Theory of Computation" would be interesting?

1Josh Smith-Brennan10dI just downloaded the 2nd edition. Thank you for the suggestion.
Josh Smith-Brennan's Shortform

Maybe read up on the concepts of outcome, sample space, event, probability space, and see what the probability of intersection of events means in terms of all that. It's this stuff that's being implicitly used, usually it should be clear how to formulate the informal discussion in these terms. In particular, truth of whether an outcome belongs to an event is not fuzzy, it either does or doesn't, as events are defined to be certain sets of outcomes.

(Also, the reasons behind omission of the "or equal to"s you might've noticed are discussed in 0 And 1 Are Not... (read more)

1Josh Smith-Brennan10dThanks for the suggestions, and now that I understand the idea that the probability values correspond to a binary interpretation of the events, it makes these areas easier to navigate for me in discussion. This definitely stands as a hard to argue against idea, and it makes sense when seen from the viewpoint of rational humans interpreting data from systems based on binary calculations and logic. Do you think it's possible that there is a better way than Binary logic to compute and reason though? Not being familiar with the literature, I wonder if it's possible that because we have relied on binary logic to compute and reason for so long, it's created a false dichotomy in our understanding of reality. Is there another way to reason that works better, based on quantum computing rational? How the next decade will add to the discussion of reality in terms of the advances in Quantum computing seems to be debatable. Translating probability values into either true or false logic is a step in binary computing that I believe quantum computing skips, and so the data returned takes in to account I think, the cases in which "...one of the events includes the other..." in a more or less straightforward way. At this point though, (I could be wrong) I believe there is still more of a front end system that runs binary to interpret the Quantum calculations of a quantum computer, because when the data returned isn't binary, we're still trying to figure out what it's good for. In relation to events and how long or little they last, this whole area of Quantum clocks [https://en.wikipedia.org/wiki/Quantum_clock] is interesting to me. We can measure time more accurately because of them, but it seems like so much of the science in common use still relies on the second as the base measurement. Maybe the second is the bottom limit of what humans can somewhat accurately perceive without aid of a tool like a watch, which makes a case for basing measurements of time using more accurate
Small and Vulnerable

The intuition pump does live at this level of abstraction, but it's a separate entity from the abstract consideration it's meant to illustrate, which lives elsewhere. My disagreement is with how the first paragraph of the post frames the rest of it. Personal or vicarious experience of trauma is not itself a good reason for pursuing altruism, instead it's a compelling intuition pump for identifying the reason to do so. Some behaviors resulting from trauma are undesirable, and it's the abstract consideration of what motivates various induced behaviors that l... (read more)

Small and Vulnerable

This is the kind of thing that feels compelling, but emphasizes a wrong level of abstraction. Personal experience of suffering is not the reason why suffering is bad. It's a bit like professing that two plus two is four because the teacher says so. The teacher is right, but there is a reason they are right that is more important than the fact that they are saying this. Similarly, personal suffering is compelling for the abstract conclusion of altruism, but there is a reason it's compelling that is more important as a consideration for this conclusion than the fact of experience. Someone with no personal experience of suffering should also be moved by that consideration.

44thWayWastrel9dI find this sentiment a little confusing, as it seems to me the subjective experience of suffering is the ultimate bedrock of any idea that understands suffering as bad? If I had no personal experience of suffering or wellbeing [https://en.wikipedia.org/wiki/Valence_(psychology)] I can't imagine how something like utilitarianism might move me. Or are you saying while yes ultimately an abstract understanding of suffering rests on a subjective experience of it, pumping the understanding of the subjective experience won't lead to more understanding of it in the abstract in the way EA needs to?

Someone with no personal experience of suffering should also be moved by that consideration.

That sounds like a fantastic reason for someone with that experience to post it, as occurred here, as a way to explain what it is like to others.

In fact, only the existence of suffering for some concrete individual justifies the abstract conclusion of altruism. Without that concrete level, the abstraction is hypothetical, and should not provide the same level of reason to be altruistic.

I don’t really get EA at an emotional level and this post helps give someone like me an... emotional intuition pump?... in a way that other EA posts do not do for me. I think it’s good that it is at the level of abstraction it is at.

Death by Red Tape

That's ambitious without an ambition. Switching domains stops your progress in the original domain completely, so doesn't make it easier to make progress. Unless domain doesn't matter, only fungible "progress".

"Progress" can be a terminal goal, and many people might be much happier if they treated it as such. I love the fact that there are fields I can work in that are both practical and unregulated, but if I had to choose between e.g. medical researcher and video-game pro, I'm close to certain I'd be happier as the latter. I know many people which basically ruined their lives by choosing the wrong answer and going into dead-end fields that superficially seem open to progress (or to non-political work).

Furthermore, fields bleed into each other. Machine learning ... (read more)

7deluks91713dIt only assumes there are a lot of domains in which you would be happy to make progress. In addition success is at least somewhat fungible across domains. And it is much easier to cut red tape once you already resources and a track record (possibly in a different domain). Don't start out in a red-tape domain unless you are ready to fight off the people trying to slow you down. This requires a lot of money, connections, and lawyers and you still might lose. Put your talents to work in an easier domain, at least to start.
Best empirical evidence on better than SP500 investment returns?

(This feels more like a dragon hoard than retirement savings, something that should only form as an incidental byproduct of doing what you actually value, or else under an expectation of an increase in yearly expenses.)

1bluefalcon17dHaha. 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.
Best empirical evidence on better than SP500 investment returns?

My expenses are well below my income; I'm done saving for retirement

Note that a simple FIRE heuristic giving about 3% chance of running out of money at some point is to have 30x yearly expenses in 100% stock index with no leverage, which is a lot more than the usual impression along the lines of "my expenses are well below my income" and is still not something that can be reasonably described as safe.

2bluefalcon19dI 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.
Best empirical evidence on better than SP500 investment returns?

Leverage can give arbitrarily high returns at arbitrarily high risk. With things easily available at a brokerage, this goes up to very high returns with insane risk. See St. Petersburg paradox for an illustration of what insane risk means. I like the variant where you continually bet everything on heads in an infinite series of fair coin tosses, doubling the bet if you win, so that for the originally invested $100 you get back the same$100 in expectation at each step (at first step, $200 with probability 1/2 and$0 with probability 1/2; by the third step,... (read more)

Covid 4/22: Crisis in India

when the institutions are bad and spread insane views, this outsourced thinking causes the trusting majority to share those insane views

Or alternatively, with the model of institutions as competent but dishonest, the takeaway from an action with an implausible-sounding explanation (pausing vaccination out of "an abundance of caution") is to make up your own explanation that would make the action seem reasonable (there are issues that are actually serious), and ignore all future claims from the institution on the topic ("we checked and it seems fine").

How You Can Gain Self Control Without "Self-Control"

The article gives framing and advice that seem somewhat arbitrary, and doesn't explain most of the choices. It alludes to research, but the discussion actually present in the article is only tangentially related to most of the framing/advice content, and even that discussion is not very informative when considered in isolation, without further reading.

There is a lot of attention to packaging the content, with insufficient readily available justification for it, which seems like a terrible combination without an explicit reframing of what the article wants to be. With less packaging, it would at least not appear to be trying to counteract normal amount of caution in embracing content of (subjectively) mysterious origin.

4spencerg2moHi Vladimir, thanks for your comments. Could you elaborate on what you would like to see more justification for when you say ("insufficient readily available justification")? I'd also be interested to know what framing seemed "somewhat arbitrary." In the section "Nine Traits of Self-Controlled Behavior" my claim is that those pretty self-evidently are traits that (i) differ non-negligibly between people and (ii) can manifest as "self-controlled behavior." Are there items in that list that you think don't differ between people, or that you don't think can manifest in self-controlled behavior? I view that list as the sort of thing that someone can check for themselves by simply seeing whether they agree with each item. Maybe you're wondering where this list comes from? It is a list of all of the traits I know of that I believe manifest as self-controlled behavior. So I don't view it as arbitrary - surely there are ones I didn't think of, but I was attempting to be comprehensive. In the section "Twelve Simple Strategies for Gaining More Control" many of those strategies have a whole research literature on them. Others are common-sense strategies. I certainly can't claim this is a comprehensive list of strategies, as hundreds of strategies exist. So maybe this list seemed arbitrary? On the ego depletion stuff, I go into a lot of detail on my thinking, that I think gives the reader plenty of information to decide whether they agree with what I'm saying or not. If you disagree I'd be interested to know. Thanks again for your comments!
What are the best resources to point people who are skeptical of getting vaccinated for COVID-19 to?

The distinction is between understanding and faith/identity (which abhors justification from outside itself). Sometimes people build understanding that enables checking if things make sense. This also applies to justifying trust of the kind not based on faith. The alternative is for decisions/opinions/trust to follow identity, determined by luck.

Impact of the rationalist community who someone who aspired to be "rational"

Naming a group of people is a step towards reification of an ideology associated with it. It's a virtuous state of things that there is still no non-awkward name, but keeping the question of identity muddled and tending towards being nameless might be better.

samshap's Shortform

Sleeping Beauty illustrates the consequences of following general epistemic principles. Merely finding an assignment of probabilities that's optimal for a given way of measuring outcomes is appeal to consequences, on its own it doesn't work as a general way of managing knowledge (though some general ways of managing knowledge might happen to assign probabilities so that the consequences are optimal, in a given example). In principle consequentialism makes superfluous any particular elements of agent design, including those pertaining to knowledge. But that observation doesn't help with designing specific ways of working with knowledge.

1samshap2moMy argument is that the log scoring rule is not just a "given way of measuring outcomes". A belief that maximizes E(log(p)) is the definition of a proper Bayesian belief [https://www.yudkowsky.net/rational/technical]. There's no appeal to consequence other than "SB's beliefs are well calibrated".

Labels are no substitute for arguments.

But that's the nature of identity: a claim that's part of identity won't suffer insinuations that it needs any arguments behind it, let alone the existence of arguments against. Within one's identity, labels are absolutely superior to arguments. So the disagreement is more about epistemic role of identity, not about object level claims or arguments.

See proving too much. In the thought experiment where you consider sapient wolves who hold violent consumption of sentient creatures as an important value, the policy of veganism is at least highly questionable. An argument for such a policy needs to distinguish humans from sapient wolves, so as to avoid arguing for veganism for sapient wolves with the same conviction as it does for humans.

Your argument mentions relevant features (taste, tradition) at the end and dismisses them as "lazy excuses". Yet their weakness in the case of humans is necessary for th... (read more)

This proves too much. Most of these arguments would profess to hold veganism as the superior policy for sapient wolves (who are sufficiently advanced to have developed cheap dietary supplementation), degrading the moral imperative of tearing living flesh from the bones.

1Ruben Tinstil2moI am sorry, I do not get your point here. Could you elaborate what you mean?
Weighted Voting Delenda Est

This is a much clearer statement of the problem you are pointing at than the post.

(I don't see how it's apparent that the voting system deserves significant blame for the overall low-standard-in-your-estimation of LW posts. A more apparent effect is probably bad-in-your-estimation posts getting heavily upvoted or winning in annual reviews, but it's less clear where to go from that observation.)

Takeaways from one year of lockdown

The stress of negotiation/management of COVID precautions destroyed my intellectual productivity for a couple of months at the start of the pandemic. So I rented a place to live alone, which luckily happened to be possible for me, and the resulting situation is much closer to normal than it is to the pre-move situation during the pandemic. There is no stress, as worrying things are no longer constantly trying to escape my control without my knowledge, there's only the challenge of performing "trips to the surface" correctly that's restricted to the time of the trips and doesn't poison the rest of my time.

Subjectivism and moral authority

As I understand this, Clippy might be able to issue an authoritative moral command, "Stop!", to the humans, provided it's "caused" by human values, as conveyed through its correct understanding of them. The humans obey, provided they authenticate the command as channeling human values. It's not advice, as the point of intervention is different: it's not affecting a moral argument (decision making) within the humans, instead it's affecting their actions more directly, with the moral argument having been computed by Clippy.

"If You're Not a Holy Madman, You're Not Trying"

The nice things are skills and virtues, parts of designs that might get washed away by stronger optimization. If people or truths or playing chess are not useful/valuable, agents get rid of them, while people might have a different attitude.

(Part of the motivation here is in making sense of corrigibility. Also, I guess simulacrum level 4 is agency, but humans can't function without a design, so attempts to take advantage of the absence of a design devolve into incoherence.)

"If You're Not a Holy Madman, You're Not Trying"

It's not clear that people should be agents. Agents are means of setting up content of the world to accord with values, they are not optimized for being the valuable content of the world. So a holy madman has a work-life balance problem, they are an instrument of their values rather than an incarnation of them.

9ryan_b2moThis is a very striking statement, and I want to flag it as excellent.
6abramdemski3moI think there are a couple of responses the holy-madman type can give: * The holy-madman aesthetic is actually pretty nice [https://www.lesswrong.com/posts/SGR4GxFK7KmW7ckCB/something-to-protect]. Human values include truth, which requires coherent thought. And in fiction, we especially enjoy heroes who go after coherent goals. So in practice in our current world, the tails don't come apart much. At worst, people who manage to be more agentic aren't making too big of a sacrifice in the incarnation department. And perhaps they're actually better-off in that respect. * A coherent agent is basically what happens when you can split up the problem of deciding what to do and doing it, because most of the expected utility is in the rest of the world. An effective altruist who cares about cosmic waste probably thinks your argument is referring to something pretty negligible in comparison. Even if you argue functional decision theory means you're controlling all similar agents, not just yourself, that could still be pretty negligible.
What are a rationalist's best tools for better decision making?

What are a rationalist's best tools for better decision making?

What are a farrier's best recipes for better pizza? Probably the same as an ophthalmologist's. What about worse pizza, or worse recipes?

Omit needless words. Yes requires the possibility of no.

A No-Nonsense Guide to Early Retirement

Investing everything in a single ETF (especially at a single brokerage) is possibly fine, but seems difficult to justify. When something looks rock solid in theory, in practice there might be all sorts of black swans, especially over decades (where you lose at least a significant portion of the value held in a particular ETF at a particular brokerage, compared to its underlying basket of securities, because something has gone wrong with the brokerage, the ETF provider, the infrastructure that makes it impossible for anything to go wrong with an ETF, or som... (read more)

Is the influence of money pervasive, even on LessWrong?

Identity colors the status quo in how the world is perceived, but the process of changing it is not aligned with learning (it masks the absence of attempting to substantiate its claims), thus a systematic bias resistant to observations that should change one's mind. There are emotions involved in the tribal psychological drives responsible for maintaining identity, but they are not significant for expressing identity in everything it has a stance on, subtly (or less so) warping all cognition.

There's some clarification of what I'm talking about in this comm... (read more)

How is rationalism different from utilitarianism?

Rationality is perhaps about thinking carefully about careful thinking: what it is, what it's for, what is its value, what use is it, how to channel it more clearly. Utilitarianism is about very different things.

How is rationalism different from utilitarianism?

It's instrumentally useful for the world to be affected according to a decision theory, but it's not obviously a terminal value for people to act this way, especially in detail. Instrumentally useful things that people shouldn't be doing can instead be done by tools we build.

There is no fundamental reason for Cheerful Price to be higher than what you are normally paid. For example, if you'd love to do a thing even without pay, Cheerful Price would be zero (and if you can't arbitrage by doing the thing without the transaction going through, the price moves all the way into the negatives). If you are sufficiently unusual in that attitude, the market price is going to be higher than that.

8weft3moYes, you are correct that the Cheerful Price could be less than my normal wage. But this is not usually the case for me. People aren't usually asking my Cheerful Price to eat some ice cream, or something similarly pleasant. And unfortunately we don't live in a world where my regular wages are above my Cheerful Price.
Is the influence of money pervasive, even on LessWrong?

strong emotional reactions

I expect being part of one's identity is key, and doesn't require notable emotional reactions.

2Dagon3moOnly processing this now, and I'd like to understand your model more deeply. I think that beliefs being part of one's identity is highly correlated with strong emotions for things that challenge or support those beliefs. There are other things which engender strong emotional reactions, as well, so I think of "part of identity" as a subset of things which are difficult contexts in which to discuss rationality. For instance, one's livelihood (even when it's not particularly part of identity) is likely a difficult topic to explore well in this forum.
The 10,000-Hour Rule is a myth

My woefully inexpert guess is that advanced cooking should be thought of as optimization in a space of high dimension, where gradient descent will often zig-zag, making simple experiments inefficient. Then apart from knowledge of many landmarks (which is covered by cooking books), high cooking skill would involve ability to reframe recipes to reduce dimensionality, and intuition about how to change a process to make it better or to vary it without making it worse, given fine details of a particular setup and available ingredients. This probably can't be usefully written down at all, but does admit instruction about changes in specific cases.

3Stefan De Young3moReducing dimensionality is the most useful cooking advice I have received. I now use a four factor model: salt, sweet, spice (heat), sour. * Is it salty enough? If no, add salt, soy sauce, or fish sauce; or reduce. * Is it sweet enough? If no, add sugar, jagery, maple syrup or caramelized onions. The essentialism is to assign characteristics to ingredients (e.g. Tomatoes are sour.) I learned this model from some south Indians, this model may be common in that culture. I'm not sure.

Let's apply Bayes formula in odds form to this example. Let = "Xavier won the lottery", , = "Yovanni says Xavier won the lottery". We have (for simplicity, let's assume that Xavier couldn't be someone who didn't even enter the lottery), . What is ? Given that someone other than Xavier won the lottery, what is the probability that Yovanni would claim that it was Xavier in particular who did? While Yovanni might have a reason to single out Xavier, Zenia doesn't, so from the hypothes... (read more)

The 10,000-Hour Rule is a myth

The 10,000-Hour Rule [...] popularized by Malcolm Gladwell [says that] ten thousand hours of practice is necessary and sufficient to become an expert.

Not having read the book, and from a cursory google search, I was unable to find a clear argument that Gladwell actually makes the claim about practice being sufficient. I did find his own statement that this is not a claim he made in the book. (My impression is that the book was criminally ambiguous and neither affirms nor denies the claim despite discussing related things at length.)

2lynettebye3moThat seems likely. I'm not calling Gladwell out - I also haven't read the book, and there's probably a pretty defensible motte [https://en.wikipedia.org/wiki/Motte-and-bailey_fallacy]there. However, it seems likely that he laid the foundation for the popular internet version by overstating the evidence for it, e.g. this quote from the book: “The idea that excellence at performing a complex task requires a critical minimum level of practice surfaces again and again in studies of expertise. In fact, researchers have settled on what they believe is the magic number for true expertise: ten thousand hours." And the rule-run-amok-on-the-internet generally assumes necessary and sufficient, e.g. this quote from Ericsson "The popular internet version of the 10 000 h rule suggests that attaining expert performance is all about getting more and more practice and experience in a given domain of activity and then reaching an expert status at 10 000 h."
Self-Criticism Can Be Wrong And Harmful

If at all possible, good activities with risks should at the very least be approached with caution and training, not outright avoided.

Taking ideas seriously is potentially harmful as ideas are possibly no good, prompting the general strategy of steering clear of ideas. The urges of asymmetric justice also pull in this direction, as application of norms, with presence of any blamable risks paralizing even when action is manifestly good in expectation.

What's the big deal about Bayes' Theorem?

The above formula is usually called "odds form of Bayes formula". We get the standard form by letting in the odds form, and we get the odds form from the standard form by dividing it by itself for two hypotheses ( cancels out).

The serious problem with the standard form of Bayes is the term, which is usually hard to estimate (as we don't get to choose what is). We can try to get rid of it by expanding but that's also no good, because now we need to know . One way to state the problem... (read more)

It's possible for two programs to know each other's code and to perfectly deduce each other's result without taking forever, they just can't do it by simulating each other. But they can do it by formal reasoning about each other, if it happens to be sufficiently easy and neither is preventing the other from predicting it. The issues here are not about fidelity of prediction.

This is not the halting problem. The halting problem is about existence of an algorithm that predicts all algorithms. Here, we are mostly interested in predicting Agent and Predictor.

However, the standard argument about the halting problem can be applied to this case, giving interesting constructions. Agent can decide to diagonalize Predictor, to simulate it and then say the opposite, making it impossible for Predictor to predict it. Similarly, Agent can decide to diagonalize itself (which is called having a chicken rule), by committing to do the opposite ... (read more)

Non-Coercive Perfectionism

What I mean by perfectionism is a desire for a certain unusually high level of challenge and thoroughness. It's not about high valuation according to a more abstract or otherwise relevant measure/goal. So making a process "more perfect" in this sense means bringing challenge and thoroughness closer to the emotionally determined comfortable levels (in particular, it might involve making something less challenging if it was too challenging originally). The words "more perfect" aren't particularly apt for this idea.

Why novel unimportant things specifically? T... (read more)

2Matt Goldenberg4moAhh interesting, thanks for sharing!
Non-Coercive Perfectionism

I thought my answer worked for that case as well: choosing the amount of time to spend on a project looks like choosing to not abandon the project when it should be continued (out of abstract consideration of what projects are important). The alternative, abandoning of projects, bears no emotional valence, so costs no effort.

2Matt Goldenberg4moDo you think that the process by which you get to rarely encountered unimportant stuff is perfect, or could you bring more perfection to the process?
What's the big deal about Bayes' Theorem?

In daily life, the basic move is to ask,

• What are the straightforward and the alternative explanations (hypotheses) for what I'm seeing?
• How much more likely is one compared to the other a priori (when ignoring what I'm seeing)?
• What probabilities do they assign to what I'm seeing?

and get the ratio of a posteriori probabilities of the hypotheses (a posteriori odds) from that (by multiplying the a priori odds by the likelihood ratio). Odds measure the relative strengths of hypotheses, so the move is to obtain relative strengths of a pair of hypotheses af... (read more)

1AVoropaev4moThis formula is not Bayes' Theorem, but it is a similar simple formula from probability theory, so I'm still interested in how you can use it in daily life. Writing P(x|D) implies that x and D are the same kind of object (data about some physical process?) and there are probably a lot of subtle problems in defining hypothesis as a "set of things that happen if it is true" (especially if you want to have hypotheses that involve probabilities). Use of this formula allows you to update probabilities you prescribe to hypotheses, but it is not obvious that update will make them better. I mean, you obviously don't know real P(x)/P(y), so you'll input incorrect value and get incorrect answer. But it will sometimes be less incorrect. If this algorithm has some nice properties like "sequence of P(x)/P(y) you get repeating your experiment converges to the real P(x)/P(y) provided x and y are falsifiable by your experiment (or something like that)", then by using this algorithm you'll with high probability eventually update your algorithm. It would be nice to understand, for what kinds of x, y and D you should be at least 90% sure that your P(x)/P(y) will be more correct after a million of experiments. I'm not implying that this algorithm doesn't work. More like it seems that proving that it works is beyond me. Mostly because statistics is one of the more glaring holes in my mathematical education. I hope that somebody has proved that it works at least in the cases you are likely to encounter in your daily life. Maybe it is even a well-known result. Speaking of the daily life, can you tell me how people (and you specifically) actually apply this algorithm? How do you decide, in which situation it is worth to use it? How do you choose initial values of P(x) (e.g. it is hard for me to translate "x is probably true" into "I am 73% sure that x is true"). Are there some other important questions I should be asking about it?
What is up with spirituality?

For explanations of popularity of religion, there's identity (simulacrum level 3), the same thing that guides dogmatic political allegiance. It probably has some basic psychological drives behind it, some of which might manifest as spirituality. This predicts spiritual experiences associated with politics.

Non-Coercive Perfectionism

Activities differ by how much time I put into them, not by effort per unit of time. I'm only choosing when to not abandon an activity. Putting more effort per unit of time is not psychologically feasible long term, while putting less effort per unit of time makes activities less enjoyable, and is only of use to quickly get unfamiliar things done (which causes me some discomfort).

2Matt Goldenberg4moAhh if it wasn't clear when I say less effort, I wasn't meaning "effort averaged over time", but less absolute effort (which in your case means spending less time)
Non-Coercive Perfectionism

There's perfectionism about results, and perfectionism about process, and these are very different. As a process perfectionist, I usually don't care about completing a project that isn't otherwise important to me, or even making much of a headway with it, only about approaching it in a well-researched way. Thus there is no systematic drain on effort towards unimportant activities, as they can be easily abandoned, while important activities are not abandoned and also get the serious attention to the process. On the other hand, I get less rarely encountered unimportant stuff done than usual (frequently encountered unimportant stuff eventually becomes efficient).

2Matt Goldenberg4moWhat would it be look to strive for perfection in your process of choosing how much effort to put into each process?
Lessons I've Learned from Self-Teaching

Instead of reading a textbook with SRS and notes, skim more books (not just textbooks), solve problems, read wikis and papers, talk to people, watch talks and lectures, solve more problems from diverse sources. Instead of memorizing definitions, figure out how different possible definitions work out, how motivations of an area select definitions that allow constructing useful things and formulating important facts. The best exercise is to reconstruct main definitions and results on a topic you've mostly forgotten or never learned that well. Math is self-he... (read more)

3Eric 'Siggy' Scott4moWhy "instead?" I find that all of those activities enhance my SRS experience, and my SRS experience (sometimes dramatically) enhances all of those activities! Particularly since I use SRS to capture concepts, intuitions, and rich relationships, not just isolated facts. Rote memorization doesn't work very well with SRS anyway--as with everything, it works best when used to understand a topic in its full intuitive glory.
Why do stocks go up?

value of the stock is $50 today and will be$5,000 ten years from now, and the rest of the market prices it at $50 today, then I could earn insane expected returns by investing at$50 today. Thus, I don't think the market would price it at \$50 today.

Everyone gets the insane nominal returns after ten years are up (assuming central banks target inflation), but after the initial upheaval at the time of the announcement there is no stock that gives more insane returns than other stocks, there are no arbitrage trades to drive the price up immediately. For n... (read more)

3Vinayak Pathak3moHmm, but what if everything gets easier to produce at a similar rate as the consumer basket? Won't the prices remain unaffected then?
3tryactions4moThis makes sense! Do you know anything about the state of evidence re: to what extent this is happening and/or driving stock returns? I'm not sure how you'd pick this apart from other causes of currency devaluation.