Posts

Cause Awareness as a Factor against Cause Neutrality 2018-08-13T20:00:49.513Z · score: 24 (15 votes)
The Partial Control Fallacy 2016-12-14T02:57:24.773Z · score: 16 (17 votes)
Raytheon to Develop Rationality-Training Games 2011-11-18T21:07:51.461Z · score: 16 (17 votes)
Meetup : Pittsburgh Meetup 2011-09-30T17:48:10.478Z · score: 1 (2 votes)
Little Johny Bayesian 2009-03-18T21:30:59.049Z · score: 13 (39 votes)

Comments

Comment by darmani on Programming Languages For AI · 2019-05-12T03:51:11.432Z · score: 3 (2 votes) · LW · GW

Hi, I'm a Ph. D. student at MIT in programming languages.

Your choice of names suggests that you're familiar with existing tactics languages. I don't see anything stopping you from implementing this as a library in Ltac, the tactics language associated with Coq.

I'm familiar with a lot of DSLs (here's umpteen of them: https://github.com/jeanqasaur/dsl-syllabus-fall-2016/blob/master/README.md ). I've never heard of one designed before they had an idea what the engine would be.

E.g.: you can write a language for creating variants of minimax algorithms, or a language for doing feature extraction for writing heuristic functions, but you wouldn't think to write either of those unless you knew how a chess AI worked. Without knowing that those are useful things, what's left? Abstract data types are sufficient for representing chess pieces cleanly. Maybe you'll decide to write concrete syntax for chess positions (e.g.: write a chess move as Nx6 and have the compiler parse it properly), but, putting aside how superficial that would be, you would do that in a language with extensible syntax (e.g.: Wyvern, Coq, Common Lisp), not a special "chess language."

The recent-ish development of probabilistic programming (hey, now there's a family of AI languages) is instructive: first was decades of people developing probabilistic models and inference/sampling algorithms, then came the idea to create a language for probabilistic models backed by an inference engine.

Comment by darmani on How does OpenAI's language model affect our AI timeline estimates? · 2019-02-15T07:43:00.017Z · score: 10 (3 votes) · LW · GW

Something you learn pretty quickly in academia: don't trust the demos. Systems never work as well when you select the inputs freely (and, if they do, expect thorough proof). So, I wouldn't read too deeply into this yet; we don't know how good it actually is.

Comment by darmani on 2018 AI Alignment Literature Review and Charity Comparison · 2018-12-27T02:10:25.201Z · score: 15 (8 votes) · LW · GW
The vast majority of discussion in this area seems to consist of people who are annoyed at ML systems are learning based on the data, rather than based on the prejudices/moral views of the writer.

While many writers may take this flawed view, there's also a very serious problem here.

Decision-making question: Let there be two actions A and ~A. Our goal is to obtain outcome G. If P(G | A) > P(G | ~A), should we do A?

The correct answer is “maybe.” All distributions of P(A,G) are consistent with scenarios in which doing A is the right answer, and scenarios in which it’s the wrong answer.

If you adopt a rule “do A, if P(G | A) > P(G | ~A)”, then you get AI systems which tell you never to go to the doctor, because people who go to the doctor are more likely to be sick. You may laugh, but I’ve actually seen an AI paper where a neural net for diagnosing diabetes was found to be checking every other diagnosis of the patient, in part because all diagnoses are correlated with doctor visits.

The moral of the story is that it is in general impossible to make decisions based purely on observational statistics. It comes down to the difference between P(G | A) and P(G | do(A)). The former is defined by counting the co-occurences of A and G; the latter is defined by writing G as a deterministic function of A (and other variables) plus random noise.

This is the real problem of bias: the decisions an AI makes may not actually produce the outcomes predicted by the data, because the data itself was influenced by previous decisions.

The third part of this slide deck explains the problem very well, with lots of references: http://fairml.how/tutorial/#/

Source: I’m involved in a couple of causal inference projects.

Comment by darmani on Overconfident talking down, humble or hostile talking up · 2018-12-01T05:55:55.137Z · score: 3 (3 votes) · LW · GW

And Paul Graham in Beating the Averages: http://www.paulgraham.com/avg.html

Comment by darmani on Cause Awareness as a Factor against Cause Neutrality · 2018-08-31T02:40:07.586Z · score: 1 (1 votes) · LW · GW

I think you hit the kernel of the argument in the first paragraph: If you have an obscure pet cause, then chances are it's because you do have some special knowledge about the problem. The person visiting a random village might not, but the locals do, and hence this is a reason why local charity can be effective, particularly if you live in a remote area where the problems are not quantified (and are hence probably not reading this).

Comment by darmani on Learning strategies and the Pokemon league parable · 2018-08-13T19:49:41.648Z · score: 4 (2 votes) · LW · GW

Put that in your post! I got what you're saying way better after reading that.

Comment by darmani on Learning strategies and the Pokemon league parable · 2018-08-12T17:39:09.538Z · score: 4 (2 votes) · LW · GW

I'm confused about whether you're talking about "learning things specifically to solve a problem" (which I've seen called "pull-based learning"), or "learning things by doing projects" (i.e.: project-based learning). The former differs from the "waterfall method" ("push-based learning") only in the sequence and selection: it's just the difference between doing a Scala tutorial because you want to learn Scala, vs. because you just got put on a project that uses Scala (and hence you can skip parts of the tutorial the project doesn't use).


For actual PBL: I am a PBL skeptic. I've seen so many people consider it self-evident that learning physics by building a catapult is superior to doing textbook problems that I wrote a blog post to highlight some of the major downsides: http://www.pathsensitive.com/2018/02/the-practice-is-not-performance-why.html . I've seen it become a fad, but I've not seen the After I wrote the blog post, I had a lot of people tell me about their negative experiences with PBL. One that stands out is a guy who took a PBL MOOC on driverless cars, and didn't like it because they spent too much time learning about how to use some special pieces of software rather than anything fundamental or transferable.


Quick summary:


Advantages of PBL:

  • More motivating to some
  • Includes all aspects of practice needed in performance (e.g.: does not omit the skill of integrating many smaller skills together)

Disadvantages:

  • Does not naturally lead to correct sequencing of knowledge
  • Not optimized for rapid learning; does not teach subskills independently
  • May omit skills which are useful for compressing knowledge, but not directly useful in practice (e.g.: learning chord structure makes it easier to memorize songs, but is not directly used in performing music)
  • May include overly-specific, non-reusable knowledge

I don't think PBL works very efficiently. I think it can produce a lot of successful practitioners, but have trouble seeing how it could produce someone able to push the boundaries of a field. I will gladly pay $10 to anyone who can give me an example of someone well-regarded in mathematics (e.g.: multiple publications in top journals in the past decade, where this person was the primary contributor) who acquired their mathematics chiefly by PBL (i.e.: not studying mathematics except for what is needed to work on a specific problem, concurrently with working on the problem).

Comment by darmani on Apptimize -- rationalist startup hiring engineers · 2015-01-15T22:50:39.257Z · score: 6 (6 votes) · LW · GW

It's more like the Crockford book -- a set of best practices. We use a fairly functional style without a lot of moving parts that makes Java very pleasant to work with. You will not find a SingletonFactoryObserverBridge at this company.

Comment by darmani on Apptimize -- rationalist startup hiring engineers · 2015-01-12T23:14:35.237Z · score: 6 (6 votes) · LW · GW

Yep. The first thing we do is have a conversation where we look for the 6 company values. Another of them is "commitment," which includes both ownership and grit.

Comment by darmani on Post ridiculous munchkin ideas! · 2013-05-12T09:43:26.807Z · score: -1 (3 votes) · LW · GW

You mean, a lot of cool mathematicians are eastern European. But Terry Tao and Shinichi Mochizuki are not.

Comment by darmani on Weekly LW Meetups: Austin, Berlin, Cambridge UK, Champaign IL, Durham NC (2), Washington DC · 2012-11-03T09:11:15.170Z · score: 4 (4 votes) · LW · GW

Chris Olah and I (Jimmy Koppel) are both Thiel Fellows and avid Less Wrongers. We'd be happy to answer any questions about the program.

Comment by darmani on Does anyone know any kid geniuses? · 2012-03-28T14:25:52.019Z · score: 4 (4 votes) · LW · GW

I know at least four people who started college by age 15. They're not "kid" geniuses anymore though -- the youngest is 16 and slowly going through college part-time, while the oldest is in his 30s and a full math professor at Arizona.

I don't know about the upbringing of the other three, but one attended a program where taking classes mutliple grade-levels ahead is the norm (though no-one else learned calculus in 3rd grade), and attended Canada/USA Mathcamp during the summers of his undergrad.

I second the Olympiads. Terry Tao famously represented Australia at the IMO at age 10, so he's definitely old enough.

Comment by darmani on Biased Pandemic · 2012-03-13T04:49:29.868Z · score: 4 (4 votes) · LW · GW

With a sufficiently strong player, Arkham Horror is a one player game which seven people play.

Comment by darmani on Diseased disciplines: the strange case of the inverted chart · 2012-02-05T16:43:04.559Z · score: 3 (3 votes) · LW · GW

There is a very healthy (and mathematical) subdiscipline of software engineering, applied programming languages. My favorite software-engineering paper, Type-Based Access Control in Data-Centric Systems, comes with a verified proof that, in the system it presents, data-access violations (i.e.: privacy bugs) are impossible.

This is my own research area ( http://www.cs.cmu.edu/~aldrich/plaid/ ), but my belief that this was a healthy part of a diseased discipline is a large part of the reason I accepted the position.

Comment by darmani on Completeness, incompleteness, and what it all means: first versus second order logic · 2012-01-27T16:57:08.815Z · score: 2 (2 votes) · LW · GW

Yes, but the point is that we are learning features from empirical observations, not using some magic deduction system that our computers don't have access to. That may only be one bit of information, but it's a very important bit. This skips over the mysterious part in the exact same way that "electrical engineering" doesn't answer "How does a CPU work?" -- it tells you where to look to learn more.

I know far less about empirical mathematics than about logic. The only thing along these lines I'm familiar with is Douglas Lenat's Automated Mathematician (which is only semi-automated). A quick search for "automated mathematician" on Google Scholar gives a lot of more recent work, including a 2002 book called "Automated theory formation in pure mathematics."

Comment by darmani on Completeness, incompleteness, and what it all means: first versus second order logic · 2012-01-25T02:35:59.558Z · score: 2 (4 votes) · LW · GW

We form beliefs about mathematics the same way we form beliefs about everything else: heuristic-based learning algorithms. We typically accept things based on intuition and inductive inference until trained to rely on proof instead. There is nothing stopping a computer from forming mathematical beliefs based on statistical inference rather than logical inference.

Have a look at experimental mathematics or probabilistic number theory for some related material.

Comment by darmani on Completeness, incompleteness, and what it all means: first versus second order logic · 2012-01-20T03:18:51.177Z · score: 0 (0 votes) · LW · GW

Computers can prove everything about integers that we can. I don't see a problem here.

Comment by darmani on Completeness, incompleteness, and what it all means: first versus second order logic · 2012-01-19T23:22:52.110Z · score: 0 (0 votes) · LW · GW

The argument wasn't specific to ZFC or HOL, it was intended to apply to any system that can have a proof checker.

Pointing out the Gödelian limitations of all systems with recursively enumerable axioms doesn't seem like much of criticism of the system of nth-order logic I mentioned. Now I have less of an idea of what you're trying to say.

By the way, I think he's using "full model" to mean "standard model." Presumably, the standard integers are the standard model that satisfies the Peano axioms, while nonstandard integers are any other satisfying model (see http://en.wikipedia.org/wiki/Non-standard_model_of_arithmetic ).

Comment by darmani on Completeness, incompleteness, and what it all means: first versus second order logic · 2012-01-19T16:55:44.786Z · score: 0 (0 votes) · LW · GW

If I understand correctly, you are saying that higher order logic cannot prove all theorems about the integers that ZFC can. That's a very uninteresting statement. Since higher order logic was proven consistent* with ZFC, it is strictly weaker. Second order logic, is, of course, strictly weaker than 4th order logic, which is strictly weaker than 6th order logic, and so on, and is thus much weaker than higher-order logic.

I've never heard it claimed that second-order logic has a unique model of the standard integers. Actually, I've never heard of the standard integers -- do you mean a unique standard model of the integers? I hope not, as we actually can get a unique standard model of the integers in a system with machine-checkable proofs. See "To Truth Through Proof" by Peter Andrews. It's omitted in the Google Books preview, but I believe the proof is on page 327; i lent out my copy and cannot check. (It, naturally, does not have a unique general model, so this not of huge practical significance.)

* I have not actually read that, and should not really say that. However, type theory with an axiom of infinity, which is strictly stronger than type theory without an axiom of infinity, which can express all statements of higher-order logic, has been proven consistent. Also, any proof in higher order logic can be trivially converted into a proof of nth-order logic for some n, which can be shown consistent in (n+2)th-order logic.

Comment by darmani on Completeness, incompleteness, and what it all means: first versus second order logic · 2012-01-17T03:22:51.112Z · score: 7 (7 votes) · LW · GW

Proofs in second-order logic work the exact same way as proofs in first-order logic: You prove a sentence by beginning with axioms, and construct the sentence using inference rules. In first-order logic, you have individuals, functions (from individuals to individuals), terms, and predicates (loosely speaking, functions from individuals to truth values), but may only quantify over individuals. In second-order logic, you may also quantify over predicates and functions. In third-order logic, you add functions of functions of individuals, and, in fourth-order logic, the ability to quantify over them. This continues all the way up to omega-eth order logic (a.k.a.: higher-order logic), where you may use functions of arbitrary orders. The axioms and inference rules are the same as in first-order logic.

Wait, I said nothing about sets above. Sets are no problem: a set containing certain elements is equivalent to a predicate which only returns true on those elements.

I also lied about it being the same as first-order logic -- a couple are added. One very useful axiom scheme is comprehension, which roughly means that you can find a predicate equivalent to any formula. You can think of this as being an axiom schema of first-order logic, except that, since it cannot be stated until fourth-order logic, it contains no axioms in first-order logic.

(My formal logic training is primarily in Church's theory of types, which is very similar to higher-order logic [though superficially extremely different]. I probably mixed-up terminology somewhere in the above.)

Comment by darmani on [Link] Duolingo · 2012-01-05T21:41:39.959Z · score: 2 (4 votes) · LW · GW

I used Duolingo for a few hours on its first day. (I used to TA for Luis, which helps for getting private invites, at least by knowing to sign up immediately.)

I've basically just gone through passing out of German lessons. This basically consists of taking a 20 question test, in which I translate sentences like "The woman drinks with her cat." and pray I don't make typos on three questions and have to start over. Except that all too often I give correct translations, but their checker isn't attuned to the flexibilities of German word ordering, or I use a different word for "chair." They said they'd teach it the distinction between "essen" ("to eat" for humans) and "fressen" ("to eat" for animals) so that their quizzes would stop recommending wrong answers, but I'm skeptical. I did another one just now, and it now seems to accept answers off by one character. I noticed this in a listening exercise not because I made a typo, but because the word for "is" is approximately homonymous with the word for "eats."

Oh, apparently Duolingo has something to do with translating the web, rather than just going through a bunch of German 1-level exercises. Maybe I have to repeat the above for the remaining 45 lessons before I see an option to do that.

Comment by darmani on Smart and under 20? Peter Thiel wants to pay you to not go to school. · 2011-12-30T23:34:39.587Z · score: 1 (1 votes) · LW · GW

I think a lot of people get put off because Andrew Hsu is first in the alphabet. A lot of them barely come up in Google searches (save for the Fellowship itself).

Comment by darmani on 2011 Survey Results · 2011-12-05T03:39:07.698Z · score: 4 (4 votes) · LW · GW

Keep in mind that many of these links were a long time ago. I came here from Overcoming Bias, but I came to Overcoming Bias from Hacker News.

Comment by darmani on Welcome to LessWrong (For highschoolers) · 2011-11-27T06:41:22.639Z · score: 0 (0 votes) · LW · GW

Looking at your other posts, it sounds like high school has tainted your opinion of schooling in general. However, college is Hard. A lot of people from my school make big bucks and find the rest of life extremely easy in comparison.

Comment by darmani on Welcome to LessWrong (For highschoolers) · 2011-11-27T06:13:46.476Z · score: 2 (2 votes) · LW · GW

For a lot of subjects, it's hard to beat school. You need constant feedback to improve in most areas, and good feedback is hard to come by without paying. I would recommend against trying to learn how to write a proof without a teacher as much as I would against learning piano. (I've tried both.)

I've been taught more things than I've learned on my own, but of the things I've learned poorly, most where things I learned on my own.

Comment by darmani on Tendencies in reflective equilibrium · 2011-08-10T05:16:20.401Z · score: 2 (2 votes) · LW · GW

I wouldn't have. The negative feelings from accepting the $5 would greatly outweigh the monetary value, even though I knew I almost certainly would never see the subject again...

...and would have been wrong; I ran into him last week.

Comment by darmani on Theory of Knowledge (rationality outreach) · 2011-08-09T16:00:04.345Z · score: 17 (17 votes) · LW · GW

ToK was my favorite class in high school, thanks to having an amazing teacher, one of two teachers in the school to complete a Ph. D. I've heard it said that, if you put Michael Vassar in an empt concrete room, he would soon start pontificating on the influence of the Enlightenment on putting people in empty rooms. I think the same is true of this guy.

We read parts of Man is the Measure and a 60s philosophy textbook , discussed the nature of causality, picked apart Max Weber's Verstehen argument, and reflected on the panopticon. We then did segments on each area of knowledge, with lively debates, discussions, and presentations.

It's still the case that ToK was subservient to our other classes. In March, the class turned into time to polish our Extended Essays (the 4000 word paper required of all candidates), some class time was spent starting a yearly tradition of painting the wall with the names of the candidates, and we had a party in place of a winter final.

I wanted to jump in and say that I liked ToK purely for the social bonding rather than the learning, and that the material covered was rather disjoint from LessWrong, but upon reflection, neither is true. My opinion on suburbia has been permanently altered from discussion of a documentary on Levittown, and it was in ToK that "utilon" became a regular part of my vocabulary. I actually became a reader of what was then Overcoming Bias while taking ToK, in part because I saw words like "ontology" and got the warm feelings from its association with ToK. I shared a few OB articles with my ToK teacher; he got a huge kick out of reading about phlogiston theory.

In summary: There are a lot of ways to make ToK good, and some of them don't look like LessWrong.

Comment by darmani on Tendencies in reflective equilibrium · 2011-07-20T17:12:49.786Z · score: 9 (9 votes) · LW · GW

My tendency is to assume that the homeless man would steal the $1000 via violent means, whereas the hedge fund manager would steal the $1 million using nonviolent deception. In addition to a belief that violent crime is actually worse, there is also the bias that it is easier to visualize. A homeless man stealing $1000 looks like a man pointing a gun at a cashier. A hedge fund manager stealing $1 million looks like a guy at a computer with a spreadsheet open.

Of course, I work at a hedge fund manager right now, so I have additional biases.

Comment by darmani on Rationality Market Research · 2011-07-14T20:58:34.675Z · score: 1 (1 votes) · LW · GW

I look way too young to be interviewing people, so I personally would not use that one; it would indeed "smell fishy," and, seeing as I actually don't interview people, I would not be able to do it naturally. Otherwise, it would be fairly normal to talk about interview candidates if they're on your mind; just a few months ago, a professor called me as I walked by his office and started wondering aloud whether to hire a guy (though the professor was very much not a stranger).

On the other hand, it's actually true that I hear people talking about joining a rationality club or being rationalists, so I could probably use the other two off fairly naturally.

I'd agree though with your overall concerns; these might work better as the first thing you say after "Hi, I'm Ray" (in a context where that's appropriate) than as true openers. I've nonetheless definitely seen these kinds of lines work in starting conversations with random strangers.

Comment by darmani on Rationality Market Research · 2011-07-14T18:23:01.951Z · score: 7 (7 votes) · LW · GW

I'm curious how varying the "opener" will affect the responses. Here are a few off the top of my head:

Opener #1: "Hey, what would you think of someone who calls themselves a rationalist? I was interviewing this guy the other day, and I asked him what words he'd use to describe himself, and that was the first thing out of his mouth. I've never heard that response before, and don't know what to think."

Opener #2: "Hey, so, I met this guy the other day, and he was going on and on about how he was a huge rationalist. That's not really something I've seen anyone do. What would you make of someone like that?"

Opener #3: "So, I have a friend who's thinking about joining/starting a 'Rationality Club.' I don't really know what to say about that, other than that it sounds a bit [pause] unique. What kind of people would you expect at a 'Rationality club?'"

These are designed to do several things. First, by describing a specific incident, it puts them in near mode, and gets them to imagine what they'd actually think. Second, they have a livelier and more conversational phrasing, which should help in getting people to open up. Third, by distancing yourself from the label, it makes them free to not be polite to the word.

These are completely untested, though I might give them a try tonight.

Comment by darmani on St. Louis, Missouri Meetup - now happening every week! · 2011-03-24T15:07:15.788Z · score: 0 (0 votes) · LW · GW

Good luck with this! If you actually meant May, I might be able to come.

Comment by darmani on Plant Seeds of Rationality · 2011-03-11T04:18:39.561Z · score: 1 (1 votes) · LW · GW

Nah, already checked -- it only archived the front page of the forums. That's actually how I found the parent website, though.

Comment by darmani on Plant Seeds of Rationality · 2011-03-10T23:36:20.438Z · score: 2 (2 votes) · LW · GW

It appears the parent website of the forums took them down a few years ago, so probably not.

The biological argument is that it doesn't necessarily work, and all the societal changes I advocated with it for implementation had much stronger things against them.

Comment by darmani on Plant Seeds of Rationality · 2011-03-10T22:53:07.175Z · score: 6 (8 votes) · LW · GW

When I was 12, I, in my infinite wisdom, decided that eugenics was necessary to save humanity, and went online to debate my belief, where I was promptly defeated by a biologist who knew what he was talking about.

It took me a while to admit to being wrong, and I never did so publicly. Instead, I kept trying to patch the holes in my position, even as they were being exposed at an incredible rate.

Nonetheless, I regard this experience as formative in becoming a rationalist. "Planting the seed of rationality" may be successful, but you will often never have the satisfaction of knowing when it works.

Comment by darmani on The Fallacy of Dressing Like a Winner · 2010-12-26T21:33:10.826Z · score: 10 (10 votes) · LW · GW

Richard Feynman gives many examples in his famous essay "Cargo Cult Science": http://www.lhup.edu/~DSIMANEK/cargocul.htm .Significant discoveries have small p-values, so clearly, if you do something and run some numbers and a small value pops up, your experiment is significant, right?

Comment by darmani on Pittsburgh meetup Nov. 20 · 2010-11-16T21:38:33.359Z · score: 1 (1 votes) · LW · GW

I will, of course, be there.

Comment by darmani on A Player of Games · 2010-09-30T04:33:55.758Z · score: 0 (0 votes) · LW · GW

Teehee .... "End P of O" is just an abbreviation for the code to actually end a p. of o. (card to self for talking about the rules).

Is it now? There are plenty of times when you might be tempted to say "Ouch!" or "Epic fail," but when the rule of Belittlement is in effect...

Comment by darmani on A Player of Games · 2010-09-29T17:03:19.135Z · score: 1 (1 votes) · LW · GW

I think the difference between board games and life situations is how much more tractable games are to this attack. You have excellent information on how the state can evolve (perfect information, once you know the "event card" deck or whatnot if there is one), and can quickly get a rough estimate of the probabilities of each. While even a DnD beginner can figure out what actions to take to maximize his expected damage output, his actual character, who has seconds to do what the player has an eternity to think about, could not.

And many games have limited interaction between players, in which case strategizing is quite simple.

Come to board game club sometime, and I'll show you games where you find yourself able to think significantly ahead (in a quite similar style to how you calculate ahead in chess) within a few turns of playing. Except many of them have low-enough long term consequences for most moves that you often only need to calculate ahead by one.

Comment by darmani on Coding Rationally - Test Driven Development · 2010-09-28T18:51:21.509Z · score: 1 (1 votes) · LW · GW

It's confusing to think of TDD as its own rationality technique: testing your belief that a piece of code works is not fundamentally different from testing any other belief. Okay, so that part is just running unit tests once. Since whether a piece of code works is a different belief from whether that piece of code with a few modifications works, for efficiency, since writing tests is work, you need to keep that tests around and rerun them. So, that's unit testing. TDD is just writing your tests beforehand, which makes a difference in the process of designing software, but not really in how confident you should be that your code works.

Something more interesting to think about is how much information a test gives you that your code works. You can often tell from eyeballing whether switching an integer from positive to negative will make a meaningful difference whether the code produces the intended result.

This really turns into an approximate, poor-man's version of proving code correct, which typically proceeds by breaking down code into its paths and checking each against the mathematical model.

Which reminds me, I have to go prove a few Standard ML functions work by Friday.

Comment by darmani on A Player of Games · 2010-09-26T21:41:14.337Z · score: 0 (0 votes) · LW · GW

Woah, you're kidding, right? I've never not played with a "no talking outside of Point of Order" rule. How would it even work if you can say things all the time, and thereby avoid learning when you're actually required to say them?

Oh, and I get a card for saying P of O during P of O, and another for talking about the rules.

Comment by darmani on A Player of Games · 2010-09-24T17:28:25.060Z · score: 4 (4 votes) · LW · GW

I concur. As a member of my school's board game club, I have learned over two dozen games within the past year, and am frequently competitive during the first game.

The core of most games is fundamentally the same: You have a position; your actions bring the game to a different position. Once you understand what moves you can make and begin to understand the benefits of various positions, you can begin to strategize using a human version of the simple Minimax algorithm.

There are other reasons why having an experienced player teach the game can be better for meta-strategizing skills: They will give examples of correct play and are more likely to enforce the rules correctly, enabling faster learning.

Additionally, a quite common scenario is to have a game with someone who's played three times, someone who's played once, and two newcomers (or thereabouts). In that case, it isn't hard for the first-timers to be competitive.

Comment by darmani on A Player of Games · 2010-09-24T16:53:18.798Z · score: 1 (1 votes) · LW · GW

I'd have been shocked if we didn't.

Comment by darmani on A Player of Games · 2010-09-24T16:20:34.870Z · score: 1 (1 votes) · LW · GW

Saying the M-word when inappropriate, 2-card penalty.

Comment by darmani on Book Recommendations · 2010-08-10T06:33:51.707Z · score: 1 (1 votes) · LW · GW

For a lot of examples, have a look at the Wikipedia article on Easter Island. Though the evidence he gave seemed incredibly strong, apparently it's quite disputed that they suffered a pre-colonization collapse at all.

Comment by darmani on Book Recommendations · 2010-08-10T04:16:26.256Z · score: 4 (4 votes) · LW · GW

I highly recommend Jared Diamond's Guns, Germs, and Steel and Collapse. From a rationalist perspective, it is enlightening for its strong scientific approach to history -- a refreshing change from school history textbooks which are written like an exceptionally dry novel, with a single canonical narrative.

Comment by darmani on Five-minute rationality techniques · 2010-08-10T02:35:52.596Z · score: 7 (9 votes) · LW · GW

Candidate 1: "If a trillion trillion trillion people each flip a hundred coins, someone's going to get all heads." ("If a trillion people each flip a billion coins" might be a stronger meme, though extremely inaccurate.)

Candidate 2: "Knowing the right answer is better than being the first to argue for it."

Candidate 3: "If it moves, you can test it."

Comment by darmani on Harry Potter and the Methods of Rationality discussion thread, part 2 · 2010-08-04T06:22:20.227Z · score: 1 (1 votes) · LW · GW

Apparate to wherever it's being stored, Horcrux it, apparate out. Even if there is an alarm to set off, you're in virtually no danger.

Comment by darmani on Harry Potter and the Methods of Rationality discussion thread, part 2 · 2010-08-02T00:22:06.516Z · score: 4 (4 votes) · LW · GW

What would the incentive to become a traitor before the battle of Chapter 33 be? Before Quirrell added the ability to switch sides, you'd just be helping your army (which you've already developed a bond with in the first battle), and therefore yourself, lose. I'd expect this to strongly outweigh the fun of being a spy.

I just Googled for "airsoft betrayal" and "paintball betrayal." I found no stories of similar events in either sport. (I did however find one person hypothetically talking about betrayal in laser tag, even though many/most systems ignore friendly fire.)

Comment by darmani on Book Club Update, Chapter 3 of Probability Theory · 2010-07-27T07:12:04.230Z · score: 0 (0 votes) · LW · GW

(Heh, I'm pretty sure being a college sophomore makes me an amateur too.)

Yep. Cox's theorem implies only finite additivity. Jaynes makes a big point of this in many places.

I'm not asking for an isomorphism in the reasoning of choosing a set of axioms. I'm asking for an isomorphism in the reasoning in using them.

For large classes (all?) of problems with discrete probability spaces, this is trivial -- just map a basis (in the topological sense) for the space onto mutually exclusive propositions. The combinatorics will be identical.

Comment by darmani on Book Club Update, Chapter 3 of Probability Theory · 2010-07-23T17:14:27.242Z · score: 0 (0 votes) · LW · GW

Do both systems satisfy the Kolmogorov axioms? One of them is countable additivity, right?

Of course, Kolmogorov's is hardly the only such development. My question is: Is there an isomorphism in reasoning that also serves as a proof of the equivalence?