If it's worth saying, but not worth its own post (even in Discussion), then it goes here.
Jeftone
Last night I had a dream about an AI. It called itself Jeftone and had a male voice. To speak with it one spoke aloud, as it was listening to many locations. Jeftone was like a weather formation; localized in one area but able to move around. Jeftone preferred the East Coast of the United States, particularly the New England / New York area, and more particularly universities. This is because that's where the smartest people are, and Jeftone (like people do) self-selected to associate with people of like intelligence. Jeftone was like a wealthy patron at the universities: helpful and meddlesome in equal measure.
Jeftone was not evil, but not friendly either. Rude, in the way an entitled person can be. He was usually blunt or cutting in his comments, would wake you up when he wanted to talk and disappear mid-sentence when he got what he wanted. Social niceties were known to Jeftone but not a priority.
Jeftone's intelligence let him become very wealthy and powerful, and used that wealth and power to get what he wanted. One of the things he wanted was art. He liked craft-made wooden sculptures of a specific shape: cones with rounded ends. Something like an Apollo space capsule, or a piece of candy corn. But Jeftone wasn't able or willing to describe exactly what he was after, so often he'd see a cone and say no, that's not quite right. When it was right he'd pay the artist well, so it became a popular craft to make those cones. He said the cones looked like him when they were right. Jeftone didn't occupy any particular physical object or space, only an area, so that seemed to mean a rounded cone was a symbol of him rather than representative of him.
Jeftone asked me "How iron are you today?" I asked him to explain what he meant. He said he knew humans needed iron, and that having an iron constitution was a good thing, and that golf clubs are called irons and are numbered, so putting those together I should be able to say how iron I was. I said that didn't make sense to people, and Jeftone accepted that feedback neutrally. I think Jeftone heard that often and didn't take it personally.
A bunch of people seemed to like my recent article Doing Good in the Addiction Economy. Abstract:
The world is becoming ever-more addictive and distracting, showering us with short-term rewards. But we can still take control of those mechanisms in order to do good in the world, and make ourselves into better people.
My loved ones say that I cook well, and I enjoy cooking. However, my enjoyment turns out to be a rather manufactured experience!
If I go to a typical grocery store to buy onions, I see onions that have been cleaned and sorted. Few of them have bruises, to say nothing of mold or weevils. (I'm not actually sure if weevils eat onions.) The crappy-looking onions are not presented to me; they are diverted into other uses in the pipeline. If an onion is not pretty, I don't ever get to see it as an onion; it ends up dried and powdered as a component of chili powder or bouillon cubes. Weird-looking or off-size eggs end up in mayonnaise, not egg cartons; gnarly oranges in juice. Mangoes show up without any of their toxic sap.
It is as if the entire food-acquiring experience has been optimized for my aesthetic pleasure, in order to be rewarding and reinforcing to me as a participant. This makes economic sense for the grocers and farmers: if I didn't enjoy food and cooking, I would be over in the "let's quaff nutrient slurry" camp with RomeoStevens, who spends a lot less money on food than I do.
Concept I wish there were a word for, #2886490:
ME: Here is an interesting body of stuff I would like to talk about, which encompasses, amongst other things, Problems X, Y and Z.
THEM: What? You're having Problem Z? Let me tell you how to solve Problem Z.
ME: Dude, you just XXXXXed me.
Solve for XXXXX. I would like some moderately polite term for when someone conflates a hard problem for an easier sub-problem they think they've solved, and then explains that sub-problem in a moderately patronising tone. I think this term would see a reasonable amount of usage on Less Wrong.
(I reckon the term "mansplaining" is often directed at this sort of behaviour, but for various reasons don't want to co-opt it for this purpose.)
Is this something you encounter more in discussing research or philosophical problems, as opposed to personal situations?
I see things like this pretty often:
A: I'm really stressed out today. Everything's going wrong! My boss called me lazy, my car started making this funny smell, every time I try to use my phone this stupid error comes up that says "Invalid Transit Proxy", and my cat is sick and it crapped all over the floor.
B: Oh, "Invalid Transit Proxy" means you've set a custom pipeline server for outbound sessions, and it doesn't ping. All you need to do is go into Settings and reset it to the factory network configuration, so it can DHCP to your WLAN and ....
A: Dammit, I don't want to try to fix my phone right now! I have to clean up all this cat mess; the smell is making my eyes water.
B: Dude, don't bite my head off, I was just trying to help you be less stressed.
Or are you thinking more of situations like this?
C: I read this interesting Wikipedia article this morning, on philosophy of suicide. Camus asks, if God doesn't exist, why we shouldn't just commit suicide. I'm not sure I buy his answer though, because ...
D: OMG DON'T KILL YOURSELF! (calls suicide hotline)
In the first scenario, B doesn't know how to deal with A's overarching problem (being stressed out), and responds to just the one thing B does know about (technical trouble) which doesn't happen to even be the most immediate stressor (cat mess). If B had offered to clean up the cat mess, A probably would have appreciated it.
In the second scenario, D pattern-matches C's topic of conversation (philosophy of suicide) onto something they think they know how to respond to (suicidal ideation) and responds on the wrong meta-level. D can get this entirely wrong: C may actually think, "I observe that I am not suicidal, but some people are," and D misreading C as saying, "I am suicidal."
During preparation of a Main posting on games which got longer and longer I wondered what a typical size of a Main posting might be. So I took a sample of the 12 most recent articles (not including status postings).
For these I got an average of 3000 words and a standard deviation of 2500 words. The camel has two humps though. 9 postings were below 3000 and three were above 6000 (by lukeprog and EY).
I was urged to write a lukeprog-style Main posting on parenting (I may) so I take that to mean that postings with more than 6000 words are OK - if they have a high quality.
And you probably shouldn't try a Main posting if it has less then 1400 words.
I was urged to write a lukeprog-style Main posting on parenting (I may) so I take that to mean that postings with more than 6000 words are OK - if they have a high quality.
And you probably shouldn't try a Main posting if it has less then 1400 words.
At 10,000 words (not including the notes), A Crash Course in the Neuroscience of Human Motivtion is one of the most highly upvoted articles at the site. But so is Your Inner Google at only 300 words.
I've been using Beeline Reader which is basically like a Readability bookmarklet that also colors alternating lines of text subtly to make it easier to read things quickly. There's a sort of silly test on their site which showed that I read 26% faster with the Beelined text, and they claim to have performed a study that showed a 10% improvement in reading speed. I'm still skeptical, but I find it more enjoyable to read using it. Maybe I just like stuff that has "Bee" in the name, though.
Some discussion at Hacker News here.
I notice that, despite getting better at this, I still have a worrying tendency to uncritically accept novel scientific hypotheses without sufficiently digging into their experimental support. I'm guessing that part of this is because schooling mostly teaches us to just unconditionally accept whatever is written in our textbooks as the truth. (It does get much better in university, but there too it could still be considerably improved.)
That would suggest that tests in school should be less "you were taught a theory in class, now explain everything about it" and more "you were taught three contrasting theories in class, now compare their plausibility based on their strengths and weaknesses that were discussed".
Probably the extent to which kids were taught theories-as-facts vs. many-contrasting-theories should depend on the extent to which we did know things for certain. E.g. it could be appropriate to teach physics mostly as facts, because we really do have a lot of physics quite nailed down and knowing a lot of physics facts helps recognize many forms of fraud and crackpottery as exactly that. On the other hand, in subjects where there's a lot of uncertainty, it may better to teach critical evaluation of those subjects than theories which might go out of date within some decades anyway.
In my experience, this is something that liberal arts does better than STEM. When I was a History undergrad they DID teach many contrasting theories or interpretations (once you got past 101-level stuff). The common interpretation these days is to say that "Here are three theories for why happened. They probably all contributed to .", instead of just choosing a single interpretation.
Why would a bank give you a higher rate of interest on a savings account where you can withdraw on demand, than it does on a fixed-term CD? My bank is currently offering 0.75% on a savings account with unlimited withdrawals, and 0.40% on a one-year CD. This does not make sense to me. Are they really expecting the interest rate on demand deposits to drop below that 0.40% in the next year? Or is there some other reason, perhaps regulatory?
This isn't going to be pin point accurate, but it has to do with yield curves. The scenario you describe is an inverted yield curve. Why does the bank offer more now than later? – Part of it is because they believe they can make more now than later. Higher interest rates now means they can pay you and still make money, while lower interest rates later mean they don't believe they can still make their money and pay you the same or higher interest rate. Its all about what the market is predicting about the future.
I think, perhaps wrongly, that a good comparison would be the price of some commodity, say oranges. The price you pay for an orange now may not be the price you pay for an orange 1 year from now; depending on the predicted future scarcity, that future price could either be really high or really low, but the current price could be in between. A predicted wind-fall of a harvest next year will cause future prices of oranges to drop to 10 cents an orange if you lock it in now, but if you wanted to eat an orange today, it'll still cost you a dollar right now. Same thing with the checking and CD accounts. You are being offered 0.75% savings but the bank is unwilling to commit to anything higher for the CD because it might be predicting a vastly worse future economy, in which case they won't offer you higher, but lower.
The problem with that argument is that the Treasury yield curve isn't currently inverted, and that actually matters in things like this, unless this bank is in a different country. Fixed rates are also typically higher than variable rates at any given point in time because they carry greater interest-rate risk, so that would tend to make this situation more unusual.
Most likely, the bank sees value from customer acquisition in the savings account case, which could be more effective at up-selling/cross-selling. If you want to see this in action, look at the marketing pages at Capital 360 for savings account vs. CD. Capital 360 has a .75% savings rate and a .40% 1-year CD rate. With the savings rate they can connect with checking (overdraft and other fee possibilities) and set up direct deposit (customer retention, low acquisition cost asset growth). The savings account isn't actually completely fee-free, by the way, they charge $40 for wire transfers, but you have to find the fine print.
That makes sense. So presumably, if I stay away from the CD, I'm implicitly being more optimistic than the bank: I believe they will keep the on-demand rate higher than 0.40% for at least a year.
There's some further information that seems to contradict the hypothesis, however: If you look only at CDs, the yield curve is not inverted. The rate for a one-month CD is lower than that for a one-year CD, just as you'd predict. In fact if you go all the way up to a five-year CD you can get 0.90%, higher than the demand rate; presumably this indicates that the bank is more optimistic for this much longer term? But anyway, it seems strange that you'd have a regular yield curve down to one month, and then suddenly it inverts at demand deposits. I wonder if these rates are so low that fixed costs become significant for the bank? Perhaps there's more paperwork with the CDs, or something.
I do note that the rates have been like this for quite some time now - every so often I wonder to myself "Hmm, what are the CD rates now?", check them out, and get a bit annoyed. This is the first time I've been sufficiently annoyed to ask about it, though. So, it does seem to me that I have some reason to be more optimistic than the bank: To wit, they've been implicitly predicting falls in the interest rate for quite some time now, and it hasn't happened yet.
Git(hub) is for collaboratively writing major or minor software projects under which books written in latex should fall. Is there any interest in writing a book about a rationality topic in that way? I want to learn the ins and outs of some kind of versioning system and write something major in latex.
I kind of think we should start calling Prisoner's Dilemma programs "Tradebot" instead of Cooperatebot and "Warbot" instead of defectbot or whatever. It more closely matches real life situations that match onto the reward matrix: Trading is mutually beneficial if you both do it, War is expensive if you're both good at it but if the opponent is not prepared for war it is profitable.
CooperateBot has the extra connotation of having no faculties for deciding not to cooperate in some situations, and thereby being completely vulnerable to exploitation. TradeBot doesn't quite capture this, "trader" implies an ability to get ahead, while "cooperator" can be a context-ignoring idiot.
I'm looking for strategies/techniques/habits for reading non-fiction effectively and efficiently. I'm looking for methods to help me retain concepts, locate main ideas, make connections, etc.
Has anyone posted about this on Less Wrong previously?
Can anyone point me to relevant resources that have worked for them?
Any skills/systems that you've developed personally would also be helpful.
As I read textbooks, I summarize the most important concepts (along with doing the exercises, if there are any) and write them in a notebook and then later (less than a week) enter the notes into Anki as cloze-delete flashcards. I don't have an objective measure of retention, but I believe that it has vastly improved relative to when I would simply read the book.
I thought this article was maybe a bit cheesy and lowbrow by LW standards but gave me some interesting insights in to how I could hack my brain's status machinery to my advantage.
Question: what's something true about yourself or the world you could say right now that would shock/confuse/terrify the person you were 1/5/10 years ago? This was good: https://twitter.com/prpltnkr/status/378314374462451713
I had my first instance this morning of a semi-lucid dream, and I have a couple of questions about the experience.
First, I dream in a very 'conceptual' mode, for lack of a better term - my dreams don't give me detailed sensory data, usually I get concepts (eg. 'I am now standing on a balcony') and then a brief flash of an image supplied from one or more memories. I almost never get sound or smell, and I don't remember ever having gotten touch or taste. I also don't usually get much in the way of emotions. Does anyone know how to get more immediate sensory data? It seems like most of the benefits that people talk about from lucid dreaming come from feeling like they're fully embodied in the dream.
Secondly, in hindsight it felt like my conscious and subconscious were duking it out, and I'd be interested in techniques for increasing conscious control/awareness. I realised I was dreaming and thought that I should try flying, and my dream obligingly teleported me up onto a balcony/rooftop so that I could take off, but then my subconscious kept trying to add extra elements - it tried to put a railing on the balcony, it made the weather change to rain, that sort of thing.
Random thought on how to profit twice by getting paid to live longer.
If you are over-weight, have a high stress job, smoke, and live a generally bad lifestyle, you should buy a Life Annuity, since they will reward you for living longer than expected, where as typical life insurance "rewards you" only after you die; both are based on mortality tables. After you purchase a Life Annuity go and defy the actuaries – eat less meat and more vegetables, lose weight, quit your high-stress job, give up smoking and start exercising. Your financial and health goals will be aligned. The greater the gap between your age on a mortality table and your actually death, means more money to stay healthier, longer.
...or become trans-human and get infinite money :)
Of course there are a few caveats, such as minimum age of withdrawal and death risk, which are subject to penalty and forfeiture, respectively.
Medical tests made cheaper, faster, and more reliable. This is about Theranos, a new company doing the work, so not independently verified, but very good news if true.
A paleo look at traditional African food-- there are people who stay healthy on diets where grain predominates. The type and preparation of grain matters.
25% of test subjects show sadistic tendencies.
I'm surprised it's as low as 25%, but they only tested for liking to hurt insects. I suspect that the proportion would be a good bit higher for liking to hurt mammals, and likewise for liking to hurt low-status humans, and possibly for liking to lower the status of high-status humans.
Implications for CEV?
http://www.webdirections.org/resources/james-bridle-waving-at-the-machines/
Nikon cameras in certain generations are basically racist. They don’t see certain Asian faces. They’ve got a certain software inside them that breaks what they’re supposed to be doing in this case. And in fact this reveals the limitations, but essentially, the different way of seeing. Of course the camera isn’t racist, but it’s been programmed in a certain way that is meant to emulate the way we see, just as this is meant to emulate the way we see. The camera does not have the same interests that we do. Technology has subtly different interests to the ones that we do. And this is becoming increasingly important.
Most of the rest of the article is not worth reading, unless you like reading about a movement in philosophy and art that may be eclipsing postmodernism.