code4lib jobs: (Temporary) Associate Information Systems Officer - Unite...
The United Nations is looking for a Java developer to fill a temporary contract working on DSpace repositories. If you're interested, please
A black and white photograph of a giant dust storm...
A black and white photograph of a giant dust storm in 1937. Photographed by Chris Johns, National Geographic
My walk along Coney Island Avenue in Brooklyn. A...
My walk along Coney Island Avenue in Brooklyn. A seven-mile stretch that includes South Asian, orthodox Jewish, Turkish, and finally Russian
New Yorkers Organize 24-Hour Read-In for Libraries!
Remember when I told NYC people to stay tuned for announcements about library advocacy opportunities during this spring budget cut season? W
A close view of the puffer fish found in Florida’s...
A close view of the puffer fish found in Florida’s warm waters, January 1922. Photograph by L. F. Williamz, National Geographic
A shelter made of antlers at Yellowstone National...
A shelter made of antlers at Yellowstone National Park. Photograph by Edwin L. Wisherd, National Geographic
Dilbert comic strip for 06/01/2013 from the official Dilbert comic strip...
The Official Dilbert Website featuring Scott Adams Dilbert strips, animation, mashups and more starring Dilbert, Dogbert, Wally, The Pointy
Clever Paper Cut Designs Feature Animals and Their Snacks
People often say You are what you eat. Well, Oregon-based artist Wendy Wallin Malinow thought it would be interesting to see that phrase in
Amish Raw Milk Farmer Vernon Hershberger Wins Big in Court
(Above: Farmer Vernon Hershberg between two of his lawyers, Glenn Reynolds and Elizabeth Rich.) “Food [...]
A German shepherd is accepted for sentry duty by...
A German shepherd is accepted for sentry duty by the Coast Guard, January 1941.Photograph by J. Baylor Roberts, National Geographic
Dilbert comic strip for 05/25/2013 from the official Dilbert comic strip...
The Official Dilbert Website featuring Scott Adams Dilbert strips, animation, mashups and more starring Dilbert, Dogbert, Wally, The Pointy
This is also why I don't really want my kids using YouTube. It's a wretched hive of scum and villainy.
Yep, this pretty much sums up how I feel about most of the internet most days. It's a downward spiral.
It's why I don't go arguing with people I disagree with in their own space unless they ping me in. My causes are never helped by going into the other side's territory and yelling at them for how stupid I think they are. Because they think I'm stupid, and nothing ever gets anywhere that way.
And it's why I now have restricted comments settings here on G+, and why I'm going to close the comments on this post as soon as I share it, because the YouTube-G+ integration still means waves and waves of trolls on my G+ shares of YouTube videos and I don't have the time or energy to deal with it.
BTW I found this while looking for an older song where a lady in a pink dress sings about YouTube comments and now I can't find it...see, it'd be useful to leave comments open in case someone else has the link, but I really just don't want to deal with TuberTrolls today.
ETA found the older one: https://www.youtube.com/watch?v=uz2jbCJXkpA
One of the roleplaying settings I am slowly fleshing out includes a culture whose ancestor worship proceeded far beyond the norm, and because they live in lush valleys surrounded by very inhospitable mountains, they have remained isolated enough for their practice of conducting necromantic rites on their deceased to flourish. Their intentions are not evil, and their extremely communal culture supports a life and undeath of service to the community.
The normal process for converting them to undead communal servants is as follows: upon death, the body is transferred to the custody of the religious order that performs the necessary ritual cleansing (but not the necromancy itself; the religious order is a silent one and does not cast vocalized spells of any kind). The cleansing takes one year, during which the flesh decomposes and the bones are preserved. On the day after the one year anniversary of death, the body is given over to the necromancers, who animate the bones and bind the skeleton to service. It is then transferred to its original family members for basic care, which includes maintaining its clothing, but its orders are pre-programmed: labor and/or defense. I hope I don't really need to go into great detail about the advantages of skeletons over other kinds of low-level undead, though I am open to some suggestions in case you have something you can substantiate.
Because the valleys are only so large, space constraints might eventually become a concern with a growing number of ambulatory skeletons. Potential solutions to this include periodic decommissioning to maintain desirable numbers, with the understood risk that recommissioning skeletons is potentially impossible and certainly much slower than having them at hand; or keeping them in deep storage of some sort, perhaps in compact bundles, at such point as no living relatives remain to serve as caretakers.
Additionally, it should be noted that the number of walking dead will inevitably come to outnumber the living. The psychological and social consequences of such an arrangement are unknown, and potentially worth exploring elsewhere. I am certain that without some religious decree to maintain positive birthrates, natural birthrates would plummet in this environment. This speaks again to the need to cull the numbers of skeletons at regular intervals.
As suggested by the title, however, this post seeks to examine some of the economic impacts of such a labor and defense force. What primarily drives their economic output in these regards are factors such as their overall durability, their capacity for autonomy, and the raw potential quality of their work. These determine what tasks suit them best and which can be left to the mortals. You might notice this is very much a question of automation.
The question of durability is one whose answer points to the set of broadly interchangeable tasks to which skeletons are applied. If they are fragile, then your options for moving heavy objects are limited to throwing more skeletons at the problem. Nowhere in the D&D literature do we encounter skeletons with superhuman strength used in such a context. Without being an expert on necromancy, my intuition suggests that getting higher strength output from a skeleton is a matter of some combination of raw material input (e.g., make skeletons from bigger things) and magical input, where more magic expended equals greater strength.
Skeletons tend to be fairly weak cannon fodder under normal circumstances, neither particularly strong nor fast. A purely economic consideration, then, is that skeletons in our mysterious valleys will vary in strength mostly according to their bone size. Only some skeletons would warrant the extra expenditure to make them stronger than average (or, for that matter, to improve any of their performance dimensions). Thus the typical calculus for skeletons as heavy lifters will include the physical dimensions of the object and the availability of any force multipliers (simple machines) for a given task.
Along this dimension, skeletons are perfect in situations that can benefit from arbitrary addition of more skeletons but less so in situations where the space in which the task is to be performed is smaller than the number of skeletons that would be required to do it. This explains why they are well suited to defense work but perhaps not for close quarters lifting tasks. Anything that falls between these two extremes is fair game, if you can get the other two dimensions right.
The other question of durability is how long they can last. As magically animated collections of bones, skeletons tend to last as long as the magic can bind them. Remove or damage too many parts, and the skeletons lose that animating force. One should imagine skeletons to be limited in absolute durability while still able to take a great deal of punishment. Do they expend hit points when they undertake feats of strength or speed that fall beyond the strength of their magic? In non-combat situations, what causes a skeleton to wear out, and can they be repaired?
What do skeletons do after they've been given an instruction? How many instructions can they hold? What do they do once they've completed an instruction or set of instructions? Here the question is one of overall capacity. Encountered under normal circumstances, skeletons tend to perform their one task relentlessly. Usually they are guards, keeping out unwanted guests such as adventurers. Outside the direct oversight of their controller, it is doubtful if they even possess enough magical awareness to distinguish between various threats. Again, it seems a simple enough task to set them loose in scripted patrol patterns with some alert mechanisms nearby as a series of defensive quick reaction forces, but what about various labor applications?
Consider various human powered machinery, things that require neither continuous access to power nor locations fixed by the presence of rivers or good access to wind. Treadle powered lathes and sewing machines. The winch at the water well and various portcullises. Hand-drawn carts, wheelbarrows, and the like. These require only the ability to grip, push, pull, and perform gross motor tasks, and can be easily performed by skeletons with minimal oversight.
The question of quality is one of discernment on the part of the skeleton. That is, at what point does the skeleton consider the task complete? With defense, the question is simple: either a hostile creature is alive within range of its magical sensors, or it isn't. The quality assurance will depend on the particulars of the magic or the presence of controller capable of ensuring the desired quality. In other words, for anything but tasks involving moving loads from point A to point B or sweeping village perimeters, the cost of quality assurance will be high.
Undead as economic automation
So let's say a typical village of 300 people has a typical outlay of skeletons. Based on some very incomplete Google searches, let's estimate that about 1/3 of those are children under 10 (so 100), half the total population of an age to perform labor and defense (150), and the rest (50) are too old for harder work. The elders in this scenario will fall into a few groups, some of which could overlap. Some will be infirm, some will be able to manage or oversee the activities of a number of skeletons, and others will serve as caretakers of some number of skeletons. Crucially, one could imagine a scenario in which an overseer directs caretaking activities performed by a handful of skeletons, while the other skeletons perform heavier labor.
My guess, based on the above, is that the village could support somewhere on the order of its own population without too much difficulty. The skeletons would be put to work in a variety of primary occupations and a number of secondary occupations. For instance, not all of the skeletons need be devoted to defense as a primary occupation, though I imagine most or all of the remaining skeletons would have defense as a secondary occupation. As examined above, the other primary occupations will include construction tasks, hauling tasks, various farming tasks, mining, and employment as basic family servants.
With skeletons providing some combination of tedious and/or hard labor in the village, the villagers themselves would be freed up to perform other tasks, especially cultural pursuits.
Are there limits? What else could they do?
Up to this point, we’ve dealt only with skeletons, who are easy enough to create, automate, and control. But I see no reason a people with generations of experience in practical necromancy might not push the limits of what can be done. Why not reconfigure the bones into machinery? Assembly lines? Factories and manufacturing plants? Combines, threshers, basic tractors? All of these are within the realm of possibility with the right incentives. The biggest risk one runs in proposing such a thing is the disgust response of the villagers. But then again, such things rarely arrive whole-cloth, instead arising through more organic means.
Here’s how it starts: the skeleton of I-- D-- has been in constant use for about a decade. The magic binding some of his parts together is in danger of unravelling. But out of a sincere desire to honor this ancestor, an engineering-minded young woman asks why I-- D-- can’t have wheels attached. And together with a necromancer, they come up with a way to achieve it. Suddenly the useful undeath of I-- D-- has been extended, and his skeleton can be put to use doing other things for which wheels are more suitable. Before you know it, the engineering-minded young woman and the necromancer have begun cobbling discarded parts together, giving some skeletons four arms instead of two, or joining them together in stationary rows.
Because the work, to keep pace, is getting more specialized and skilled, it’s no longer sufficient to oversee the machines with humans. And so the necromancers look further up the chain of available undead and figure out that they can chain together other kinds of undead to create additional processing and management capacity. The result is potentially a set of computational devices, powered by necromantic magic, by which most, if not all, economic activity is performed by the undead. In short, you get technological unemployment without even having gone through proper industrialization.
I leave it as an exercise for the reader to determine if I’ve missed anything important, though I hope I at least got the obvious things.
With that in mind, a brief typology, with an introduction to common problems:
(1) Policy engines.
These systems are basically deterministic, and are are simple enough that you could implement them in a spreadsheet. You take some numbers and booleans and input them into the system, and it outputs a particular result.
So, for instance, "if a user has documentation of citizenship, and makes less than $20k per year, and isn't otherwise in the system, then approve their welfare application." Simple, and (naively) not particularly failure-prone.
Inevitably, you will run into problems that require human discretion. People have their identities stolen, or enter incorrect information, or have problems that can't be reduced to a simple rule. What if you know your estranged husband is making money, but he's moved elsewhere in the state and applied for welfare, and you're still listed as married? You get screwed.
(2) Analytic aids.
These systems are designed to take large volumes of data and reduce them to a human-readable format. These range from extremely low-level query languages like Google's Dremel, up through GIS packages, to high-level analytic software produced by companies like Palantir or SAP. (Or the NSA, for that matter.)
The main problem problem here is garbage in, garbage out. If your data is fouled by statistical noise (see, e.g. http://goo.gl/UM1Jdb), then you're going to get bad results. Worse, if you ask a malformed question because you don't understand the noise in your data, you will get nonsense results. And even if you don't, paraidolia is common: people pattern-match in even very sparse datasets, meaning that people will see trends where none exist.
You might not think that you've entrusted your life to these sort of analytic aids, but for critical issues like zoning, limited business licensing, or tax auditing, you almost certainly have.
(3) Machine learning.
This seems complicated. But conceptually, some of it is pretty simple. I'll use support vector machines as an example, because they're pretty easy to explain. Here's a simple example:
You have a huge biometric database which includes things like height, hair length, hip and waist measurements, and inseam length. You take all of these numerical values and graph them out in a multidimensional space. You then look for the largest gap between two groups, and draw a line through it. Inevitably, you'll find two groups -- and you can label them "men" and "women."
At its base, it's pretty simple. It gets much more complicated when you add nonlinear classification or need to classify more than two groups or need to reduce some fuzzy value to a number, but you can still do it.
There are two serious problems here.
First, once you've run your classifier, the reason for your outputs is completely inscrutable. The best you can say about the results is "well, the result is on this side of the hyperplane, so it's an apple rather than an orange." Which is not particularly helpful.
Second, if you give a system out-of-band data, it will return garbage results. So, for instance, if we have a simple apple / orange classifier, and we give it a bunch of grapefruits, it will tell us whether grapefruits are apples or oranges. Because that's what it classifies. It won't tell us "uh, these things aren't oranges," because "not an orange, but also not an apple" is not a concept that it understands.
(4) Diagnostic systems.
These are a narrow subcategory, but you encounter them daily. If you're unfortunate, you've run into this problem on Google itself.
So, when a link is first uploaded, Facebook sends a bunch of browsers-in-sandboxes to crawl it. More than likely, it's a computer in a datacenter which is running a simulation of another computer, which is running a simulation of another computer, which is running a browser. Then it runs diagnostics on the innermost emulated computer to see whether it's doing anything interesting.
Where "interesting" means "trying to bypass security and gain access to anything it shouldn't." If the malware detection system finds out that something's trying to escape its sandbox, it bans the link.
The problem here is that all software is broken. By which I don't mean that people are intentionally building bad software, but that most code that compiles still has unanticipated side effects. Some of which look like security issues.
From the comments (in answer to the question posed by the collection name):
>> "This just in: ThinkGeek to no longer carry PC inspired products and wil now require pre-order on all new merchandise!"
Consider first the difference between two masks: putting on a brave face (say when you’re in adult trapped in a dangerous situation with a child, where you cannot admit you’re scared or worried) and political correctness (say you’re at an office party where you cannot be completely candid). Both are voice masks; masks you put on when you have to pretend to agree with a sentiment you actively disagree with. You relieve the strain of voice masks by moving to a social context where you can speak more freely, and express your real emotions more completely. In the former case, it would be nice to have another adult around — making it a larger group — to share fears and anxieties with after the child goes to bed. In the latter case, it would be nice to retreat for drinks with a couple of trusted friends — a smaller group — to have a more candid chat about current workplace politics.
Now consider two other kinds of masks: pretending to enjoy yourself (say at a family gathering or a graduation ceremony where people who care a lot more are deeply immersed in the proceedings) and pretending to be interested (such as when listening to a boring but influential person drone on in a situation where leaving would cause offense and repercussions). These are exit masks: masks you put on when you have to pretend to care. You relieve the strain by moving to a social context where you don’t have to speak or fake an emotional intensity you don’t feel. Again, in the first case, you might relieve the stress by moving to a larger group that affords greater anonymity (such as a big city) and in the former case by retreating to a smaller group (perhaps going for a walk alone to unwind and get the bullshit out of your head).
These are opposed drives: moving in social space to speak and emote more versus moving in social space to speak and emote less. Of the two, exit masks are more basic: it is only hard to pretend to agree when you care. If you don’t care to begin with, pretending to agree adds no additional strain. You’ll nod along to whatever.
You can stop pretending to care once you are alone, but you can only truly stop pretending to agree when you are with a group with whom you can voice your disagreement to sympathetic ears.
Dissent exits are driven by feelings of anti-belonging or repulsion that drive you away from groups whose sacred values are profane to you. Once you’re out, you start seeking the sacred once more.
There is a basic tension in the human condition: we explore best alone and heal best in togetherness.
As a voice mask starts to break, you get angrier and more cruel, or break down in tears. As an exit mask starts to break, you might get increasingly callous, reckless, contemptuous or coldly logical.
Money, drugs, communication technologies, impersonal organizations. The Big Four technological means for extending your elastic limit.
These technologies also work to lower voice-mask strain.
Technology drains power from those who impose masks on others.
An example should hammer home the point. A major class of situations where we need masks is waiting. Powerful people make others wait around all the time, often in states of ceremonial silence or physical discomfort. Punishment for relieving the stress either through complaining or leaving is high.
But smartphones drain the authoritarian power of making people wait.
Advanced rocket engines might have gotten us to the moon, but money, alcohol, cigarettes, typewriters, phones, copiers and meeting procedures helped the humans involved build those engines without murdering each other in the process.
To get to the Moon, you need to stop pretending to care what others think a lot more than you need to stop pretending to agree with what others think.
This is why the industrial age organization was designed to enable specialization: giving everybody intellectual turf to retreat to. Somebody interested in electronics did not have to pretend to care about somebody interested in book-keeping.
Technology introduces an asymmetry into the age-old tension between exit-masks and voice-masks.
Under normal circumstances, human communities are self-stabilizing at some size, just like ape troops. Wander too far, and voice-mask strain draws you back in. Get too deeply immersed in community life and exit-mask strain pushes you back out.
The problem is that technology has an exit bias.
It is much easier to develop technologies that allow us to lower our exit masks further and wander farther off, than to develop technologies that allow us to lower our voice masks and come closer together.
Wow, it really is hard. I mean difficult. Yes. Difficult. Because of all of the unintended peens in the world.
Possibly NSFW even though none of them should be.
It's one thing to circulate a current Amber Alert, ensuring that all of the information is there, that it comes from a proper source (ie don't just reshare, CHECK THE LINKS), and that the answer is to call police, not just a random number. And then, of course, update your post (and let those who have reshared know) when it is resolved.
But "missing" posts that tell you to call someone other than police, that have no date or location, or are in any other way suspect should not not be shared and possibly should even be reported. This is how stalkers can find their hidden victims. This is how abusive parents reconnect with children that have been taken from them for good reason.
Go ahead and assist police on actual missing persons, but don't pretend you're a cop or detective or on the crime show of the week by taking it upon yourself to effectively dox someone who is trying to stay hidden for reasons you may not be aware of.
Let's face the facts: some of the people you know online are carefully constructed personas, and some of them lie. That nice guy you've been interacting with on social media for ages? Do you think he'd tell you about how often he punches his wife in the face? Of course not, so when he tells you she's missing and you put the word out and a friend of a friend saw her, you're unwittingly part of her further abuse and possibly even murder.
That nice lady you talk to on that game you play, who always sends more lives or helps with your farming or whatever...do you think she's going to tell you about how she beats her teenager and is about to ship him off to bible camp to beat out his gayness? Hell no, she's not going to tell you about his "shame"! So when he runs away and she puts out a plea to call her because she just wants her precious boy to come home and you help her find him, you are not helping him at all.
Imagine if someone has escaped a personal hell after months of brave but terrified planning, and you're the one who points a big digital finger to say, "HERE THEY ARE!"
We all want to help our friends when someone they know is legitimately missing. We all want abducted people found fast and hopefully relatively unscathed.
But THINK before you post, and THINK before you share. Check the details. You don't need to be rude about it, but you can be responsible. When in doubt, don't reshare.
- St. Edward's University
- Odd and off the wall stuff
- Robotics and Artificial Intelligence
- Making and DIY
- Literature and poetry
- Roleplaying games
- Random musings and thought experiments
- Language learning stuff, especially English <--> Spanish
- Cooking, especially vegetarian and vegan
- New York City
- Social Justice, especially feminism
- Have some public posts
- Preferably posts that interest me in some way
- Hint: it helps if they aren't all links to the same page/blog/subject; I like well-roundedness
- Interact with me in a meaningful way
- It also probably helps if you have any idea what my tagline comes from. Just sayin'.