Vector Calculus of Argument

and so I must say: I agree

(Maybe I should stop posting things based on dreams I remember at 3AM)


Making Do [ Review & Speculation ]

I recently got an opportunity to speak with Steve Daniels, author of "Making Do", about making a business out of making things out of waste in Kenya. You should read it, its free online. This was Mr. Daniels senior honors thesis at Brown University, and I am impressed.

I'm unfamiliar with the social sciences, and have never done anything as ballsy as travel to the poorest parts of the world to study their problems and potential solutions first hand. So, utmost respect to Mr. Daniels, and I doubt I will be able to contribute much to this particular topic, but here are some thoughts.

My absolute favorite line in the thesis is the following :
"microenterprise efficiency comes not from the individual firm, but from the dynamics among similar enterprises in collective geospatial clusters. In fact, through clustering the jua-kali economy displays a critical property of ecosystems that Western economies lack: it produces virtually no waste."
Daniels illustrates how large industrial centers have arisen spontaneously in Kenya, capable of processing local scrap into useful goods. Everything is recycled and re-used. Items as complex as welding tools and metal lathes are made locally from improvised scrap. These tools can then be used to process more scrap, and to improvise even more tools, creating an organically growing and self reproducing means of production.
"... the linkages among microenterprises form dense networks of activity. Take a stroll through Gikomba, and one can’t help but think of the informal sector as a living organism with intricate systems that form a concordant whole."
Daniels goes on to elaborate on what industrialized nations can learn from this production scheme, in terms of building an efficient economy that wastes nothing and can run on constrained resources. Daniels also notes that there are a number of barriers slowing down the technological progress of this informal manufacturing. Most notable is that subsistence lifestyles leave no room for risk taking and credit is scarce, so there is no room try to develop new technology. There are a number of clear and very doable suggestions in the text, and the work represents one of the few satisfying "and this, in detail, is how we will save the world" answers that I've come across.

In some sense, this form of production is similar to an idealized future-society that we've been considering some time. That is, a society that can run on very low resources, produces zero waste, fits into the natural environment, is local, is self reproducing, and is highly mobile. Now, scrap processing in Kenya does not fit all of these criteria and presents with some serious drawbacks. It would be beneficial to consider weather the flaws evident in the from scrap production in Kenya generalize to our idealized science-industrial complex.

In Making Do, Daniels notes that production of uniform, high quality machinery in the improvised production environment is not cost effective. The specific example cited is an irrigation pump that could not be made affordable to local farmers until its production was outsourced to China. I believe that this could be a fundamental flaw in aiming for local, robust economies that contain a complete industrial basis. At the end of the day, the most efficient means of production is always to produce en masse, for the entire planet, and in a giant optimized and computer controlled assembly facility. Globalization is not just inevitable, it is the necessary and most efficient means to produce high technology.

That said, the optimal means of production my be a hybrid of the local-cellular approach and the global-centralized approach. It makes some sense to claim that the level of sophistication in a technology is directly related to the extent that globalization of production is necessary. As a simple matter of resources, technology that requires neodymium, sapphire, gold, silicon, arsenic, germanium, must necessarily involve global trade networks. Simpler technology that requires only rocks and plants can be manufactured in your backyard. Scrap metal re-use in Kenya that can produce lathes and welding torches and stoves requires an intermediate level of trade integration.

I suspect we can reason about this formally. If we look at a graph representing the transportation cost between population centers, we can get some insight as to where to place the global-local trade off. If transportation costs are low, the whole world is highly connected anyway, so in a sense everything is local. If transportation costs are infinite, no global activity is possible. Reality lies somewhere on this spectrum. It also seems likely that the transportation cost has been kept artificially low due to the temporary abundance of fossil fuels, skewing our means of production to have more globalized characteristics than may be sustainable.

So why was I ever set on some science fiction of the self-reproducing assembly machine that could produce everything I'd ever want in my garage (this being the epitome of hyper-local high tech production)? I remember it has to do with Star Trek and those "replicator" machines that you carry around in space and will make you tea, lasers, whatever you want. If we are to achieve off-world colonization, it is necessary to reduce our industrial basis to something that can escape Earth's gravity well. If we can figure out how to make computers in a slum in India, from purely local resources, maybe we can start working on how to produce them on Mars.

So, with that selfish interest in mind, it sounds like improving the availability of micro-finance in poor regions, coupled with Fab Labs to distribute prototyping technology and skills, is a promising course to follow both for alleviating world poverty and for purely academic advancement in local manufacturing. Read Daniels' thesis for the full list concrete solutions for expanding improvised manufacturing in the third world.


Mathematical Fish

Nervous systems, a company that applies algorithmically generated patterns in design, has posted a collection of animal coat patterns that can be created by reaction diffusion equations. As someone who spent most of undergrad modelling stem cell differentiation and pattern formation, I really appreciate this company. Their products are uniformly beautiful. Image here : a snapshot of the patterns formed by the reaction diffusion inspired hallucination model.



I am not at all qualified to discuss issues of brains, but I do have some thoughts the previous post about society that deserve their own thread.

The definition of self is fluid and culturally defined. Western cultures are individualist, the person is defined by internal characteristics. Eastern cultures are relativistic, people are defined in relation to their role in a social structure.

But as it relates to eusocial humans, we need to think about how groups of people make decisions. A classic approach is authoritarian, decisions are made by some central authority (the king, the church, tradition), and radiates outwards into society. The democratic model has individuals rationally evaluating options according to their own metrics, voting, and then the group abiding by that vote.

A more modern model sees individuals as linked into networks, influencing each other through a non-rational transmission of ideas. Decisions are made by consensus, some critical mass breaking towards a policy choice. Consensus selves are aware of their membership in free associations, and do not have the same concern for internal intellectual growth as individuals, instead picking and choosing from the general cultural memepool.

Authoritarian decision-making is totally discredited, but the question for us is about the distinction between democratic and consensus models. We're in a hybrid state right now, but can sufficient technology create a pure consensus model? Do consensus models break down, and do they actually produce better decisions?

I have a creeping suspicion that consensus fails in any group larger than few hundred. Democracies have many roadblocks to making a decision, forcing deliberation. Those who lose out have the satisfaction of being a loyal opposition, and waiting for the time when issues will swing their way.

Consensus models may seem to invite participation from all, but in reality they are driven by memetic blitzkriegs. The winners are those who can mobilize and make decisions before an opposition can form. Debate becomes more acrimonious, and facts and history are thrown by the wayside (see the Ground Zero Mosque for a great example). Quality of decision-making suffers.

Consensus forms people into cliques, and because the policy-making group is not well defined, losers have an incentive to splinter off, into smaller communities with more narrow intellectual spaces, again harming the quality of debate.

Of course, the question of how close Twitter is to a true eusocial system, and what upgrades might be needed, remains open. But I would definitely argue that consensus, while flawed for the reasons mentioned above, is not inherently unsalvageable. Democracies can work spectacularly (the American Revolution), and fail spectacularly (the French Revolution). The same kind of serious intellectual debate that lead to the American constitution should also be applied to the architecture and norms of consensus based decision-making processes.


Why are some aggregates "smarter" than their individual components and others are not?

So I had a very broad question which I had been dimly aware of for some time, but I've never asked it. I'd be interested what some people have to say... I'd almost consider trying to post this on a site like Math Overflow and seeing what people say, even though its not really what they go for I think (I would probably phrase it differently). Please chip in your 2 cents.

In many many fields we have this idea of simple units acting together in cohesion to create some very complicated aggregate body. In some cases, for instance, Brains, the behavior of a single neuron is thought to be quite simple, and "unintelligent", while the overall body displays substantially more complexity of behavior, and capacity to adapt favorably to various environments and situations. On the other hand, maybe the clearest example of the opposite is the "stupidity of crowds". A crowd of people is thought to have dramatically less problem solving ability than a single person. A single person is frequently able to able to monitor their spending and manage a budget appropriately for instance, while the California legislature is not.

Other examples of systems may not be so clear cut. For instance I'm not really sure if an electron is smarter than a cloud of electrons. In areas like probabilistic combinatorics, we can frequently create large probabilistic systems composed of very simple components which are coupled together in simple ways, but about which we can say almost nothing in terms of the behavior of the whole system. In statistical physics I suppose it is the opposite -- predicting the motion of a particle, given its local environment, is thought to be extremely difficult, and typically modeled perhaps using Brownian motion, yet we can deterministically model the evolution of the gas as a whole, and develop useful statistics to "characterize" the macrostate.

I suppose that in most of science, the only time we study aggregates of "smart" components are where the components are people or animals. Perhaps we do not recognize other components as smart?

Its not clear to me what precisely is different in the way that societies are built of humans and the way that brains are built of neurons. You might suggest humans can move freely for one, but in most cases, it seems that people establish some local network of people they trust and respect, through which they receive information, and these parameters of trust and esteem become established and then fine tuned as life progresses, perhaps not unlike neural network weights. Obviously the aggregator function is substantially more complicated.

Perhaps you might suggest that for some problems, networks of humans are quite effective -- for instance, we can design space shuttles, and a single individual probably cannot do that. Its only political problems that individuals fail at. Perhaps you can point out the problems that brains fail at by analogy... certain long term risk reward tradeoffs? drug addiction?

Of course the answers you get will depend heavily on the formalism you choose for computational ability. What I would like to ask is this:

1) Are brains organized from subunits in terms substantially different from societies / other schemes?
2) Given subunits of a certain design with a certain computational power under some formalism... VC dimension which can be learned efficiently? Topological Entropy of the analogous dynamical system?.... how much computational power does the aggregate possess, when formed under one connection scheme vs. another?

Obviously 2 is going to be pretty hard to answer... and will depend on your answer to 1 which may be contentious. Pitch in your 2 cents.

Are nation-states becoming irrelevant?

This article claims that they are, and that cities are going to be in the driving seat in the 21st century. I feel like actually it's misstating its own thesis, which seems to be that the wealth of individuals and their voluntary associations will once again outstrip states' ability to control them. (Otherwise, what does it mean for a city to want something or do something?) It supports this by pointing out worldwide urban growth and the growing gap between urban rich and rural poor in countries like China, but it doesn't really make clear why a city is a more coherent entity with more potential for controlling things than, say, a patch of countryside which shares some political allegiance or ethnic identity. What I want to say to fill this gap is that the urbanized masses are actually organized into corporations and similar organizations with common interests, and that it is those, and not cities per se, that are doing the controlling; whereas the countryside is generally composed of individual farmers and such who are mostly concerned about their immediate neighborhood. This is a very libertarian future this guy is predicting, though I'm not sure he realizes it.

I'm still skeptical, but I don't have anything clever to say. Thoughts?

A Blog of Interest

I really appreciate the journalism and photography of this acquaintance. I feel like the sum of her recent posts definitely rival F's Modern-Day Colonialism essay.

To try to relate this back to WeAlone, at one point I asserted that it's easier to invent a new technology to save the third world than it is to fix the political situations leading to poverty. I have been told that India's problems stem entirely from government inefficiency and corruption, impairing resource distribution and planning for off seasons, in a country that is otherwise resource wealthy. Is this actually the case ? If so, what sort of technology, if any, could tackle this problem ? Is this fundamentally a social, rather than a technological, hurdle ? And, if so, do we necessarily need to overcome such hurdles before we can claim to have moved into a new technology-dominated phase of society ?


Newsy Tuesday

In lieu of actual content (maybe later), I'd like to post something cool from Singularity Hub.

vOICE is an innovative augmented reality system that lets people see with sound. Images from cameras are translated into tones, which are played through headsets. The system has surpisingly good fidelity, and implications for brain research.

What’s amazing about the system – and what makes it a hot item for neuroscience research – is that it appears to restore the actual subjective experience of vision (visual qualia) to blind users, rather than just teaching them to correlate objects and sounds. Users have reported the return of experiences like depth and the sense of empty space in their environment. The restored vision is not the same as normal visual experience – one user described it as being comparable to an old black and white film, while others report vague impressions of objects as shades of grey. Research is now underway to understand how the vOICe system might be rewiring the brain to achieve this effect.

Holy crap, we can see with sound. No inplants required, no electrode arrays on the tongue, just off the self tech and clever algorithms.

The neurobiology of psychedelic drugs: implications for the treatment of mood disorders


I would not mind living inside this fractal

Youtube user zsirhorcsog has posted some sort of crazy 3D fractal. It more or less looks like a gothic cathedral made of math, when you zoom in. Its something I would not mind at all seeing converted to a nonspecific place-of-worship-esq building.


The Information Monopoly

From the New York Times Magazine.

...As broadband brought millions of facts, the fantasy of perfect factuality and the satisfaction of fact-checking to everyone. Soon — and astonishingly — Google became much more than trusted; it became shorthand for everything that had been recorded in modern history. The Internet wasn’t the accurate or the inaccurate thing; it was the only thing.

Those of us who think seriously about the creation of facts in society should be concerned about the totalitarian effects of the internet, search engines, and in particular Google on the search for truth. The internet is the first, and often last stop on any quest for information. That information is filtered through search algorithms by a handful of companies, Google, Yahoo, Microsoft (somewhere between 60%-70% is Google alone).

For an Orwellian overlord, the power of this technology is obvious. No longer will tyrants have to burn books, they simply have to drop dangerous information off page one. Dissent is like a wild-fire, and like how a fire can be stopped by removing one of the sides of the Combustion Triangle, dissent can be squashed by sucking away the oxygen of information. You need not remove it all, just make the cost of access to information high enough that the energy that fuels revolution disappears. The crude techniques of airbrushing and book burning are obsolete, just switch up an algorithm and watch the inconvenient truth disappear.

As a society, we must resist the centralization of access to information under a few massive providers. We should demand openness on algorithms, so that search results may be fairly judged, and we should form an organization devoted to looking for such Orwellian modification, cataloging them, and making the public aware of what is hidden.


Questions in Subjective Experience

Earlier today I was trying to generate an answer to "what would it mean for Neuroscience to succeed". This post is not about that. This post is about one of those dumb questions from philosophy "What makes red ... Red?".

Far from being a meaningless or rhetorical, this question takes on a concrete meaning in neuroscience, although we may lack tools and data to give a satisfying answer at the moment.

Light at a wavelength of 700 nanometers is not quite the answer, since color perception is relative and one can be prompted to experience illusory colors. But, even if we account for all these idiosyncrasies of perception, do we have a satisfying answer for why red feels like red and green feels like green and they seem obviously non-interchangeable ?

Perhaps. There are ( I simplify ) three channels of information sent from the eye to the brain. One is information on light/dark contrast. Another is yellow/blue contrast. Finally, we have red/green contrast. Simple black and white vision evolved first. Later, yellow/blue color vision appears to differentiate between different types of "white". Some completely red/green color blind individuals perceive color in this space. The Red/Green channel evolved most recently, in old world primates, and distinguishes between different types of 'yellow'.

So, trivially, red is "not green, not blue, bright"

Perhaps this is enough to satisfy you for a definition of red, but I am unconvinced. For instance, I feel completely satisfied with a definition of left as "a direction that is orthogonal to up and forward, and is not right". This is because, for most of my life, I have been largely unable to tell the difference between left and right. Now that I know how to read and have learned other skills with a left-right asymmetry, I'd notice if you flipped left and right, but I'm not convinced I would have noticed as a small child. However, I think that waking up one day to a world in which the red-green axis had been reversed would be quite alarming.

Red-green is something more than a mere sign convention within "not blue, bright". This axis represents some perceptual eigenfunction, and it needs to have more information than 'not green, not blue, bright'. If you accept the axiom that the experience of information is equivalent to the existence of said information, then the asymmetric quality of the red-green axis implies that either red or green, or both, contains some bits of information beyond simple color contrast.

One, somewhat boring answer is simply that, since the absorption spectra of green and blue cones overlap more, "green" necessarily contains some component in the "blueish" direction, more-so than red. So, red is simply the "less blueish" side of yellow.

Does this explanation also satisfy you for the yellow-blue axis ? It is possible that individuals without a red cone really do perceive color space as just having these two directions, like left(blue) right(yellow), and that the qualitative difference between them is slowly acquired once one is accustomed to the sky being blue and people being brown ? I'm not colorblind myself, so I'd need more data to answer this question.

Assume for a moment that this is not the case, the red, green, and blue, aren't just vectors in colorspace, but have some sort of intrinsically different representation in the brain. By this I mean that the extra bits of the 'red' vector are hard-wired into how the visual system works. This would require, at some point, 'redness' and 'greeness' to have representations in the brain that are more distinct than color opponancy. Speculation is that at some point "redness" is behaviorally relevant in ways that "greenness" is not, and so color perception is deeply related to some sort of behavioral cues. I stop short of speculating as to what those cues might be.

If neuroscience has any hope of mathematically describing experience, it will eventually address this question seriously. Eventually, we will be able to say yes my red is the same as your red, see, I proved it with math. Hopefully its just a matter of bits.

America's Ruling Class

Recently I read an article in The American Spectator entitled America's Ruling Class-And the Peril's of Revolution. I disagreed powerfully with Angelo M. Codevilla's conclusions, but thought that many of his arguments made sense, with strong logical foundations. A dangerous contradiction, and one that I've not experienced so strongly since reading The Unabomber Manifesto (and like The Unabomber Manifesto, America's Ruling class goes from "yeah, I agree," to "wtf are you smoking" fairly quickly).

Mr. Codevilla's thesis is that the political disenchantment in America, as most contemporanious represented by the Tea Party, comes from the growing power of a new ruling class since World War II, that class' arrogance, and its failure to govern. The common people of America, sick at this unnatural state of affairs, are beginning to rise up in revolt.

Who are America's ruling class?

Today's ruling class, from Boston to San Diego, was formed by an educational system that exposed them to the same ideas and gave them remarkably uniform guidance, as well as tastes and habits. These amount to a social canon of judgments about good and evil, complete with secular sacred history, sins (against minorities and the environment), and saints. Using the right words and avoiding the wrong ones when referring to such matters -- speaking the "in" language -- serves as a badge of identity. Regardless of what business or profession they are in, their road up included government channels and government money because, as government has grown, its boundary with the rest of American life has become indistinct... Hence whether formally in government, out of it, or halfway, America's ruling class speaks the language and has the tastes, habits, and tools of bureaucrats. It rules uneasily over the majority of Americans not oriented to government.

The ruling class encompasses Democrats and Republicans. At its top are Senators and Congressmen, its ranks are swelled by public servants, teachers, union members, and those who suckle at the trough of public money, ranging from welfare queens to non-profit activists to research scientists.

I took immediate offense to this, but as I calmed down, I realized that the reason I got angry was that Mr. Codevilla was right. I am a member of America's ruling class. My education at a state university is subsidized by local taxes, and I plan to find more government grants soon. My friends are liberals who speak in a certain language, many of whom work in the non-profit sector. And I am skeptical of the ability of business to help America, and am sanguine about the power of intelligent government intervention.

America does having a ruling class, and no one can deny the influence of the federal government, and its followers. "By taxing and parceling out more than a third of what Americans produce, through regulations that reach deep into American life, our ruling class is making itself the arbiter of wealth and poverty... By making economic rules dependent on discretion, our bipartisan ruling class teaches that prosperity is to be bought with the coin of political support." Nor can I deny that this ruling class has failed America. Since WW2, we've seen four useless quagmire wars in Asia, the decline of American education, immense increases in pollution, a decline in global prestige, massive government debt, an expansion of regulation without a commensurate expansion in safety etc etc. The ruling class is separated from its intellectual foundations, and is less intelligent and less coherent than ever before. They have failed to make a case for their position, and now in the everyday process of politics, denigrate their own ability to govern.

The Center Cannot Hold
Codevilla posits as a counter-weight the forgotten America of ordinary people, what he calls the "Country Class." These plain folks are fed up with an intrusive, top heavy, inefficient and unresponsive federal government. They believe in local autonomy, the family, and the Bible, and their beliefs are cruelly mocked by the ruling class.
Describing America's country class is problematic because it is so heterogeneous. It has no privileged podiums, and speaks with many voices, often inharmonious. It shares above all the desire to be rid of rulers it regards inept and haughty. It defines itself practically in terms of reflexive reaction against the rulers' defining ideas and proclivities -- e.g., ever higher taxes and expanding government, subsidizing political favorites, social engineering, approval of abortion, etc. Many want to restore a way of life largely superseded. Demographically, the country class is the other side of the ruling class's coin: its most distinguishing characteristics are marriage, children, and religious practice. While the country class, like the ruling class, includes the professionally accomplished and the mediocre, geniuses and dolts, it is different because of its non-orientation to government and its members' yearning to rule themselves rather than be ruled by others.

Wait, what? This class doesn't match any group I've heard of. Social conservatives are more eager than anyone short of radical Marxists to re-engineer society, in their case along Biblical lines. Libertarians want freedom to pursue their own rights, but trample over the ability of people to collectively face challenges. There is no such thing as a 'redistribution of wealth', all distributions are redistributions, and even hunter-gatherer tribes share food. The kind of society that Codevilla envisions is not Industrial, or old agricultural (as in Europe), it appears closest to the historical idealization of the American West. This lifestyle was contingent on a specific set of historical factors, primarily open land, that will not appear again. My personal experience of the country class, from the Tea Party, is that they are not rugged individualists, they are proles. They are worn out cogs in the social machine.

The country class wants local autonomy to choose their own school boards and curricula, but this is suicidal in a modern economy. Global competitiveness is dependent on a set of skills that can be assumed nationwide. Companies need to trust that their employees in Los Angeles have the same skills as their employees in Fargo, lest the company collapse through inefficiency and misunderstanding. They decry the intrusion of the Federal government, then demand agricultural subsidies, social security, and disaster relief. Without pork-barrel projects like the interstate highway system and rural electrification, the countryside would be a desolate wasteland of third world poverty.

Mr Codevilla shows his true colors: "Restoring localities' traditional powers over schools, including standards, curriculum, and prayer, would take repudiating two generations of Supreme Court rulings. So would the restoration of traditional 'police' powers over behavior in public places." What are these 'traditional police powers', not having to read suspects their Miranda Rights? What about immigrants? And where is the Military-Industrial Complex in Codevilla's society, the generals, spies, and defense contractors who justify so much of America's dysfunctional foreign policy in the name of 'national security'. I do not wish to call names, but under Codevilla's rejection of the Democratic-Republican establishment is a desire to raise up a new Military-Christian alliance, in short, fundamentalist fascism. Cities, the main drivers of American prosperity, would see their complex institutional networks disbanded in the name of 'autonomy.' And individually, small towns would slip into religious hysteria and xenophobic paranioa; Arizona's laws against illegal immigrants, and the Salem Witch Hunts.

America's ruling class has failed, but the solution that Codevilla advocates, a return to local autonomy, and "[An] attack on the ruling class's fundamental claims to its superior intellect and morality in ways that dispirit the target and hearten one's own is in effect the break-up of the American Empire (and we are an empire, even domestically). He may believe that America's grand democratic experiment that he claims to defend can survive this assault, but I feel differently. We are tottering on the edge of the knife. We must unite to survive. Policy debate and disagreement are vital to the health of our democracy, but the empty repetition that "Washington is broken" has to end.

Next Week on We Alone on Earth: America's New Ruling Class.

(PS: Apologies for typos, this was banged out in a hurry and I don't have time to proof read right now.)


Science, Technology, and Society Movie Night

I'm chilling at the CSPO start of year conference, which I can only describe as fantastic. The quality of the people and ideas here is like nothing I've ever seen. One of the activities was everybody writing down one thing they'd be willing to, and I volunteered to run an STS Movie Night.

A preliminary list:
Koyaanisqatsi-grab hold of your brainstem, because your view of man and nature is about to explode. Literally.
Zardoz-Some people say that this is a movie about life, death, immortality, the technosocial elites, and the nature of an apocalyptic ecological collapse. Other people say this is a movie about guns, penises, and Sean Connery in a red leather diaper. Both are wonderfully right.
eXistenZ-Cronenberg could compose a series all his own, but I like eXistenZ because its a higher budget film, and deals with virtual reality and levels of reality.
The Fly-mad science
Altered State-human subject experimentation
Jurassic Park-Chaos, man! And dinosaurs.
Hackers-Fact-25% of CS majors are a young Angelina Jolie. A look into the wacky computer underground of the early 90s.
The Terminator/Terminator 2-Military robots, technological inevitability, and of course Ahnold.
Idiocracy-A very funny movie about social Darwinism and the future of intelligence.
Gattaca-Genetic engineering, social stigma, transhumanism and equality. Also a damn good film.
Minority Report-Crime and thought. Note that DHS is working on 'multimodal mind-reading sensors' to detect terroristic intent. Can you charge someone for being nervous in an airport.
Eternal Sunshine of the Spotless Mind-Memory and love, who we are when we can cut away parts of our souls.

I've only scratched the surface. Loyal readers, can you suggest more movies?


Mill PCBs using Adobe Illustrator

I elected to design a couple circuit boards using a vector graphics editor rather than Eagle cad. This allows for smooth organic traces and arbitrary graphics, but can make fabricating the boards difficult. I would recommend Eagle cad or some other program that can output gerber files if you are designing a board that you might some day want to mass produce. These instructions and scripts pertain specifically to the "Modela 3D plotter" devices. These instructions are absurd and more of a satire of actual PCB milling.

Quick preview
0 - sketch it up
1 - rasterize
2 - screw around in Gimp
3 - convert to mill file using python
4 - adjust and mill board
5 - export to SVG & use python to turn SVG into drill file
7 - manually adjust and drill board

Software links
cad.py (for generating mill code, from here)
circle.py (for extracting circles form .svg and converting them to modela drill file)
modela-move.py (for manually positioning the mill)
modela-send.py (for sending a .rml to the mill with buffering)
realign.py (for applying affine transformations to .rml files)
stats.py (for getting .rml plotting bounds for sanity check)
- files not guaranteed for correctness, may generate commands that break mill, be careful
- realign hates negative numbers
- note to self : hosting on dropbox, move later
- requires python serial libraries and BeautifulSoup packages to run

0 - Draw the board. I set up a 0.25" grid which is good for 0.1" pitch components. Use circles only for component pads, and put a circle at the center of each pad. This makes it possible to strip out the circle positions with a script to generate a drill file. I connect pads with smooth splines and add custom artwork. Routing and positioning is all done by hand.

1 - Export your file and process through "cad.py" script from fab-lab to generate code for your milling machines. In fact, "cad.py" crashed when I tried to process SVG directly. Instead, rasterize the SVG at 600 DPI and process it as a .PNG. A few things to watch out for : Adobe Illustrator uses a default 72 DPI, while Inkscape uses 90 dpi. That is, units in the exported SVG file will be in 1/72nds of an inch from Ilustrator, and 1/90ths of an inch from Inkscape. Hence, opening a SVG at a DPI other than what it was created at may cause problems.

2 - I never thought I'd be using The Gimp for circuit board design. Open the 600DPI rasterized file. First, make the background is opaque, and invert the file such that traces are white and the background is black. Then apply gaussian blur followed by threshold to round out the traces, and manually corrected close traces with the paintbrush. Gaussian blus+threshold can also be used to shrink or grow the size of board features if you are having trouble with traces being too thin or too close together, but watch out for accidentally merging traces. This is absolutely absurd, but it works, and cad.py liked .PNG more than .SVG.

3 - We wrote a collection of utility scripts for the Modela milling device. These include

realign.py for apply arbitrary affine transformations to a mill file (works on standard in/out, redirect files to/from these IO streams to read/write from files). This is useful for moving the file co-ordinates to just the right spot in mill co-ordinates. This is also useful for manually aligning drill files or two sided circuit boards, since they seem to be a bit off sometimes. It can't handle negative coordinates at the moment.

The file stats.py simply prints out the number of plotting commands (redirect files to standard in), followed by the minimum and maximum x and y coordinates of the milling file. This is good for a sanity check to make sure your file is in the right location. It can't handle negative coordinates at the moment.

The file modela-move.py accepts x,y,z coordinates and moves the mill to that position.

The file models-send.py accepts a .rml file piped from standard in and sends it to the milling machine. This script is necessary to avoid overflowing the command buffer on the mill.

4 - Use realign.py to move the output of cad.py to the correct position for milling. Use stats.py to double check that your board is where you think it is. Calibrate and position the mill. You probably want to make sure the x,y,z speeds are set slow. I have been using the slowest speed of "1" because I broke way to many bits going faster. Read up on the Modela mill or ask the mill owner to train you in its operation. TODO : collect some useful links here ?

5 - Prepare the drill file. To do this, redirect a SVG of the board to the "circles.py" script. This script will extract the centers of all circles, remove duplicate circles, and create a semi-sane set of drilling commands. However, it does not use exactly the same coordinate transforms as cad.py, hence drill files are off anywhere from 5 to 40 mils ( the machine uses 'mil' as the basic unit, which is 1/1000" ). Whatever transform you applied with realign.py to the board file, do this also to the drill file.

6 - Manually adjust your pen down height to be 50 mils above the board. Run a test run to see how far off the drill file is, and use realign to move it to the correct position. I really wish I had a better procedure here. When the first hole looks aligned, set pen down to -80 mils and drill. Make sure there is something underneath the board so you don't dril through the bottom of your milling bed.

This actually works. I milled nine boards the the past couple days. Designing circuit boards in drawing programs, and using gaussian blur and paintbrush aren't generally part of the workflow for circuit board production. You can also just print off the board and use the toner transfer method, then drill by hand.

The results look nice, though 0.1" pitch is as small as it can go for through-hole work. SMT probably has a higher resolution since you don't have to worry about the holes.

p.s. : I dislike how google sends the blog post out to the RSS feed when its first created, such that subsequent grammar corrections are not reflected in people who read this blog via RSS reader. I guess this just means I should proof-read better.


MakerBot Grows Up

I've always been doubtful of rapid prototyping as the basis for some sort of home industrial revolution. As my co-author Everett has demonstrated, MakerBot is far from ready as an appliance, and even if it were, I doubt that there is much of a market for tiny plastic trinkets.

Now this is where it begins to take off. The army has deployed self-contained machine shops to Iraq and Afghanistan to build replacement parts as close to the front line as possible. Now, CAD-CAM milling machines are nothing new, but as a self-contained unit, something like this is unique. I think to become a successful technology, RepRap has to become truly self-replicating, which means determining the minimum number of materials necessary to support a RepRap, and a way to build all the necessary modules using the RepRap itself.

The end goal of a self-contained, self-replicating industrial complex may be impossible sort of magical nanotech (or alternatively, the factory-machine is just huge.) But the idea of a device that can be dropped into the wilderness with a minimum of rare elements, and proceed to construct all the necessities of technological civilization is too attractive for me to just give up on.


Thought of the Day

"If we intend to practice anticipatory technology assessment either as inventors or as policy analysts and scholars, we should not approach [The Future] as a surfer would the waves but perhaps as an oceanographer might."
--Sean Hays

"Prediction is very hard, especially about the future."
--Yogi Berra

How should we approach futurism, with data or with intuition? We are (supposedly) serious scholars, and that implies some kind of data. Certainly, our review boards with be happier with dense footnoted and figured papers, but I'm far from certain that data driven research produces better results.

Futurism is not about specific predictions, timelines and events. Instead it is a way of looking at contemporary policy choices, of broadening the framework and implications of our choices. The job of futurist is not to predict, but to scenario build, to give people a sense of what tomorrow is going to look like, and why that vision of tomorrow matters today. I've read a fair amount of futurist work, and I will say that almost anything by Bruce Sterling is more visionary and more true than the best of Kurzweil or Buckminster Fuller.

This is not to say that data isn't important. Futurists have to keep current on the latest scientific advances. Historical analogies are the basic building blocks of futurist construction. But while a rigorous, data driven approach may look impressive, I doubt that the results are any more useful.

We use imagination to explode the present.


Simple light-diffusing plastic with materials available at the local drugstore

I have been experimenting with lazy ways of making the hallucination goggles. At the moment, its fairly easy to mass produce electronics, so thats not an issue. The main annoyance is building suitable goggles themselves. The only requirement is that enough diffusing media needs to be placed between the eye and the LED in order to create a uniform stimulation of the entire visual field. Today I experimented with pouring various mixtures available at your local KMart into cheap swimming goggles, also from the local KMart.

The overall idea is to get some clear liquid plastic that can be poured into the goggles, and to get some sort of powdered or granular substance that can be mixed with the liquid to make it diffuse light. Technically, any clear granular substance would work since you only need to create a heterogeneous index of refraction in order to scatter light. I tried both PLA ground in a coffe grinder and table salt. For liquids, I tried silicone and clear 5 minute epoxy (5 minutes is quite necessary if you want to be able to get a smooth finish. Anything else cured before I could pour it. Make sure your epoxy mix is clear, not yellow. ) You don't need much of whatever you're adding to the liquid to make it diffusive. I would say no more than 25% of the total volume. I used about a teaspoon for one pair of goggles.

Silicone mixed with PLA that had been powdered in a coffee grinder did not work. The silicone was much to viscous to pour or mold properly, and the PLA flakes were too large and hard to mix.

Epoxy mixed with table salt seems promising. The table salt doesn't seem to effect the curing of the epoxy.

Initial tests are totally awesome suggest that the epoxy+table salt solution works extremely well, and is much faster and more sturdy than the ping-pong ball construction method. If you find that the slight yellowing skews the white balance, the software in the original goggles post should contain some variables for adjusting the white point. Adjust the location of the LEDs so that the perceived light inside the goggles is as uniform as possible, which may appear off-center on the outside of the goggles.

As an added bonus, these goggles are still totally waterproof. Hello do-it-yourself submerged neutral bouyancy sensory deprivation chamber. If I had to do this again, I would pour less epoxy, presently these goggles collide with my eyelashes. Not to mention the fact that epoxy cant stick to flexible surfaces like the seal on the goggles, so if you pour it up there its just going to flake off and make the whole job look messy.

The exterior could be more stylish though. Perhaps there is some way to put the LEDs inside the goggles, or to make some nicely fitted PCBs.

You might want to rinse the goggles in water after the epoxy sets. This will dissolve the few grains of salt that managed to be exposed on the surface.


Reply to Jaron Lanier

A few days ago my colleague Everett posted an NYTimes op-ed by noted futurist Jaron Lanier. Lanier is the kind of person who gets placed next to Ray Kurzweil and Bill Joy; an autodidact VR pioneer musician artist free radical philosopher. Summarizing Lanier only does him injustice, but one could say that he's a transhumanist, with the emphasis on 'humanist.' ((EDIT: He'd vehemently disagree with this characterization, but I think it fits. His advocacy of radical biological alterations to human beings is pretty transhumanist.))

In his editorial, Lanier argues against mixing theology and technology. He posits that AI research has taken on psuedo-mystical overtones: we are creating new life, we are evolving, we are supplanting the human race, that is profoundly dangerous to technological progress and human values. Scientists risk a Frankenstein backlash from a majority rightfully concerned with the direction of society, wherein as computer become more human, humans become more machine-like.

Lanier makes two points in support of this claim. More and more of what we do is act as information processing nodes, hyper-linking, retweeting, sending on without thought or commentary. The signal talent of the 21st century is curation, sorting the gems of information from the dross of data we are deluged in, and as far as human ability is concerned, this is a sorry use for thought. Like, dis-like, access, archive, delete, repeat. When sorting information becomes our major activity, creativity is devalued.

But more critically, "What all this comes down to is that the very idea of artificial intelligence gives us the cover to avoid accountability by pretending that machines can take on more and more human responsibility." Evading responsibility appears to be one of the major spectator sports of the 21st century. Disasters happen again and again, and any human who might be responsible slips away from the clutches of retribution. Artifacts, derivatives, blowout preventers, are blamed for what are ultimately human failings. Instead of operating machines, we become supervisors of algorithms. With each degree of automation, we cede both responsibility and control. It is a common observation that events apear to be spiraling out of control, and that no one is at the helm. Maybe its true, maybe the ship is steering itself, and rocks and shoals be damned.

Lanier's solution: "When we think of computers as inert, passive tools instead of people, we are rewarded with a clearer, less ideological view of what is going on — with the machines and with ourselves. So, why, aside from the theatrical appeal to consumers and reporters, must engineering results so often be presented in Frankensteinian light? The answer is simply that computer scientists are human, and are as terrified by the human condition as anyone else."

Well, that is all very well and true, but are computers inert, passive tools? Individual machines are, but computing as a technology is far from inert. It moves and shifts according to market forces, it reacts to changes in technology and economics without clear paths of human responsibility. In short, computing and computers in aggregate already behave like a living organism, and short of a Butlerian Jihad, that organism cannot be stopped. Contemporary AI is still far removed from the question of 'what is a person?'. At best, it is a collection of specialized tricks for limited situations. Emotional inner states are the province of hairy-eyed radicals. But we are not alone on Earth. Our machines are with us, with their own inertia and desires, and while it is true that humans are the sole entities with agency, we do not interact one-on-one with the other six billion inhabitants of Earth, but through a technological medium that is very much an active participant.




colorizing notebook doodles


Gossiping Ourselves to Death

I recently read Neil Postman's excellent Amusing Ourselves to Death. Postman offers a critique of the corrosive effect of television on American discourse, education, and culture. Television is of course the dominant media of the 20th century, and as Postman describes it, "The problem is not that television presents us with entertaining subject matter but that all subject matter is presented as entertaining, which is another issue altogether." Anything presented on television is evaluated first by its potential to entertain, its capacity to inform or enlightened is nearly irrelevant. When entertainment become the central virtue, politics becomes talking heads shouting at each other, rather than a considered debate over the issues and merits of policy. Religion is reduced to a public spectacle, as in Evangelical mega-churches, without any spiritual or moral dimensions. Sesame Street and its ilk turn students away from any knowledge that does not grip their sense of fun, to the detriment of higher learning.

Amusing Ourselves to Death is one hell of a diatribe, its only flaw is that it was published in 1985, and so does not cover the defining communication medium of the 21st century, the internet. What kind of results to we see extending Postman's methodology to computers?

The technology of the internet is very good at two things, interpersonal communication (email, IM, forums), and personal publishing (blogs, Facebook, Youtube). The best thing that can happen to something on the web is that it goes viral, that people feel compelled to send it on to all their friends. Intrapersonal communication, a need to forward the best stuff, and a forum for public display; strip out the jargon and what you have is old fashioned gossip.

Email and IMs make gossip faster and easier, but don't alter the fundamental nature of gossiping. On the other hand, the whole point of Facebook is to carry out the social functions of gossip without any human intermediation whatsoever. We stay in touch with our friends, find out how their lives and relationships are going, check out their favorite bands and TV shows, and all from the comfort of our home. Moreover, when we use Facebook, we willfully invite other people to gossip about us, we become exhibitionists. In the future, even introverted people will need to maintain some kind of public presence.

What does gossip imply for society and politics? The more salacious a piece of gossip is, the more engaging it is. Instead of being entertained, we are disgusted and titillated, the truth of a rumor is basically irrelevant, and fact-checking often strengthens misinformed beliefs. The damage to public discourse is obvious. Gossip has always played a role in politics, but when it plays the preeminent role, honest evaluation of our leaders, their beliefs, and executive ability become impossible. Politics becomes little more than innuendo, name-calling, and black propaganda.

I don't have an easy antidote. We never came up with one for television, and a glance at the cable channels will show you it's worst than ever before (see the Ghosts-and-Loggers, I mean History Channel.) On an individual level, we can steel ourselves against internet rumor, but that's just a band-aid on a sucking chest wound. I think rumor is embedded in the technological architecture of the current generation of internet technologies. We need a new communication medium that penalized those who spread rumors.