Productive Recursion Proven
Nathan Cravens
Days ago I read Neil's very down to earth book, 'Fab'. Neil believes we'll have the equivalent of Star Trek replicators in twenty years time. I want this to happen, but there is one problem that us post-scarcity types are well aware, but perhaps have not put into these words: Incentives to produce while operating with less and less money (capital gain) over time is one obstacle that needs addressing as we go from the "minicomputer era of personal fabrication" to the personal computer era of fabrication. If the economic pitfalls of the present Industrial system are ignored, people will not have enough labor value to purchase the rights to print anything nor will firms have the incentive to invest in personal desktop fabricators to produce. If the capital gains issue is not addressed in our minicomputer era of fabrication, that issue alone may extend Neil's twenty year time frame unnecessarily.
Production methods lower the cost of products when simplified for rapid replication. That is called productive recursion. Understanding productive recursion is the first step to understanding how we need to restructure Industrial economic systems in response to this form of technological change. If Industrial systems are not reconfigured for productive recursion, they will collapse before reaching anywhere near full automation. I hope this writing helps divert a kink in the proliferation of personal desktop fabrication and full productive automation generally.
Here are cases that prove the theory of productive recursion in practice.
Distributed Production
Grass roots Indian TV cable operators charge roughly $2 per month for service, depending on the operator and number of clients. In short, this greatly reduced cost for service is due to reverse engineered satellite receivers by means of distributed production.
From Neil's 'Fab'
" A typical village cable system might have a hundred subscribers, who pay one hundred rupees (about two dollars) per month. Payment is prompt, because the "cable-wallahs" stop by each of their subscribers personally and rather persuasively make sure that they pay. Visiting one of these cable operators, I was intrigued by the technology that makes these systems possible and financially viable.
A handmade satellite antenna on his roof fed the village's cable network. Instead of a roomful of electronics, the head end of his cable network was just a shelf at the foot of his bed. A sensitive receiver there detects and interprets the weak signal from the satellite, then the signal is amplified and fed into the cable for distribution around the village. The heart of all this is the satellite receiver, which sells for a few hundred dollars in the United States. He reported that the cost of his was one thousand rupees, about twenty dollars"
The satellite receiver is made by an Indian company: Sharp.
" This Sharp turned out to be an entirely independent domestic brand. They produced all of their own products, although not in that room--done there, it would cost too much. The assembly work was farmed out to homes in the community, where the parts were put together. Sharp operated like a farm market or grain elevator, paying a market-based per-piece price on what was brought in. The job of the Sharp employees was to test the final products.
The heart of the business was in a back room, where an engineer was busy taking apart last-generation video products from developed markets. Just as the students in my fab class would learn from their predecessors' designs and use them as the starting point for their own, this engineer was getting a hands-on education in satellite reception from the handiwork of unknown engineers elsewhere. He would reverse engineer their designs to understand them, then redo the designs so that they could be made more simply and cheaply with locally available components and processes. And just as my students weren't guilty of plagiarism because of the value they added to the earlier projects, this engineer's inspiration by product designs that had long since become obsolete was not likely to be a concern to the original satellite-receiver manufacturers.
The engineer at the apex of the Sharp pyramid was good at his job, but also frustrated. Their business model started with existing product designs. The company saw a business opportunity to branch out from cable television to cable Internet access, but there weren't yet available obsolete cable modems using commodity parts that they could reverse-engineer. Because cable modems are so recent, they use highly integrated state-of-the-art components that can't be understood by external inspection, and that aren't amenable to assembly in a home. But there no technological reason that data networks couldn't be produced in just this way, providing rural India with Internet access along with Bollywood soap operas"
"There isn't even a single entity with which to partner on a join venture; the whole operation is fundamentally distributed"
Personal Fabrication
" For example, the ground resistance meters that were used for locating water in the era cost 25,000 rupees (about $500). At Vigyan Ashram they bought one, stripped it apart, and from studying it figured out how to make them for just 5,000 rupees. Those meters could then be operated to earn 1000,000 rupees per year in a business. Another example arose because they needed a tractor on the farm at Vigyan Ashram, but could not afford to buy a new one. Instead, they developed their own "MechBull" made out of spare jeep parts for 60,000 rupees ($1,200). This proved to be so popular that a Vigyan Ashram alum built a business making and selling these tractors."
Open Source Production?
" Anil runs the "Honeybee Network." Modeled on how honeybees work--collecting pollen without harming the flowers and connecting flowers by sharing the pollen--the Honeybee Network collects and helps develop ideas from grassroots inventors, sharing rather than taking their ideas. At last count they had a database of ten thousand inventions.
One Indian inventor couldn't afford or justify buying a large tractor for his small farm; it cost the equivalent of $2,500. But he could afford a motorcycle for about $800. So he came up with a $400 kit to convert a motorcycle into a three-wheeled tractor (removable of course, so that it's still useful as transportation). Another agricultural inventor was faced with a smilar problem in applying fertilizer; his solution was to modify a bicycle."
I hope these segments add to the vision and motivate your work.
On Productive Recursion
http://openfarmtech.org/index.php?title=Productive_Recursion
Nathan Cravens
Effortless Economy
Bryan Bishop
> "[F]ab labs are likely to have far greater impact on the stability of the
> planet by helping everyone else acquire the technology they need for thier
> survival."
>
> economic pitfalls of the present Industrial system are ignored, people will
> not have enough labor value to purchase the rights to print anything nor
> will firms have the incentive to invest in personal desktop fabricators to
> produce. If the capital gains issue is not addressed in our minicomputer era
> of fabrication, that issue alone may extend Neil's twenty year time frame
> unnecessarily.
I'll admit that I have long put off reading Neil's book, but on the
other hand, I've had enough contact with people who have read it, and
people that work with him daily, or something, to know what it's
about, and enough flipping through it while in the used book store to
also know that some of the discussions on the open manufacturing list
have already gone a little bit further than what he might have
explored in the book- if not, I'd be happy to go take a look at the
book of course, please cite and ref etc.
I do not think it's much of a problem of "labor values" or "economics"
here. The motivating statement for your email was supposedly your
opening quote: "... by helping everyone else acquire the technology
they need". How can we make this happen? By building the tools. How
can we make tools? Instead of being endlessly caught up in the
concerns of economics, the question of how can we make tools is a
rather technical issue- though not an impossible one.
An email on the topic of fabricational recursion
http://groups.google.com/group/openmanufacturing/browse_thread/thread/113d5a39898e061a?hide_quotes=no#msg_2000b6278e1af0ea
Dependencies, steps, reliability engineering, statistical models of
debian's ecology
http://groups.google.com/group/openmanufacturing/msg/9761b60732d5da4b
It will, however, require a shift away from thinking that current
businesses must- for some ideological or maybe some implicit or
axiomatic reason- or future businesses- support the closing of the
loop of technological dependencies and empowering individuals. There
is no (valid) sophistry here, no prose or politics, just the
undeniable truth of whether or not those fablabs *work*, whether they
are doing their job, or doing what they look they are doing- a topic
which Eric has poured many brain-hours into. For instance, do fablabs
actually shorten dependency paths when their hardware is mostly
proprietary-- what are we giving these "others" around the world in
these fablabs if they are unable to build their own tools, if they are
unable to 'replicate'?
http://groups.google.com/group/openmanufacturing/msg/2fccdde02f402a5b
http://groups.google.com/group/openmanufacturing/msg/f456aebde5952d03?
http://groups.google.com/group/openmanufacturing/msg/cc2256bf7814a1e4
Although I will not deny it would be nice, perhaps amazingly nice, for
what Nathan calls the "pitfalls of the present industrial system" to
take some positive collaborative interest in assisting individuals in
strengthening the tool availability coefficients and amplificable tool
building abilities,* I am not sure that their involvement would be the
catalysis or the "secret ingredient" for what we're hoping to
accomplish, so their non-participation would hardly push things back
20 years.
((I suspect there /are/ no secret ingredients overall anyway, which is
why it's so important to be proactive in terms of building, designing
and programming, since there are, in fact, things that need to get
done.))
Come to think of it, putting estimates in terms of years until some
vaguely definable technological state - like the technological
singularity - reminds me of Paul's commentary on some of Kurzweil's
books-
And recently I got wind of this abstract in my inbox for a paper that
I was vaguely involved with:
"""
Title: Changing the frame of futurism: From story-telling to
heavy-tailed, high-dimensional probability distributions
Keywords: Bayesian, futurism, heuristics and biases, decision support,
black swans, heavy tails, dynamic Bayes nets, stochastic differential
equations, continuous-time Markov processes, structural uncertainty
Abstract: Most futurism (e.g., Kurzweil) offers a single predictive
scenario, or, at best, small collections of qualitative scenarios.
However, results from the biases and heuristics program of research in
cognitive psychology (in particular, results concerning the
conjunction fallacy, the availability heuristic, and expert
overconfidence) suggest that focusing on single, intuitively
plausible, story-like predictions about the future is likely to lead
to inaccurate predictions and failure to account for extreme, unusual,
or unprecedented events. These systematic errors in both the public’s
and policy-makers’ judgment are especially dangerous as technology
increases in power, making possible unexpectedly large effects and
unexpected new interactions, on scales up to and including human
extinction.
Recent advances in probabilistic methods and computational statistics
allow an alternative. We can now build software tools that allow
individuals to consider, not single storylines as to how the future
might go, but structured probability distributions over huge spaces of
alternative future trajectories. More specifically, we can break the
future state of the world into distinct aspects, each described by a
set of quantitative variables. We can then allow users to specify
conditional probability distributions which describe how each aspect
of the world (e.g. computing power) influences the time-evolution of
its own future state and of the states of other aspects of the world.
In this way, users are naturally led to see the interplay between
their causal intuitions and the joint probability distributions which
those intuitions necessitate, and acquire a realistic view of how much
fine-tuning is needed for the world to follow any single preconceived
storyline. Heavy tails arise from user uncertainties about the
strength of each interaction and from combined effects of many
uncertain background variables.
We present an early web application, The Uncertain Future, for
allowing users to model timelines until human-level AI. Our software
allows users to input probability distributions with a visual,
click-and-see interface, and it presents collated expert opinions on
each relevant sub-issue so that users can see the range of expert
beliefs as they enter their own probabilities. We also discuss
possible extensions.
"""
Anyway, perhaps that would be an avenue to explore for those who would
be uninterested in technical details and exploring topics on their
own? But on the other hand, it's really not the same thing as dawning
your sharpie and drawing some microchannels to fabricate polymers into
specific shapes, or something- playing with statistics and
uncertainties and feelings and beliefs, versus actually, you know,
pushing atoms and potentials around in a more direct manner.
- Bryan
http://heybryan.org/
1 512 203 0507
* I've just made that up, please forgive me
Nathan Cravens
I'll admit that I have long put off reading Neil's book, but on the
other hand, I've had enough contact with people who have read it, and
people that work with him daily, or something, to know what it's
about, and enough flipping through it while in the used book store to
also know that some of the discussions on the open manufacturing list
have already gone a little bit further than what he might have
explored in the book- if not, I'd be happy to go take a look at the
book of course, please cite and ref etc.
I would say we've discussed mostly what's in the book. In many ways we've gone way ahead; in others we're still behind.
Open Manufacturing, including Factor e Farm, is still not as developed as a Vigyam Ashram or Honeybee Network. Those organizations, both located in India, have been around longer and largely developed out of necessity. The next step then is to invite members of these groups into conversation to help provide knowledge and business models and plans to accelerate OS fab lab construction.
I'll admit that I have long put off reading Neil's book, but on the
other hand, I've had enough contact with people who have read it, and
people that work with him daily, or something, to know what it's
about, and enough flipping through it while in the used book store to
also know that some of the discussions on the open manufacturing list
have already gone a little bit further than what he might have
explored in the book- if not, I'd be happy to go take a look at the
book of course, please cite and ref etc.
We've discussed things like this:
"The final frontier in rapid prototyping is to introduce functional as well as structural materials, in order to print complete working systems. Powders and plastics that conduct electricity can be used to print wires, there are printable semiconductors that can be used to deposit logic circuits, motors can be made with magnetic materials, combinations of chemicals can store energy, and so forth. Printable inks containing each of these types of materials have been developed and demonstrated in the laboratory. Integrating all of them into a printer is the most promising route toward making one machine that can make anything."
I could go on and on with quotes that have addressed topics on this list. In short, 'Fab' provides an introduction and foundation for discussion. Neil's our boy...
I do not think it's much of a problem of "labor values" or "economics"
here. The motivating statement for your email was supposedly your
opening quote: "... by helping everyone else acquire the technology
they need". How can we make this happen? By building the tools. How
can we make tools? Instead of being endlessly caught up in the
concerns of economics, the question of how can we make tools is a
rather technical issue- though not an impossible one.
I want to ensure the opening quote happens by addressing a foreseeable economic issue. That quote is mentioned because, in my view, it encapsulates the book as a whole.
Its the technical details that conflict with present Industrial economic models that I'm concerned with. Here's why.
In terms of business:
1) Self replication. This is good, profits can be made from this. The Green corporate model aims toward full automation, self assembly, cradle-to-cradle, ect.
2) Open Source. PROBLEM!! How do we make a return when everyone can have self replicating fab labs to make their own? I'll go bankrupt! And hey! I need to stay in business, because these labs do not make everything I had before for free. I still need to make money until everything is made at home on my desktop. People will not ask, "where can I buy that" to "how can I make that" to "I like that, I'll make that." soon enough. How can I ensure my well being when it cannot be fully supported without scarcity?
Sam Rose seems to anticipate this when he mentions Open Business Models provide a way for other businesses to continue after bankruptcy. People will be out of jobs and companies will go bankrupt faster than they can profit as businesses and the tasks that produce money generators self replicate without earnings or profit.
http://www.communitywiki.org/en/SamRose#OpenBusiness
An email on the topic of fabricational recursion
http://groups.google.com/group/openmanufacturing/browse_thread/thread/113d5a39898e061a?hide_quotes=no#msg_2000b6278e1af0ea
Dependencies, steps, reliability engineering, statistical models of
debian's ecology
http://groups.google.com/group/openmanufacturing/msg/9761b60732d5da4b
It will, however, require a shift away from thinking that current
businesses must- for some ideological or maybe some implicit or
axiomatic reason- or future businesses- support the closing of the
loop of technological dependencies and empowering individuals. There
is no (valid) sophistry here, no prose or politics, just the
undeniable truth of whether or not those fablabs *work*, whether they
are doing their job, or doing what they look they are doing- a topic
which Eric has poured many brain-hours into. For instance, do fablabs
actually shorten dependency paths when their hardware is mostly
proprietary-- what are we giving these "others" around the world in
these fablabs if they are unable to build their own tools, if they are
unable to 'replicate'?
I may well be jumping ahead of other issues to be solved, but I discuss economics because I see a potential developmental bottle neck when...
1) Half the products people use are made for free.
2) The other half are made by businesses that will eventually go bankrupt over time as production methods are further distributed and later personalized.
http://groups.google.com/group/openmanufacturing/msg/2fccdde02f402a5b
http://groups.google.com/group/openmanufacturing/msg/f456aebde5952d03?
http://groups.google.com/group/openmanufacturing/msg/cc2256bf7814a1e4
Although I will not deny it would be nice, perhaps amazingly nice, for
what Nathan calls the "pitfalls of the present industrial system" to
take some positive collaborative interest in assisting individuals in
strengthening the tool availability coefficients and amplificable tool
building abilities,* I am not sure that their involvement would be the
catalysis or the "secret ingredient" for what we're hoping to
accomplish, so their non-participation would hardly push things back
20 years.
If fabs are and remain open source, easy and affordable to replicable, and assume the role of what business and banks presently do until scarcity is dissolved -- we'll be fine.
I will develop this idea further here:
http://www.appropedia.org/The_Triple_Alliance
((I suspect there /are/ no secret ingredients overall anyway, which is
why it's so important to be proactive in terms of building, designing
and programming, since there are, in fact, things that need to get
done.))
I agree, we just got to get to it. Now what to get to? The stuff needed to make stuff still costs money. (?!)
Here are two proposal sketches that come to mind. Of course, to solve building fab labs will require more than a sketch.
1) Locate the most expensive item in demand that is most easily reverse engineered and assembled: sales of this item fund the purchase of more fabrication tools to make more complex objects.
2) The Ponoko model. Users send designs to be fabricated. The earnings go to generate more tools to make more complex designs: like jet engines.
Bake these plans until all fabrication tools are open sourced and fabs are replicated at the cost of intent made from your desktop.
Will you use these models to support your lab, Bryan? This can provide more tools to use when renting out the workspace. When its not rented out, with other models in use, stuff will still sell on Fabay, cash from PayFab, ect.
Hi Ted!
Come to think of it, putting estimates in terms of years until some
vaguely definable technological state - like the technological
singularity - reminds me of Paul's commentary on some of Kurzweil's
books-
http://heybryan.org/fernhout/
And recently I got wind of this abstract in my inbox for a paper that
I was vaguely involved with:
"""
Title: Changing the frame of futurism: From story-telling to
heavy-tailed, high-dimensional probability distributions
Keywords: Bayesian, futurism, heuristics and biases, decision support,
black swans, heavy tails, dynamic Bayes nets, stochastic differential
equations, continuous-time Markov processes, structural uncertainty
Of course!
These systematic errors in both the public’s
and policy-makers’ judgment are especially dangerous as technology
increases in power, making possible unexpectedly large effects and
unexpected new interactions, on scales up to and including human
extinction.
"up to and including human extinction." ROFL!
We present an early web application, The Uncertain Future, for
allowing users to model timelines until human-level AI. Our software
allows users to input probability distributions with a visual,
click-and-see interface, and it presents collated expert opinions on
each relevant sub-issue so that users can see the range of expert
beliefs as they enter their own probabilities. We also discuss
possible extensions.
Yes, the Dubai method. Silly, that.
* I've just made that up, please forgive me
I only suggest that you reverse the amplification loop via attentuation when needed. Automate it and call me in the morning with a damage report...
Nathan