Productive Recursion
Production methods lower the cost of products when simplified for rapid replication. That is called productive recursion.
From an email by Nathan Cravens of Effortless Economy
A recently adopted word, "recursion," has been useful in considering what is needed to create "a thing." Recursion is something touched on when describing casting the metal for LifeTrac. Some may want to purchase the product of the cast and save construction time. Others, knowing it is well worth the time toward the effort itself, will go "one recursion down" to reduce the financial cost of construction. The further up the constructive recursion, the greater the financial cost; the lower down in the constructive recursion, lesser financial cost follows.
As a general example, 4 hours to produce a design will mean not laboring for 60 days / 8 hours a day to purchase the same part for construction. The identification and presentation of contrast along these lines will, I am at liberty to suspect, will fuel this work into a widespread revolutionary movement. This example would also further dampen Luddite critique which argues "technology as toil." Contrasting design construction time with labor market time at minimum wage in addition to stressing the usefulness of the design in diminishing toil will assure the 'technological transfer acceleration' of the OSE format.
Productive Recursion Formula
Based on production rates in a foundry - $1000 per day of value generated (see Factor e Live Distillations Part 6) - and labor of $25 per hour for that time - or $200 - and one sees a 5:1 ratio of value generated to labor used.
Thus, any given device - say $2k in material costs - can be recursed to $400 in labor, or cost/5. To that, one needs to add the value of raw feedstock - say $200 if the $2k device weighs 2000lb, and we assume that scrap steel is 10 cents per pound. On top of $2k, one typically has cost/2 in labor - as for example with the CEB press, we are expecting the machine to require 20-40 hrs in labor at $25/hour - or $500-$1k.
So the price formula evolves to:
C = C_s + C_L + C_F
where C is the total cost, C_s is the cost of scrap steel, C_L is the cost of labor, and C_F is the cost of foundry labor.
We've observed that C_s is typically C_d/10, where C_d is the cost of a virgin materials for a given device - such as $2k for the CEB press.
C_L is typically C_d/2, as observed empirically from the CEB press. C_F = $200, or one day's worth of labor at foundry duty, which can produce at most 2400 lbs of cast parts per day, and can produce the necessary steel for most objects in one day - assuming say 8 shots, with multiple castings per shot.
For most electromechanical devices of low complexity (Tractors, CEB presses, steam engines, etc), the cost structure is therefore:
C = C_d/10 + C_d/2 + 200 (dollars)
= $200 + 3 C_d/5
Or more generally -
C = $200 n + 3/5 C_d
where n is the number of days to produce the steel castings.
For devices of 2000 lb or greater weight, the formula is essentially
C ~ 3/5 C_d
For the CEB press, we come up with C = $200 + 3/5 * 2000 = $1400 for the recursion cost where we went down to metal casting of components.
This is the most basic formula that includes one recursion, but with several recursions (say not only structural metal, but also components like hydraulics, or components of components ) - the formula may get more complicated. However, we know that the upper limit of the 3/5*C_d term is C_d - the industrial cost.
On a side note, this theory is worthless, even if the simplest case - until we gain empirical evidence. For example, we should be able to show that we can produce the CEB press at $1400 if we start with scrap steel - which would be a testable experimental proposition.
Summary
The above section has shown that instead of paying $3000 for at cost production of the CEB press ($2k in materials, $1k in labor) - which is already a factor of 10 lower than the competition - we'd be producing it at $1400, or about half of that - for factor 20 cost reduction over the industrial counterpart. If typical industrial product is 4-6 times above cost, we are saying that we can produce items at 8-16 less cost than the mainstream industrial production. This is a hint at abundance economies, though it requires physical proof for validation - by building the item in question from scrap steel in a flexible, community-supported fabrication facility.
Social Dynamics
Who organizes the method of production above? The productive agent is someone like Factor e Farm - who organizes crowd support of a fabrication facility. The price above (8-16 cost reduction) applies to turnkey products, which may be sold as kits. DIY fabrication allows one to produce infrastructure from scrap.
Who is accountable for the production facility? There is no mystery there. The producer maintains the production facility - the crowd supporters reap the benefits of its existence. But the supporters are not required to maintain the production facility, so that accountability is placed in the correct hands.
Critique
The Gingery books show an example of constructing an entire metal workshop from scratch. This is a sad case, however, for replicability - as it takes many months to accomplish this feat, while ending up with relatively light-duty equipment.
This is NOT what we are after.
The open source Multimachine is a more replicable example. This project is technically sound from the heavy-duty perspective - but it also lacks a level of standardization that makes the project truly replicable. So does MechMate, the free router table plans - where specialized parts require a laser or plasma cutter to fabricate.
The question of recursion feasibility boils down to absolute simplification of design and optimization of casting technique to make it feasible. These are the two issues that need to be solved for the recursion formula to be proven.
In particular, what is the most effective and quick route to metal casting? Molten metal can be had, but it appears that the ability to make quick molds for casting is the limiting step. It is worth exploring whether casting can be as effective as suggested above.