How far away are we from a 14mt clanking replicator architecture?

We are one step ahead.
The question is not how far away we are, but how small we can make it for practical purposes.
And how well we have documented the plan for tech-seed to grow.

The real growth of our technologies started in late 19 century and the beginning of 20th century.
The amount of information and technologies was manageable trough 20th century to about 1950–60. But then, since about 70-x it started to explode. And it began to be more complex since then, however, the basic core things do basically the same things as they did in mid 20 A.C.

Cars ride the roads, ships travel river and seas.

A lot of complexity was involved due to reducing human labor and increasing efficiency of the production — energy wise, speed wise etc.

We even can’t say that quality of the products is increased during the time in terms of the durability of the products. Some did, some did not.

As an example, a 100 years old lathe can be perfectly useful today for what it was created back then and there are such examples. Not sure we will be able to say the same about some 5D CNC, 100 years later. We have bargain simplicity and durability of our tools for different things like flexibility, less maintenance, speed, more precision with less human interaction, fewer materials requires to build them, higher energy efficiency, etc. But we still use relatively simple tools to get good precision when it is requires, simple old school tools — surface grinder, surface plates, precision stones, gauge blocks, optical flats, optical measurements.

We did get new tools like lasers, plasma cutters which speedup some things we do, extends our possibilities a bit, and overall improves our lives. This was a great journey and it gave us more than 12 million chemical compounds, it gave use 14-nm transistors and processors which use them, it gave use iPhones and Samsung S7, space rockets and space probes, games and Linux, air bearings and ferromagnetic fluid sealing’s, fighter jets, Concord, genetically modified life/organisms/bacteria, robots picking berries(not mass used, but they do exist), Boston Dynamics and its menagerie of robots, quadcopters and vlogging and all sorts of goodies and we are back to electric cars and utility blimps which were used 100 years ago.

Most important the journey gave us the knowledge. Knowledge how to do things differently, and how to do them better in certain ways. And both of those aren’t the same.

And by using those new toys we got the ability to develop other concepts which would take millennia to develop if our new toy would not be mass produced, cheap and faster than our old toys. By that, I mean big data, neural networks, physical modeling, numerical analysis, finite element method, tools for designing our new toys(different CAD’s and CAD systems), databases, and holy Internet. And it has to be understood that those new toys are a big part of the current R&D process and significant (but less significant(*)) part of production processes. 
((*)I’m not expert but it like compare G-code with the tool which has generated that G-code for the CNC or 3d-printer or another tool which uses that G-code to do the job)

South Bend 9-inch lathe, produced ca 1951

The knowledge and tools we obtained and created during the explosive industrial revolution allow us to look differently at the problems which production had in late 19 beginning of 20 century. One of the funny things is that to modify a 1900’s lathe for our digital world is easier than do the same with a lathe made in 1970’s. With the second it requires removing a lot of things they did back then to make their lives easier. We do use basically the same equipment as they did in 1900’s just addons are different. So, the question about clanking replicator — we do have it, as our tools are capable of producing parts of themselves and which with the help of other tools and human’s help produce those tools. We do have it, as basically, a stone-age knife has built (as a father) the everything we have. But unfortunately we not quite sure how it did that, and a lot of people are excited to watch his return and lessons on as an example Primitive Technology youtube channel.

Our bad memory, and that it depends on the will of owners of particular technologies(in a lot of cases) and that we didn’t have the incentive to make with our production factories the same we did with production itself — moving from hand made to assembly lines and mass production — is one of things which separates us from having clanking replicators. The reason a lot of people may think it isn’t there might be the current state of computers and processors production, the complexity of manufacturing facilities and of the processes thems selfs. But it is not an entirely true statement, it is true in parts of complexity, but it is not the reason because vacuum tubes computer will do basically the same as today’s equipment, but slower, and it will require a lot more materials to build and a lot more energy to work. Or the same with 1cm big transistors instead of 20nm transistors. Producing 1mm transistors is way much easier than 20nm, producing 0.5mm transistors also a lot easier, and basically, it is the technology which gave mass access to the computers for a lot of people — Z80 processors.

CPU Zilog Z8S180, Konstantin Lanzet — CPU collection (wiki:source)

Those processors are 2 orders of magnitude slower than current top notch CPU’s, but they are definitely capable of doing a lot of things. Since we hit the limits of Si-processors, and now we are moving in multicore direction with processors and graphic cards we do develop algorithms and tools to create software which utilize the multicore capabilities for the tasks where it is possible(not all tasks) and some tasks and problems could be solved on a multicore Z80 platform in the same way and time as they are solved on GPU cards and multicore platforms today, just more energy/electricity will be spend in Z80 case.

So, potentially we could use our tools, to design the system which is back in time technology level, but using tip&tricks we learned recently. It possibly will be less efficient, it will not surprise(positively) us with its speed, but it will produce itself and everything else what it can produce. We can do that even to replicate current technologies, but it will be a big setup and those are already in sufficient quantities as they are needed, so as long as demand for them isn’t there, there is no real incentive for big players to do so. (There is a history of such attempt by Japanese back then, but they definitely were not prepared to do so, in the first place because of the absence of adaptive algorithms, which we do use today, since not so long ago).

And the situation shows us another problem.
Yes, we are more sophisticated in designing tools (software), and they have become more accessible in last 20 years, and more mature, but we still not there in therms of accessibility and quality of such software which is available to general public. Big players do not have the incentive, and small players have no tools. Such project could probably cost $10’s millions to develop(or might be less) to demonstrate and potentially it could be done with help of the community over time, but so far projects like http://opensourceecology.org/wiki is one of the top achievements in the field. (btw, they have a nice list of CAD software with commentaries about them http://opensourceecology.org/wiki/List_of_CAD_Programs , my personal favorite: ‘NX — Price is “it doesn’t matter if you ask or not — you can’t afford it”. This is what they use to design nuclear submarines.’)

It is not only about software by also about the ecosystem of the development of such system, there should be online storage's and processor times devoted to such development, not everyone can have a super cluster or GPU farm and those might be needed to analyze the systems and find solutions. Thus before answering the question — “How far away are we from a 14mt clanking replicator architecture?” — it has to defined:

Which kind of technology’s level we do like to have?
How efficient we like the system to be?
Which subset of technologies should the thing be capable of creating besides replication?
Which environment it should work at?
How big it allowed to be?

Those are important questions.

If we do consider the moon or space like a place where such technology could be useful then efficiency could be not a problem in the case, because there is plenty of solar energy accessible by simple means, some troughts about that can be found there.
If we have some goal to achieve — as an example to have machines which we can teleoperate, it will be not important for us which technologies they do use, so long as they do the things we would like to be done.

So we definitely not there yet to offer universal solutions, to offer compact and efficient solutions for our technological level, but we definitely capable of generating more specific designs combining our knowledge and give old technologies a second life.

Group