Prompted by a mixture of thinking about the wider research that the Indie Manufacturing project falls under, the recent Maker Assembly, and yesterday’s celebration of 40 years since the Lucas Plan, I’ve been pondering the idea that we should aim to “ship bits, not atoms”.

It’s a nice concept. Rather than transport physical goods round the world from centralised factories we should send the instructions on how to make the goods over our digital networks to the place (or a place much nearer to) where the good is needed, and only “convert” it into physical form close to where it will be used. That way we minimize the transport cost - usually the environmental cost reduction is more important than the monetary cost, but you would reduce both. There could also be savings in waste - because goods are only realised as tangible items when they’re wanted, and so there’s no oversupply; and you might also reduce warehousing costs as there’s no longer a need to keep stocks of product spread out geographically in order to keep delivery times short.

For certain classes of product it works. Opendesk is a good example. The furniture is constructed from commodity, well-understood materials and the assembly process is reasonably straightforward. The customer can easily see the quality of the wood their desk will be made from (and there are established industry grading systems) and most customers probably have the requisite skills to build one from years of experience assembling Ikea furniture.

For more complex products, such as consumer electronics, the advantages of just-in-time individual assembly are less clear. Electronic circuits are made from a much larger range of source components and although you can encourage designs to pick from a stock library of parts, there will always be one or two parts which are specifc to that one device. Plus electronic components get cheaper pretty quickly as the purchased quantity rises. And assembly is also more complex - both in the tools required and in documenting the steps involved in building the finished item.

In these cases, I can see three possible futures playing out:

  1. The master craftsperson. In this version there are networks of craftspeople who have the requisite skills to produce a well-polished finished product the first time that they make one. They’ll have the plans and instructions to assist them, naturally, but they’ll need the attention to detail and conceptual skills to realise how the instructions translate into actions in order for the item to come together in the way the designer intended.

  2. Lots of “good enough” or “not quite good enough but we can live with it” products being made. Where the workshops manufacturing these haven’t made enough of them to learn the tricks in putting them together properly, and so end up making things that work, but the fit-and-finish leaves a lot to be desired.

  3. The completely de-skilled assembler, who can’t fail to assemble the product successfully.

I assume anyone proposing the completely distributed factory of the future is aiming for either the first or the third of those.

The first is very appealing but will surely result in expensive products. It feels like an updated version of the Arts and Crafts movement. William Morris with a 3D printer…

My guess is that most people think that the third is the most likely (or the best compromise between achieveable and desirable). In the ideal world the assembler would own their own manufacturing equipment, so while it’s not necessarily taxing work, there’s more variety in what’s produced and they get to profit from the output of the machines.

I think this does a great disservice to the tacit knowledge accumulated by the workers as they go about their tasks. I remember a conversation with one of the members of DoES Liverpool, who worked in the Halewood plant of Ford (and then Jaguar Land Rover) his entire career (while desiging and building CNC mills in his shed in his spare time). We were discussing workers moving between different jobs in the workshop in order, I proposed, to make things more varied and so less boring.

While agreeing that was true, he countered that such an approach would noticeably slow down their work-rate and they’d much rather stick with the particular machine with which they had the greatest expertise. They all knew how to use the different tools, but would have particular aptitude or just a finely-tuned familiarity with that particular instance of the tool as to be much more productive on it than their colleagues.

I’ve experienced it myself. As you become more experienced with the tools in the workshop, even on such “hands off” items as a 3D printer, you’ll be able to tell just from the sound it makes when something isn’t right. Or you’ll spot that the filament isn’t quite extruding smoothly and tweak the temperature, or nudge the bed to minutely raise or lower it in order to get the first layer to properly adhere.

The tools will get better and better (particularly when the people operating them have the skills to modify and improve them), but it will remain the case that the hundredth item you make of any product will be better than the first.

The fully-automated replicator is an alluring dream, but it feels as though it comes from a naivete about working with materials and the real world. The pursuit of that dream has undoubtedly moved our technological capabilities on immensely, but does the aim of replacing the workers blind us to the greater possibilities of augmenting their abilities?