Efficiency and complexity will be points of vulnerability, but that will not stop the manufacturing of some products or services from being complex, nor will it drive people away from the essential value of efficiency. What we see as efficient in a more complex calculation with multiple values, may of course change.
The advent of natural language processing (NLP) has enabled artificial intelligence to explore the world of human culture and learning without guidance. Computers are no longer bound simply by their original instruction set and initial data. From here on knowledge, instruction and design will be stored, processed and developed semi-independently of homo-sapien intelligence. The more structured the language of a discipline the more amenable it is to NLP. Areas like recipes, law and technical vocabulary are likely to be the easiest to adopt by NLP systems.
This codification of instructions, of craft and practices has already allowed machines to outperform humans in basic motor skills and the co-ordination of technical documents. Whether a jury is ready to listen to the arguments of a robot lawyer without discrimination is another thing. Software artists, authors and musicians are emerging. There is barely a physicist left in the world who’s research is not mediated by a computer[i].
This trend will greatly increase the rate of knowledge production, and should in theory have an even greater impact on dissemination of design and best practice.
This ability to transfer instructions increases the capacity and efficiency of automated algorithmic production. The pressure on logistics, demand for personalisation and increasing minaturisation points to a greater modularisation of the fabrication machines as well as end products. One way to compete with large manufacturers with a high level of invested capital is to sell small versions of the machines used direct to the consumer. An example can be seen in the way Cinema moved to TV, and then mobile phones united both the production and consumption equipment in one device.
Manufacturers, large and small already offer customisations of products before purchase over the web. In these systems, the consumer is bought closer to the factory through information sharing and individualised delivery, rather than simple mass production. Retail becomes replaced by logistics.
This again produces a neural like configuration in the economy, with nodes of production and consumption dependent on communication and feedback between the elements. The only thing that prevents monopoly domination by the most efficiently tooled manufacturing centres are logistics issues.
This suggests that the degree to which local or domestic fabrication can replace logistics will in a large part be a principle question of political economy, such are the subsidies, explicit and implicit to big business across the world. Germany stands out as nation with an effective insitutional capacity to build and suppor SMEs and a number of countries offer support to small farmers. But by and large existing economic infrastructure is geared towards suppor for mass production and centralised distribution.
If the capacity to localise production is applied in an economy the likelihood is that some successor of the 3D printers, fab labs, or even RNA printers will all bring manufacturing capacity into our homes. This will be predicated on the production of smart materials with the flexibility to be used in many applications.
More complex production by end consumers will doubtless add to the complexity of waste collection and pollution (e.g. the printing of disposable tooth brushes, let alone mutagenic viruses in the rivers).
It may also be the case that if the focus is on biological of pseudo-biological manufacture (mushroom houses, algaes, bacterial assemblages) that there is an advantage in localised production.
Food processing is an area which already has wide range of domestic devices. It may even be possible, as with the mobile phone, to localise production and processing in one small unit. This might be relevant to artificial meats or bacterial feedstocks.
Another area where minaturisation combines with information transfer to distinct advantage would be in personalised medicine, the precise composition of which would in many cases be specific to variables within the individual’s body including time of day and what they last ate.
The reduced capital cost of machinery may produce a trend in some sectors to return to vertical integration of supply chains (raw material, manufacturing and sale managed in one organisation). It may evenly allow primary producers to add value early in the supply chain and go direct to consumers. We may see a new “artisan” movement that becomes an exercise in sharing instructions for “artisan machines”.
In the world of Financial data gathering, electronic currency and parallel computing the capacity should in theory exist to allow the movement of every penny in the economy to be charted. Predictive spending and credit could lead to completely different economic pattern with competition between players pushed several steps into the future.
Against this illusion of near certainty in currency flows we have to set the prospective of rapid and ongoing disruption. That is, the pattern of resource use and distribution is changed significantly by a new market entrant. Disruption by technology and fast moving technology companies has become an industry gold standard. The launch of mass market company backed by good code and massive advertising spend is what most “unicorns” seek.
Disruption will arise not only from innovation, it appears to be principle emerging area of soft competition between multi-national actors, the closure of borders, denial of services, disruption of logistics, often in a way that has been calculated to inflict a “valuable” disruption to the other party and presumably, often planned with the aid of strategic AI looking several steps into the future. The very ability to forecast brings with it the ability to infer causality in the pattern and thereby change it. So while every data point in the economy may be trackable, black swans will appear and the river may burst its banks or meet unexpected dams from a fallen tree or crashed truck.
There may also be disruptions through climate change (droughts with knock on effects), seismic events and social collapse likely to follow from ecological constraint.
Many of these trends point to arrangements where manufacturing will be more distributed, leading to processed commodities and smart materials to have more small customers rather than delivering to large centres. Or more probably, small well tooled producers and consumers meshed together with large centres of primary production and mass manufacture.
Given that economies of scale are a simple fact of maths, as is efficiency, these values will always be pursued almost by definition, by large organisations. In this multi-polar world the size and efficiency of an organisation becomes a point of vulnerability, most multi-national actors will need a State sponsor to protect them in a world where their competitors are increasingly taking on the appearance of instruments of foreign policy and resource acquisition for the host nation. This was to some extent, always the case with colonial enterprises, and a parallel can be seen the travails of national champions during the era of great power competition in the second half of the 19th century.
This dynamic itself suggests that the political economy that arises will necessarily be communist, corporatist or oligarchic, with the government needing business to acquire and distribute resource and large businesses needing government support to do so in the face of competition both at home and overseas.
And all the while in a political economy of rapid innovation, politico-economic powers will be looking for disruptors, those people who can efficiently apply a new idea or technology such that it can significantly change resource flow and market share. This means helping companies to scale novel technology rapidly, which itself becomes perhaps an economic disruption in the re routing of materials and a supply crunch caused by novel demand, or worse, an existential dilemma.
To scale rapidly to global production ignores the precautionary principle, and risks releasing a new uranium or thalidomide. To not scale a solution rapidly in a time of ecological crisis provides crisis through omission rather than action. Ideally, real time experiments in the economy should be monitored in real time, and if we are honest in the research base, seemingly far off, downstream effects of the change might be brought to light in simulations that can investigate everything from impact on local environment to chemical and metabolic change overtime into the future. An obvious case of this can be found in Climate science.
[i] There is the problem in a post Kantian world where we look out from instruments built by giants that if any of the underlying theory on which our instruments are dependent is wrong we can only built a grand edifice of nonsense. And this is problem that will in future lead us to the question of trust in AI. For if AI is bringing us both the theory and the data, if it is involved in our very perception and construction of the universe, then we have to trust that it is a competent an honest guide. What if the guiding AI of a branch of Science turns out to be the silicon equivalent of a giggling fraudster?