Engineering Consumption – Mattereum – Humanizing the Singularity

Deming’s Key Principles

• Accountability came from statistical process control.
• Transparency came from Deming’s emphasis on a culture of openness and clear communication.
• Trust came from the non-violent, non-destructive correction of systematic errors, leading to goods and services that consumers could trust, because the people working together inside of organizations to produce those goods and services could trust each-other.

These are Mattereum’s values. We hope we have learned from the best.

This is how our civilization was made and is maintained. It was made by teams of people who learned to trust each other enough to admit what is wrong, to correct systemic problems, and to work together in enormous numbers and at vast, unbelievable scale, to manufacture the complex artefacts required to run the operating system of modernity. Without the social transformation, statistical process control alone does not deliver the desired results.

If we could figure out how to get more effective cooperation spanning all aspects of the supply chain — just as we did for all stations on a production line — what might be possible? Why do investment, production, consumption, and waste have to be so split up across organizational boundaries without effective information sharing arrangements? Why can’t we share information all the way down the value network associated with an object, rather than just along the supply chain it was manufactured on? When I buy something, why isn’t all the information about the object — right down to the CAD files it was cut from — transferred to me as a standard consumer right?

This whole process, this machine, the modern industrial supply chain is without a doubt the pinnacle of human civilization, the rich technical mycelium which gave rise to the moonshot and the curing of polio, the green revolution and everything else. It is everything from tractors to GPS satellites. It is the closest thing we have to a replicator. But in its current form it no longer serves our interests well.

Except for a very lucky few, most of us feel like things slipped off the rails somewhere in the past. Something went wrong with our culture. Exactly when, and how, and what went wrong depends on your point of view. The diagram which shows worker wages stop rising in line with worker productivity from 1971 is the thing I show people when I want to talk about “what went wrong”, but it’s only a symptom. Others might point to the Treaty of Versailles, or the ban on personal ownership of gold in America in 1933, or any one of fifty dozen points in history. But, Sometimes it seems like it does not matter how many good seeds we sow, what comes out of the ground is spears.

We have to negotiate with reality in a new way to get traction on the situation.

The mechanisms we have been using to get good outcomes are not working any more: the mechanisms we expect to help the system self-correct all seem to point the wrong way at the wrong time, and the levers of change appear to be connected to the 8-track player in the dashboard, not to the wheels. We can change the music, but we can’t seem to change the direction. We are skidding.

Big Tech was supposed to help us with this, but it wound up largely as an outgrowth of the surveillance state. The blockchain was meant to fix this, but that political ambition has been almost entirely diluted out by the dreams of taking over the financial services industry (which might be a worthy short term goal, but let us not confuse it with saving the world). And saving the world is table stakes these days, as you might have noticed in the news.

We manufactured the wrong things

Capitalism is pretty dumb. Price signalling is a very limited stream of data between buyers and sellers, and learning in capitalism is notoriously slow. If I see a bike on sale for $400 and buy it, is that because I’m desperate for a bike RIGHT NOW, or because I look at the price, say “well, that’s a pretty good deal”, and then get rid of my perfectly functional old one? Would I have paid $500 for a model with a slightly better seat and a better color than grey? Would I have paid $3000 for a carbon fibre model, were one available?

If all we’ve got is price signalling, the only way to move forward is to manufacture all the alternatives, advertise them widely and see what people buy. This is an evolutionary “bloom and prune” approach, and fields like market research which attempt to divine or anticipate buyers’ needs are often very inaccurate because people’s self-reporting about what they want is often very inaccurate. Just asking people what they want doesn’t cut it either. It is hard to fine-tune capitalism to manufacture what people want, and it is even harder when the advertising loop starts to take control of the process, not just telling people what is available, but actively trying to make them want things they did not previously want. Do you really want this thing, or do you want this thing because we made you want it? If we created the demand we are serving, are we helping anybody at all by meeting these imaginary needs?

That vortex, that infinite regress, has distorted the feedback systems inside of capitalism to the point where nobody knows what they want any more in any kind of solid, consistent, clear way. It’s created a huge and poisonous semantic fog which has taken away our ability to know ourselves, because the human mind was not made to reason clearly when fed 5000 ads per day. We evolved in a relatively slow moving, information poor environment without the written word. In a fast moving environment, dominated by marketing messages and skillfully composed advertising copy, and images produced by some of the most technically competent artists in the world, is it a wonder that we can’t think straight about what we want?

In fact, ironically, the only place statistical process control is applied to consumer behavior is targeted advertising, in which some of the best minds of our generation collude to gather vast portfolios of data about our personal lives, and use it to try and drive buying behavior without any fundamental model of people’s needs or wants, only their expressed preferences.

Is it a wonder that we’ve soaked up the entire capacity of the wish fulfilling tree of industrial mass production making fashionable junk that nobody needs, and still can’t seem to find a way to get everybody access to the drugs they need to stay alive, even basics like antidepressants or insulin?

We are squandering this plenty that statistical process control and quality control gave us, and it is wasting our lives, and killing the world.

Slow progress on optimizing the waste streams

Massive process engineering work has been done on the Production function of our society, over centuries. We have a name for it: the Industrial Revolution. Because it was a revolution. Quality control was also a revolution, but a quieter one.

In finance, something similar happened. The investment function we discussed earlier also absorbed statistical concepts to manage how money moves around. Over time this became known as Quantitative Finance, and started to hoover up a disproportionate and frightening number of the brightest minds in physics and math. Enormous efforts have gone into building these systems, and they are uniformly amazing. They’re competing head to head, so there are winners and losers, but the actual quality of the work in quantitative finance is phenomenal.

But what about the consumption and waste functions?

Start with waste. Has landfill gotten radically brighter and more efficient over the past 40 years? Maybe a little. But compare it to what has happened in manufacturing over the same period, and essentially our waste management is unchanged. There’s a lot of talk about recycling, and bins everywhere, but the actual reuse of that material in ways which prevent further raw materials being pulled out of the ground is a lot more complicated than people hoped when the recycling movement got started. It’s too early to call post-consumer recycling a failure, but all too often it just means dumping in poorer countries.

Recyclers simply do not have the tools or resources to combat the sheer scale and complexity of the waste. Although we have lots of incremental progress on pulling value out of industrial and post-consumer waste, are these systems really massively more efficient than what we had before? Do we measure, weed out variation, and make maps? Only here and there, and only in certain industries. Steel is pretty well recycled, plastic not so much.

But compared to the sophistication of the industrial processes that produced the plastic bottles we are throwing away, the recycling side is standing still in comparison.

Targeted advertising isn’t much help on consumption

Consumption is hardly better. Yes, there has been progress, but most of that progress is on selling people things they don’t need, ever more efficiently. The level of consumption has outstripped all imaginable process improvements in making that consumption efficient. Let’s talk about some of the measures taken, and the impact they have (not) had.

So the first question is how do we know what we want. You know the general theme: your entire click stream is used to model who you are, so that advertisers can compete at auction to show you signals to control your behavior. And because we are not evolved to deal with these kinds of cognitive attacks (yet!) they are partially effective, enough to pay for progress and improvement in the fundamental techniques. They are still getting better at this.

And let’s not forget, targeted advertising is something like 90% of the profit at Google and Facebook, and also a contributor to Amazon’s revenue stream. How do they know what to recommend to you? A huge part of the economy of the internet is using statistical methods to understand and change consumption patterns, but in the crudest and least-effective possible way. “You get what you measure” remains the dominant fact of life, and measuring click streams and credit card purchases only measures one step of a four step process.

We have optimized only half way through investment, production, consumption, and waste, and no further. We should not be surprised that there are problems. We should be measuring consumer satisfaction and environmental impact, too. Then we would get a better world.

But, ugly and adversarial as the advertising attention-parasitism game is, and dangerous as these information dossiers on us all are, it is still an attempt to apply statistical process control to the consumption system, and it is happening on a truly enormous scale. It takes exactly the same kind of reasoning which was used to make the manufacturing system efficient, and applies it to the consumption system, it just doesn’t push far enough into the consumption system to measure “consumer satisfaction per unit of environmental impact” which is really the sort of thing we ought to be using all this fabulous machinery to measure. And it does not push into the waste system at all. Odious as it is, it may turn out to be the right approach, just applied with too-shallow insight into the problem domain.

The targeted advertising system does what it does with a very partial model of the system it is trying to optimise, and only the most crude and short-term definition of its goals (evolutionary in the worst possible way, non-cognitive at the lowest level). But it does sell product, and it’s paying for the creation of enormous datasets about people and what we think they might want. It just doesn’t ask those all important questions: “Are you still using that thing you bought? Do you like it?” Nor does it know what the thing is made of.

Of course the problem is that it measures the wrong thing: measure spending, get spending. Measure satisfaction, measure progress towards our stated life goals, and maybe get those things instead. Price in environmental damage, and get another thing again, a green economy. The problem is that spending is irrational, and behavioral economics factors thwart the eudaemonic potential to turn large scale datasets about people into the common welfare. We have made computers just smart enough to feed on us like attention parasites, but not smart enough to be good and faithful companions like dogs or horses. I cannot yet meaningfully say to Google “find me a good book to read, but make it a little outside of my norm, the last few suggestions have been a trifle timid” and get any useful intelligible response, but the damn thing won’t stop showing me adverts for books it wants me to read. The stupid system is incapable of partnership, it only knows how to hustle and distract us. We have not gone far enough to get positive results, only negative ones. The system is building momentum, but it is still below the threshold of revolutionary change.

This ad targeting system is a relic: it’s 1960s Mainframe era ideas about production and consumption, running in the 21st century.

Let’s look at other areas where we have large-scale successful attempts to use statistical process control to optimise consumption. Uber tries to put cars where it anticipates there will be riders. It uses price signalling to encourage drivers to turn up during periods when it expects things to be busy. Uber plays the game. Still they seem to have wound up in the same trap as (for example) Apple hardware manufacturing: a truly great service for some people, at the expense of the labour rights of others. The algorithms optimized resource allocation in the pursuit of everyday low prices (or good quality goods which defy our expectations about what phones and tablets can be like, year after year after year). But these systems still wind up treating human beings like machines at every level, and this trend has to be identified and banished before we wind up in Marshall Brain’s vision of dystopia.

Amazon is optimizing what is in the warehouses closest to you, based on the ability to get most of what you order to you same day or next day. This is a fantastic example of using statistical process control to optimize outcomes and serve people’s needs: whatever it is you bought, it’s more valuable to you if it arrives quickly. This is an unmitigated good, again extracted at a very heavy social cost: Amazon warehouses really have picked up the reputation of treating people like a cheaper version of machines. Amazon would certainly automate all the way, if they could.

And that’s basically it for using statistical process control and quality control to optimise consumption. We haven’t made any strides on the scale of the industrial revolution in the consumption system. Not even close. We need this to happen.

Production is already pretty lean, at least for many of our bigger systems. But making consumption lean? The work has not even begun yet.

How might we begin to make consumption processes as efficient as production processes?

Consider the quantity of stuff we buy in the course of our lives for what amounts to experimenting with our identity — goods we purchase to understand ourselves better, or to foment personal growth. We may want to take up the guitar, try your hand at fly fishing, or start learning Tae Kwon Do. As beginners, we often have no idea what the best kit is to start off. So we end up spending large amounts of time tracking down reviews and recommendations, many of which are contradictory. We wind up either getting the most expensive equipment presuming that price equals quality, or the cheapest with the best overall reviews, knowing the goods will have to be upgraded it if the hobby sticks and becomes part of our identity over the long run. All we want to do the experiment. What we don’t want is to be left lugging around the gear if we don’t like the results.

This process of experimenting with identity is entirely natural. But make no mistake, this process leaves residue: each of us has that stash of stuff sitting unused in closets, attics, and garages; the remnants of hobbies and identities which didn’t quite fit, and were set aside. We may manage to move some of it along by selling it online or at a yard sale, but for the most part it all just sits there. Old clothes. Ice skates. A helmet or two. Silk paints. Did I ever really wear this much tweed? Apparently so.

The transaction costs of getting rid of this stuff are too large for us to take action: the stress of making the decision to sacrifice a slab of our investment and just sell the damn clutter, the time it takes to list on an auction site, the physical logistics of posting it to a new owner, and managing the customer service overheads involved in the entire process often leave the goods (and therefore the capital they represent) stranded.

The other approach, just writing them off and throwing them in the dump, seems wrong — both a waste of money, and of the materials embedded in the objects. So instead of doing the rational thing and getting rid of it, we wind up using nearly every square inch of the Boomer generation’s basements, attics, and closets as a sort of informally specified, unsearchable, distributed warehousing solution as the massive superabundant flow of goods from our hyper-optimised production system hits the analogue slackness of our consumption systems, and simply pools in a huge lake of underutilized or obsolete things. There are tens of millions of metric tons of this kind of waste in America, and it all has value — if only we can find it.

The production systems of the world run on Enterprise Resource Planning (ERP) systems, of which SAP and Oracle are probably the two best-known examples. Similar systems exist in the world of finance to manage capital inside of banks, and to allocate resources in private equity firms. This is the software which runs civilization’s arteries and veins, its digestive system and its lungs. It’s the nervous system of industrial capitalism, and without it, we would almost all be destitute.

But these systems are corporate, intimately tied to the investment and production phases of society, but only very weakly tied to consumption and waste management. They are, essentially, direct descendants of the mainframe paradigm: one big computer that rules the whole organization.

And these systems interoperate only with great reluctance; the world is not run by a big, interwoven, interoperable mesh of big ERP systems seamlessly talking to each-other to make optimal decisions. It’s all still largely stuck in the mainframe phase, on arcane standards that are impossible to parse, and worse to debug. In short, these systems are due for an upgrade.

What we need is ERP for the People.

We need smaller, more flexible software systems to help individuals manage the same kinds of tasks that ERP systems handle: physical assets, time, money, commitments and more, as integrated systems. We would all really benefit from having tools that bring the power of knowing what you’ve got, where you have it, what you paid for it, and what it’s worth to somebody else right now. Imagine how much it would change if it was all at our fingertips in a series of dapps which help us optimize our personal relationship with matter itself, mediated by the marketplaces we all participate in, plus new marketplaces for information about the quality, provenance and value of physical objects.

Our working title for this model is Effective Abundance Platforms: platforms which help us manage our relationship to the abundance that industrial capitalism produces, while optimising the hell out of the inefficient capital allocation mechanisms which are represented by error-prone purchasing and reselling behavior among consumers. It’s clean, it’s green, and we think, with Mattereum in the lead, it could be extremely profitable as a new class of businesses.

The global challenges posed by climate change and resource scarcity are driven by many factors including imprecise capital allocation, a financial system with poorly defined boundaries, dependence on polluting energy, and outdated methods of industrial production. Mattereum is working to solve these problems by creating digital twins of material objects and using blockchain smart contracts to automate all aspects of how material things are traded, owned, and combined. We aim to squeeze out these systemic inefficiencies and more accurately allocate capital to activities which promote wellbeing.

Our universal naming system for physical objects enables the creation of efficient markets for information about the composition and qualities of physical things. This accurate information will let society use the same statistical process control techniques which revolutionized manufacturing and investment over the past two centuries to completely transform consumption and waste management globally.

Over the last few decades, the revolution in production efficiency has changed the world. We know a similar revolution in consumption efficiency will follow suit, if we set out on the right path forward now.

And we need to squeeze every last grain of efficiency that we can out of the global economy, because people are still hungry, and structural waste on a finite planet is the enemy of everything that lives. If the internet has a purpose, if the blockchain has a purpose, this has to be part of it: we aren’t just fighting against authoritarianism, we are also fighting against entropy.

Food rotting in the back of the warehouses does not have to happen. We just need efficient systems to connect hunger to food, and at least half of that problem is just bad software which harms the sellers as much as the hungry buyers.

We are all on the same side against waste, and bad software. It’s all of us, against entropy.