The Problem of Production - and why tinyML can help

Ernst Friedrich Schumacher was living in a very different world when he published the controversial economic study Small is Beautiful. In 1973, the internet was far from ubiquitous. Combustion engines were unchallenged by electric powertrains. The first ever cell phone call had just been made – something so commonplace in 2023 that it's far from noteworthy.

Yet the world of 1973 and 2023 is also very much the same. People witnessed an energy crisis much like the one we are experiencing due to the Ukraine conflict. The Watergate political scandal in the White House had stirred the media, in many ways similar if not superseded by the slew of political scandals we see and hear about nearly ever day. And perhaps most importantly, newfound fears of overpopulation and climate change had captured the public imagination, causing a rift in economic thinking and planning. 50 years later, one may say things are very much the same.

Why is this important? Why should anyone care about an almost obscure economic text and the future it described? There is one simple reason:

In Schumacher’s own words: “the problem of production is NOT solved”.

The greatest misconstruction of the industrial evolution, improperly amplified by the age of information, is that we can treat use of natural resource as income. That is to say, the production of goods made from earth’s assets with the unimaginable efficiency of modern systems has given the improper illusion of income. Natural resources are taken as expendable and replenish-able, when in fact, they are not. Goods produced from these resources, which all things are, from plastic wrap to propane tanks to smartphones and TV shows, should be treated as capital. They are finite, and once spent are gone forever.

This is why the problem of production is not solved. It was not solved in 1973 nor 2023. Yet the growth in population and acceleration of environmental exploitation, under the guise of technological modernisation, has made us less aware of this problem.

If this all seems too macroeconomic and theoretical, imagine the following:

Your income can be spent, as by design you will work to earn more. Income you consume does not reduce your overall financial health, so long as you have capital, like savings, a house, stocks, or other assets that are maintained in balance. Now, let’s say you start to pay for things using your savings, mortgage your home, sell your stocks at a loss and must liquidate other assets to keep afloat. It’s not hard to see how this will soon destroy your financial health.

At the global level, this is precisely what the modern economy is doing. We are spending our capital: minerals, water, petrochemicals, soil, forestry and nearly everything the earth gives us like it is income that can be replaced. A data center providing computation consumes vast amounts of energy to power semiconductors made of materials mined from remote parts of the world to power a vast array of activities our lives depend on. Some might say this problem of production has been solved, for do we not have nearly democratic and affordable access to the unbelievable power of internet-connected smartphones? Surely this is true! However, we are spending earth’s capital, its finite assets, to do so. By treating these goods as income, we sully the very gift that makes production possible. The problem is it will take lifetimes to see the capital run out. The problem is that to modern economics, people don’t matter. The smallest unit in this grand system is overlooked as the most important player in the game. This is why Schumacher felt compelled to contradict decades of economic thinking.

Bigger is not always better.

Modern progress is predicated on growth.

Yet bigger is often more fragile.

Nothing can grow forever.

I am not an economist, nor an expert in climate change or financial systems. I am, however, a devoted member to the field of energy-efficient machine learning. As an AI entrepreneur, Co-founder of Deeplite, and the latest chairperson of the non-profit, tinyML Foundation, I’ve spent almost my entire career trying to make incredibly powerful AI- products more accessible and affordable. So, when I picked up a dusty book with a funny title after the recent tinyML Summit in Burlingame, California, I have to admit I was scared shitless by what I read. Not only because of the immediacy and significance of Schumacher’s thesis, but because I couldn’t see a way to solve it. Big Oil is bigger than ever in 2023 SaudiAramco & ExxonMobile profits are at all time highs). The world’s most valuable companies produce vast amounts of batteries and semiconductors that exploit the earth’s capital to produce ungodly income for powerful corporations that create modern life. The biggest thing in my own field – Generative AI, epitomizes the insane consumption of resources we treat as income instead of capital. What could ever turn this tide and stop Mother Nature from hemorrhaging?

I don’t think there is any one answer. Sustainability, ESG and carbon accounting has indeed been embraced by many businesses in varying magnitudes and manners. But as Schumacher said, we must treat economics as if people mattered. As if every person mattered. In this way I found some reprieve. As a person who knows the benefits of tinyML, having just seen the best experts, startups, and companies present their latest developments on how to make AI and ML more energy efficient at the 2023 tinyML Summit.

tinyML is more than just adding AI to sensors and small devices. It’s a paradigm shift in how we can build modern systems that are more mindful of resources and natural capital. In fact, it’s a steppingstone into a future where technological, economic, and environmental progress are jointly possible.

The problem of production may never be solved. The world we live in today will likely be very different from the one in 50 years’ time. It may also be very similar. I simply hope that one thing does become true: we treat our environment and economy as if people mattered. It’s quite possible that alone may solve many problems.

As Schumacher said: "“Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius — and a lot of courage — to move in the opposite direction.”?