Not so long ago most of our furniture was made by hand. A craft person would think of an item that could be carved out of wood, one that might be desirable and useful for other people. The carpenter would consider the material, the time it took to produce a piece, how to transport it, who may buy it, and how much should it be sold for.

That process was as slow as it sounds, and resulted in less pieces being made, of higher grade and less fluidity in their design. In modern terms, the carpenter was the investor, designer, distributor and seller.

To perform a net positive set of decisions the carpenter would need to rely on a wide gamut of nuanced signals. All the way from how the skin of a tree looked, to what ornament was trendy at the time, and how much weight can a carriage handle with the current state of its wheels.

More over, all of those decisions were not transferable to the next piece of furniture the carpenter will design and sell. Each piece came with its own questions and answers. The intuition to navigate the future unknown made a carpenter successful, much more so than any one quantitive decision over another.

You can imagine a pile of half–made dining tables and unsafe chairs behind the carpenter’s workshop, the result of trial and error needed to build their skill, and that core intuition.

Fast forward a few decades and we humans have now invented tools that take those decisions, the result of trial and error and various know-how’s  – and programmed them into machine’s instructions.

Finding a tree for our table has become industrial as well. We now have a consistent stream of material, cheap to produce and very compatible with our product line and packaging. Distribution deals have been sorted through various partnerships, and the company–we’re working as a company and not a mere sole proprietor now–is constantly optimizing every taylorist step between the factory and the end costumer.

With time we started using computers and sensors to monitor more aspects of the process. We could now see where we spend most dollars in getting to our customers, and what changes could we make to our lines of coffee tables to make them more desirables. We have data on where people hear of our products, how many sales did a certain ad bring in, and how much time a delivery track cost us with that wrong turn.

The question stands: what do we do with all of this data? Remember that our entire operation hinges on the knowledge that the carpenter rendered in chopping a tree, prep’ing wood and building a table. Give or take. We took that knowledge, quantified it, put it on an assembly line and mathematically optimized it.

When we had less computers and less data, that was the best we could do. There was no feasible way to project the same attention and thoughtful to each individual piece of furniture. So what we were left with were incremental improvements, of questionable importance.

On the one hand we have our line of products, factories and staff – and on the other we had the market, with consumers, advertising and competitors. Our job was to find the most sound mathematical formula to slice and match the two. To be more successful we might engage in recursive thinking, questioning what has changed since the last time we sat together and adapting portions of our operations accordingly.

We were industrialists because we had to. There was no way to produce anything any other way. You could not make an affordable chair, and stay in business without optimizing every aspect of your operation.

This brings us to today. I propose us to consider a future of decentralized assembly lines, reminiscent of the carpenter, the tree and the dining table.

Just that this time we would not need to do any of the work. We now have AI, ML and other data tools capable of the flexibility needed to address different instances of instructions.

We are arriving at the limits of where linear, industrialist and mathematical thinking can take us. Making is becoming cheaper by the day, deep interoperable and creative thinking will be the next wave of everything.

This article originally appeared at Nitzan's blog HERE

Nitzan Hermon is a designer and researcher of AI, human-machine augmentation and language. Through his writing, academic and industry work he is writing a new, sober narrative in the collaboration between humans and machines.