The Technical Is Political

When someone frames technology as apolitical, remember the saying of how the devil’s finest trick was convincing the world he does not exist.

The difference between knitting with your bare hands and using a machine to do so is not so much a matter of natural versus artificial, but a matter of self-sufficiency versus dependence.

Your bare hands might not win contests of speed or precision. Yet their allegiance will always lie with you. Frail as they may become with injury or old age, you know they will draw their full strength to support you.

The knitting machine does not feel like an extension of your nature. Not because it is not natural, but because it is not yours. Its allegiance lies elsewhere, with the investor who funded it, the engineer who built it, or the entrepreneur who lent it.

Yet rejecting machines would be the wrong conclusion. It is likely dependence you want to reject. Yes, machines often encode power asymmetries, as they require significant coordination and resources, with centralization being a cost-effective way of supplying these.

But cost-effectiveness depends on how you define cost and who gets to bear it. Loss of sovereignty is also costly, for instance. Once you factor in such losses, the top-down approach stops screaming efficiency.

There is this project called OpenKnit, and its poetic significance is obscured by the myth of apolitical technology. It is a knitting machine whose design can be freely accessed and forked into new versions, a machine that can be assembled for cheap from off-the-shelf parts and disassembled back into them, and one that allows users to “print” exactly the clothing they need, locally.

The false dichotomy between centralized technology and no technology has been weaponized at length. “This is an important technology, whose benefits far outweigh the costs, so our company should put it at your disposal, because missing out would be backwards.”

The conclusion here does not follow. Yes, we might want to have technology at our disposal, but that does not mean that it should effectively be governed by a select cadre of corporate leaders, irrespective of their nationality. Even more so if the technology is truly important.

Technology builds on other technology, so to reclaim a technology you have to reclaim the stack it sits on. You start with the interstices, placing seeds in the cracks of the stack, then gradually extend roots deeper into the technologies below.

OpenKnit requires some parts to be 3D printed by default. Initially, you may concede to using an off-the-shelf 3D printer to get things started. But over time, you may venture to break down the underlying technology and extend your roots a step further by building a Voron Trident, a 3D printer whose design can be freely accessed and forked into new versions…

To start with a heap of sand and try to speedrun your way to ChatGPT is unrealistic. Much more sensible to start with an open model on rented servers, then gradually reclaim the server using bare metal accelerators, then gradually reclaim the accelerator using reconfigurable chips, then gradually reclaim the chip using fab shuttles, then gradually reclaim the fab using a homegrown process.

Start somewhere familiar and you will have a usable alternative at any given point along the arduous journey of reclaiming the stack. An alternative whose authenticity only grows over time, as the roots extend deeper and deeper.

When someone argues that they should make decisions on behalf of everyone instead of everyone making decisions themselves, then they are either naïve or lying. It is sometimes hard to tell.

Language and aesthetics can help make explicit the relationship between technology and sovereignty. Technology whose many underlying ingredients have been reclaimed could be described as being deeply rooted and grounded, as having roots that run deep. The cloud is the complete opposite of that, being ungrounded, nebulous, adrift. Somewhere in between are technologies that have taken root, in the sense of having started being reclaimed, yet their roots still run quite shallow. Bare metal servers do bring a partial sense of grounding, yet their parts remain closed until our roots pierce them open.

When a technology is being leveraged to gain dominance by estranging people from it, we could describe that as technology being uprooted. Managed services, proprietary extensions, and regulatory capture are popular ways of gaining moats. “Embrace, extend, extinguish.” Lacking the vitality of open participation, uprooted technology eventually starts to decay and wither away, a process often called enshittification.

Roots do not crudely transform whole volumes of soil into wood. They are sparse threads that weave through the soil to efficiently interface with its resources and anchor themselves securely. Reclaiming technology calls for a similarly rational approach to growth into the depths of the technical ecosystem. Dependency trees should be constantly pruned in order to remain lean, but they should be adapted to resources that are available from many sources in order to remain resilient. The chosen ingredients should be a handful, but they should each be plentiful.

Look up and take a few moments to scan the space around you for objects big and small, from furniture to devices. Really, just take a moment to glance around.

Sit briefly with the possibility of genuinely reclaiming a few of these, of understanding what makes them tick, of being able to manufacture them yourself. We live our lives relying on objects that we are completely detached from, many of which magically appear on our doorstep as the end result of a nebulous process that probably involves some workers abroad and some factories somewhere.

Now take the previous exercise and replace objects around you with pieces of software running on your devices. Really, just take a moment to browse your app gallery or installed programs.

What did it take for one of these to get here? What is it made of? What lies behind this interface? Even if we often uncover an intimidating amount of complexity when we draw back the curtain, we might occasionally find that things boil down to simple parts. Either way, it is healthy to interrogate our surroundings and deepen our relationships with the systems we interact with day after day. Even the furniture you currently sit on may otherwise slip through as almost virtual, an abstract notion with favorable properties.

To grow to new heights without falling over, you must first grow into new depths.