When people ask me what I studied for my PhD, I tend to mumble something about “computer science” or “engineering”, but although it contained elements of both, neither is strictly accurate. In truth, my field is Cybernetics, the little-known and rarely trendy study of circular and self-regulating systems. Taking its name from the Greek word for the man who steers a ship, it encompasses control systems, robots and, yes, mechanical limbs, but also less obvious things like endocrinology and the population dynamics of rabbits. What unifies these very different systems is their circular, self-referential nature: they control their own futures.

Self-regulating systems are defined by their negative feedback: they resist attempts to push them out of equilibrium (positive feedback, despite the name, is usually bad). The homeostasis that keeps your body from freezing, overheating and drying out relies on this. Systems of this kind will resist any attempt to destroy them, even if it means warping beyond all recognition. This runs counter to how we usually think about built objects, which are very often engineered to be hard, but brittle. From reinforced concrete and steel plating to firewalls and encryption, the things we make are resilient only until broken, so prioritise not breaking at all. This is not a strategy you will see often in nature, because no living thing can afford it. No exoskeleton can be thick enough to deflect every possible tooth, claw and stinger that nature has to offer, so each organism must be able to recover from injury.

This does not mean healing like Wolverine: most creatures cannot regrow a leg, but they will adapt to their changed body and keep living. This is an unattractive principle to most designers because it means losing control over your creation. It is clear that a crab remains a crab so long as it scuttles around, but when does a bridge cease to be a bridge? If it “loses a leg” be technically traversable, but it certainly isn’t the same structure that passed its risk assessment. Easier to think of artifacts as simply intact or broken.

But self-regulating systems don’t so much break as die, and this includes any system humans are part of. A company takes much more effort to found than to run, and even more effort to salvage. The humans within a company are its feedback loops, making constant tiny adjustments to keep it running, and they can do this long past the point where it would otherwise have collapsed. As a result, a company’s material outputs tell you very little about how close to the brink it is: one can be perfectly profitable while only a single sick leave or redundancy from disaster. Trying to fix a spiralling company via lay-offs is like trying to cure disease by bloodletting: a complete misunderstanding of the mechanisms at work.

In fact, the modern world seems almost hostile to self-regulation. The recommendation algorithms at the core of almost all social media, online stores and streaming sites are classic positive feedback loops; reflecting and amplifying the things you have engaged with before. That these systems would be toxic and unstable is, frankly, obvious to anyone to anyone with a basic education in dynamics or control, but the designers of these systems tend to be drawn from a much narrower pool of knowledge. We also can’t ignore that these algorithms are very good at encouraging compulsive engagement, which looks very good on graphs, even if the route to sustainable profit is much muddier.

I am not suggesting we recreate the Cybersyn project (look it up though, it’s fascinating), but cybernetics continues to have great applicability even today. Thinking not only about what our machines and organisations do, but how they keep doing it in spite of an uncertain and changing world is only to our benefit. It may not be as glamorous as implementing flashy, radical solutions every couple of years, but it saves everyone a lot of time, money and stress.