There’s a common saying that “it’s easier to break than make something”. But is it?
To understand whether this statement is true or not, we should probably first explain what those terms “break” and “make” mean actually. What does it mean to create something, and what to destroy it? We should also try to understand why we would want to make or destroy something, what are motives and our intuitive understanding of actions that we want to take to do so. Then we are able to tell why we think one is easier than the other, and whether it really is.
To start understanding terms “create” and “destroy”, we should start by saying first what it means “to do something” at all. Universe has current state. Current state, with all of its forces that act within it, determines the next state that the world will end up. That’s quite deterministic and not very helpful. Taking into account the randomness of the nature itself (yeah, starting from the bottom – quantum world) and leaving the possibility of existence of the free will, we might say that we will do something that will transition us from the current state into the future state so that we are more satisfied/happy with that future state than we are with the current. For example, you are now hungry, so between deciding whether to eat or not, you realize that you’ll feel better if you eat, and that is why you do action of “eating”. You compared the possible outcomes of several actions that you know you could take, and you chose to do one of them. So you did it. So, with this example, we could say that “doing” is guided by our estimate of what’s the best thing to do regarding the future that we want to find ourselves in.
It’s also important to somehow measure how hard the action is or how demanding it is. In a purely physical world, that would probably be the total change of energy between the two states. For us humans, energy is not the only measure. We also have to care about time required for action to finish, social aspects of our doing, how it makes us feel while doing it… For computers, it’s more close to physical aspect, as it probably comes down to time required and power consumption, so – energy. So, we are generally concerned about the hardness of our work as some measure of energy and time put into it. Let’s call that a cost of a certain action.
So far we’ve decided what it means to do something (intentional state transition that leads to more satisfied and happy future) and how to compare our actions (by comparing their costs). The next thing is to define what does create and destroy mean. There is also one extremely important thing to note: all actions share one thing in common, and that is whatever the current state happens to be, and so the conditions within it. So both creative and destructive action start with whatever the initial state is before those actions happen. I’ll use this fact later.
Creating something starts with our vision what is it that we want to create. So, by estimating what’s the current state (blank canvas) and our goal (Mona Lisa), we determine the steps that we would need to take in order to transition ourselves towards our goal. We have the previous knowledge what are very small and controlled sub-actions that we can do that would help us converge to final state (brush strokes with selected colors). We only do those actions because they are safe way to get to our goal, and doing very controlled and small actions takes time. Usually, a lot of time, and thus energy. But here’s the real question: If we knew that a random splash of paint across the canvas at the exact moment in time would result in a Mona Lisa being drawn on that canvas, would we do it? Of course we would! So what stops us from doing so? Knowing the conditions. If we knew that at that exact moment we were supposed to throw bucket of colors onto the canvas, the mixture of colors in the bucket, their viscosity and other properties were perfectly matched so we end up with a Mona Lisa on the canvas, we would paint it that way. But we don’t have such information. We are unable to predict an outcome of such complicated event happening as splashing the bucket of colors is. And that is exactly why we take the more safe and “deterministic” approach – painting with many, many, tiny, well placed brush strokes. Those are the events that we’re able to predict, and so by taking one small step at the time we are sure that we would converge to our desired state even though it would probably takes us more time and energy to do so.
Usually, splashing the bucket of color is considered as “destructive” act. Simply because it’s really improbable that such action would have an outcome that satisfies us. Imagine throwing bucket after bucket of colors onto the canvases, and maybe 1 in a 10.000 would create picture that makes sense to us, and that an artist could say that he’s satisfied with because it’s good (not talking about other types of art where the mess and randomness are actually the desired result, I’m thinking more towards photorealistic painting and such). And it’s exactly that not knowing of the entire conditions within the current state and thus not knowing how probable (or should I say improbable) a certain outcome is that make a difference in our decisions. Destruction is doing some actions that we know are not likely to produce something meaningful, and so we do them in order to quickly come to our goal state, without much thinking about the steps we’re taking. But, still, destruction is a transition from a starting state to a goal state, but just with less precision, less thinking and thus more “random” and “messy” sub-actions. Those sub-actions require less time and thus energy, and that’s why we perceive them as easier to do (well, usually).
Destruction doesn’t care and does messy actions because the probability of them being meaningful or likely to get us even further apart from our desired goal is improbable. Creation means taking smaller and more controlled steps that we know will probably help us converge to final state if a predictable fasnion. Our actions are guided by our estimate of what is good for us to do to get to the desired state. And small and meaningful creative steps that take time are indeed harder (more costly) than destructive steps. So, the statement is (usually) true. If we only knew that random splash would be even better…
plus and minus give plus
It should be obvious that creation and destruction are inherently the same process, but just with different estimates and probabilities for their outcomes. As I mentioned before, both of those share the same initial state. What decides whether our next action is of a creative or a destructive kind? Here’s one example also: imagine painting an image and an artist makes a really messy stroke across the canvas. Seems that he has ruined the painting. But later in the process, he uses those messy shapes to create final shapes out of it more easily than he would if he hadn’t done that messy stroke and instead drew each shape independently. By “destroying” things he actually warped the conditions on the image (color mixes and shapes that exist on the current image) so he could benefit from them later on. Such thinking leads to question whether we can combine the destruction and creation?
One thing comes to my mind when talking about deterministic (creative) and random (messy, destructive) combination – pathtracing.
Pathtracing is a very weird mix between the two, which has positive outcome in the end. It is also interesting the better the random generator, the better the convergence. I’ll leave this as it is and let you think about it. But the answer, if we can combine them is indeed positive.
destruction is creation
I’ll mention again that both of these are the same thing, just different in their estimate (that’s why we can combine them). There’s also one example that I found to be extremely compelling for giving a proof that those are the same thing.
Computers know no destruction. They only have “constructive destruction”. Take for example objects from OOP concept. To create them, we follow very strict set of instructions to help us “create an object” (constructor). But to destroy and object? Well, we do the exact same thing, we execute the part of the code that “destroys an object” (destructor). Funny, isn’t it?
practical application (frameworks)
Another thing that I wanted to mention is creating frameworks. In computer programming terminology, those represent “a layered structure indicating what kind of programs can or should be built and how they would interrelate” and what interactions are to be expected between its components. What does a painter see as a framework? Different types of brushes and colors, of course.
We are capable of predefining and creating components that will help us create more complex things in the future. We expect a certain behavior and follow the rules of their usage and we get the expected results.
But, could we run a random generator that would output a sequence of bits that would represent the most advanced computer program ever written? Could we generate an artificial intelligence that way? Could we solve any problem by randomness and destruction? Well, it’s not random anymore if you can predict the outcome, but yes, if we happen to run that random generator that has the chance to output a code for any program that we want just by starting it with good parameters (good seed), then we would write programs that way. But because it’s very unlikely that such thing would happen, we write code line by line, following many programming rules and patterns, because we are sure that those will eventually get us to our goal. But, the randomness is always waiting for its chance.
Creating and destroying are the same process. Creation can be “harder”, but since those are the same thing, we are simply asking a wrong question. We should ask if action A is harder than action B, and that’s it. From now on, don’t say for yourself that you are a “creative individual”, but better say “destructive but ending up with desired outcome”.