I've been contemplating how advancements in AI will transform society and our way of life.
As AI takes on more intellectual labor, it might seem that humans will no longer need to think. However, I believe that a different kind of thinking than what we've traditionally considered intellectual labor will be required of humans.
This is similar to how humans were largely freed from physical labor through mechanization, yet were required to engage in different kinds of physical activities.
These different kinds of physical activities involve delicate work with hands and fingertips. This could be skilled labor like that of an artisan, or operating computers and smartphones.
Similarly, even if we are freed from intellectual labor, we cannot escape the intellectual task of thinking.
So, what kind of intellectual activities will be required?
In this article, I will introduce my thoughts on the paradigm shift in software development in the age of AI, and explore our destiny as beings who must think.
Process-Oriented Software
I propose process-orientation as the next paradigm, moving beyond object-orientation.
This is an approach where the central module of programming is a process. A process is triggered by events or conditions, processed by various roles according to a defined sequence within the process, and eventually terminates.
Thinking of this entire flow, from initiation to termination, as a single unit fits human intuition.
Because of this, software and systems can be understood primarily through processes, from requirements analysis to implementation, and even to testing and operation.
After implementing the main processes in a system, auxiliary processes or processes for adding new functionalities can be plugged in.
Some additional processes may start with events or conditions independent of the main process, while others may start when conditions are met by the main process.
However, even in such cases, there's no need to modify the main process. It's sufficient to define the added process to start when the main process meets its starting conditions.
Furthermore, because a process is treated as a single module, the definition of the process includes all the processing it performs.
Not only that, but a process also holds the aforementioned starting conditions, as well as variables and data areas for writing information required during processing.
Since processes are treated as unit modules and contain all necessary processing and data areas, there's a high possibility of redundant implementation of processing and structured data across numerous processes.
One option is to make these common modules, but it's not wrong to instead steer towards allowing redundancy.
Especially with AI assisting programming, it's conceivable that having many similar but distinct implementations across multiple modules could be unproblematic.
The commonality of processing and data types primarily aims to reduce the amount of program code in the developed software, making it easier to manage and understand.
However, if the cost of managing implementation code is significantly reduced by AI, the necessity for commonality diminishes.
Therefore, the policy of avoiding complexity in software structure due to commonality and instead defining all processing and data structures individually for each process, even with much redundancy, is perfectly reasonable.
This signifies a shift from a mindset of global optimization to one of individual optimization. This is because not having commonality allows for individual tuning of similar processes across different modules.
Individually Optimized Society
Similar to software applying process-oriented thinking, in a society where AI-driven automation leads to high efficiency and productivity, the mindset shifts from global optimization to individual optimization.
This is a phenomenon that can be called an individually optimized society.
Our society has various common values and standards, such as rules, common sense, manners, and general knowledge.
However, if these are strictly applied to all situations and circumstances, inconveniences arise in many exceptional cases.
Therefore, while emphasizing common values and standards, we allow for flexible judgments depending on individual situations and circumstances.
These might be explicit exception clauses in rules, or rules stating that judgments should be made on a case-by-case basis. Furthermore, even if not explicitly documented, they can be implicit understandings.
For example, various exception clauses are explicitly stated in laws. In addition, even if they cannot be explicitly stated in law, sentencing is influenced by individual cases through the judicial system. Extenuating circumstances are precisely the idea of reflecting individual situations.
Looking at it this way, it becomes clear that the concept of individual optimization, which originally involves carefully checking the individuality of all situations and circumstances and making judgments based on that individuality, is already deeply ingrained in society.
On the other hand, it is certainly inefficient to judge every single thing individually and carefully. Therefore, in an era where high efficiency is crucial, global optimization is sought.
However, as society becomes highly efficient through AI, the value of pursuing global optimization diminishes. And an individually optimized society, where careful judgments are made for each individual situation and circumstance, should come to fruition.
Subjective Philosophy
Making individually optimal judgments based on the situation or circumstance means that instead of immediately applying common judgments, one must deliberate.
I call this ethical view, where the act of deliberating itself holds value, "subjective philosophy."
Every event always possesses a unique individuality "now" and "here," distinct from other events. A corresponding responsibility is imposed on "me" when making a judgment that takes this individuality into account.
Making a standardized judgment that ignores individuality and fits a mold, or abandoning deliberation and making a haphazard judgment, is unethical, regardless of the outcome's quality.
Conversely, even if the judgment's outcome leads to unintended consequences and something bad happens, if that judgment was sufficiently deliberated from multiple perspectives and accountability is fulfilled, the judgment itself is ethical.
Thus, as we become able to move beyond the concepts of efficiency and standardization, we will enter an era where on-demand individual optimization, or subjective philosophy, is demanded.
Framework Design
Whether in philosophy, society, or software, a framework—a conceptual structure—is crucial for optimization.
This is because the direction of optimization changes depending on the perspective from which each subject is viewed and how it is evaluated.
From the perspective of global optimization, a framework needs to abstract various things highly and make them as simple as possible. In this process of abstraction, individuality is lost.
On the other hand, in the case of individual optimization, it is desirable to grasp and evaluate events or subjects from multiple perspectives, tailored to that specific event or subject.
In the case of global optimization, only a handful of people were sufficient to consider what kind of framework should be used to understand various things.
Most people could simply understand, evaluate, and judge things according to the framework devised by that small number of people.
However, in the case of individual optimization, many people will need to devise a framework for each individual matter to appropriately grasp its individuality.
Therefore, the ability and skill to design frameworks will be required of many people.
The Destiny of Thought
Organizing things this way reveals a future where even if AI takes over the intellectual labor humans have traditionally performed, we cannot stop thinking.
We will be freed from intellectual labor for productivity and material wealth. However, the individually optimized society and subjective philosophy will simultaneously demand that we design individual frameworks for each matter and deeply deliberate.
This places us in a situation where we must continue to think, perhaps even more so than in current society.
AI can perform intellectual labor and make judgments that anyone could make. But for matters for which "I" must bear responsibility, AI can only provide information, present judgment criteria, or offer advice.
The final judgment must be made by "me." This is the same as how, even now, one can consult with authority figures, parents, or friends regarding various personal decisions, but cannot delegate the judgment itself.
And in an era of highly advanced efficiency, not engaging in deep, individual judgment will become unacceptable. This is because the excuse of being too busy to think due to the demands of life will no longer be valid.
In such an era of advanced efficiency, we will not be able to escape the destiny of thought.