# Conceptual Inflation > As the lower bound of what society in general expects from a given technology gradually rises, the meanings of terms we have long used gradually change as well. We start to wonder why a car is called an automobile, or why an AI washing machine is called AI. How will the terms we now use as a matter of course come to feel in the future? As the lower bound of what society in general expects from a given technology gradually rises, the meanings of terms we have long used gradually change as well. We start to wonder why a car is called an automobile, or why an AI washing machine is called AI. How will the terms we now use as a matter of course come to feel in the future? ## Automatic Programming and the Automobile In the old days, automatically converting code that was not machine language into machine language was called "". Put simply, anyone who uses a compiler or interpreter is already doing automatic programming. But to our ears now, that term sounds very awkward. *That* counts as automatic? If you think about it, calling an automobile an automobile is also strange. It was called an automobile because it moves without being pushed or pulled by the force of a horse (a carriage) or a person (a rickshaw), but now that too sounds very awkward. *That* counts as automatic? ## Artificial Intelligence The term "artificial intelligence" is similar. In the 1980s, if something had a microchip built into it, people would first and foremost call it "artificial intelligence" or "smart." For example, if a washing machine had a microchip and provided a simple function like "delayed washing," it was called an AI washing machine. If we go back further, to the era when mechanical calculators had already been commercialized yet "computer" still referred to people who made money by performing calculations as a profession (a mere 100 years ago), arithmetic operations were regarded as an intelligent act that only humans could perform. A small number of mathematicians were researching "calculation by mechanical procedure," but for the overwhelming majority of the public, calculation was considered an intelligent act that only humans could carry out. But now, if you say a desktop calculator has "intelligence," it sounds very awkward. *That* counts as intelligence? Someday, the spectacle of people being enthusiastic about 2022's [ChatGPT](https://wiki.g15e.com/pages/ChatGPT.txt) 3.5 will feel strange. For some people, it may already feel awkward now. The impression that one could [feel the sparks of AGI](https://wiki.g15e.com/pages/Sparks%20of%20artificial%20general%20intelligence%20-%20Early%20experiments%20with%20GPT-4.txt) in ChatGPT 4.0 will soon come to feel awkward too. > Even if a system satisfies the definition of AI, once we understand how it works and the problem it is trying to solve, it is no longer regarded as AI. --1985, , [Software aspects of strategic defense systems](https://wiki.g15e.com/pages/Software%20aspects%20of%20strategic%20defense%20systems.txt) ## Agentic Coding When "automatic" and "artificial intelligence" come together, they become what in 2025 we call [agent-based coding](https://wiki.g15e.com/pages/Agentic%20coding.txt). I suspect expectations for this term will rise twice as fast. Systems that we now regard as "agentic" may before long no longer feel agentic at all. What sort of awkwardness will people in 2030 feel when they look at the "agentic coding" systems of 2025? My thought is that, for starters, we may have to replace "coding" in "agentic coding" with something else. Won't the layer likely move higher than it is now, into something like "agentic software production/operation"? Having machines handle coding on their own will be far too taken for granted. What about code review? Right now the code review stage is the bottleneck (even before agentic coding, code review was originally a bottleneck, but now it has become even worse), but going forward it will probably seem natural to think there is no longer any need for humans to review things. ## Agentic Management What if we move up one more layer? Separating out software production and operation and thinking about them alone is, from a business perspective, no more than local optimization. To achieve global optimization, do we need a concept like "agentic management"? In that era, will managers be pushed out if they "cannot prove to the board why a human manager, rather than an AI, is necessary"? As they are being pushed out, will managers have thoughts like this? > Back when I used to force developers to "prove why a human developer, rather than an AI, is necessary," I truly had no idea that would become my own situation. I thought that at least managerial decision-making required the advanced expertise of humans.... If we are very, very lucky, a world may come in which all production is automated by robots and AI and just distribution takes place—but before that world arrives, what processes will we have to go through?