Taste Is All That's Left

In the world of software development, the author reflects on the evolving nature of creating software. They note that the hardest part was not making the software exist, but rather dealing with its unpredictability during development. The author recounts their journey through building web applications and the countless hours spent on typing, reading manuals, and resolving API issues.

The point of no return, they call it, was production. It was the ultimate hurdle that separated the skilled developers from those who merely talked about it.

Fast forward to the present, and the author observes a significant change. The barrier between idea and artifact has collapsed to an almost negligible distance. The once laborious process of building software has become faster and more efficient. However, the author warns that this progress comes with a hidden cost. While it may seem that the value created by mastering the initial development process is gone, they argue that it is merely relocated.

The skills and judgment honed through the arduous climb of production walls still hold immense value.

The author questions the relevance of asking whether machines are reliable. They argue that the output generated by these machines is generally good enough, and that is the problem. The notion of "good enough" dissolves the drive to improve. The author emphasizes that the economics of software development have changed. The cost of making things is no longer a filter that prevents the creation of a thousand mediocre variations. Instead, the decision of what is worth keeping now relies on the developer's own judgment.

The author introduces the concept of "taste" in software development. They describe taste as the intuitive, wordless verdict that developers reach when evaluating multiple versions of the same function. Taste is not a matter of personal preference, but rather a gut feeling that separates the right from the wrong. The author draws parallels to the abilities of a good mechanic or editor, who can recognize problems before they even know the reasons behind them.

The author explores the origins of taste and its development. They suggest that taste is not something that can be acquired through osmosis by consuming excellent work. Instead, it is built through repeated experience of making mistakes, learning from failures, and internalizing those lessons. The friction of the development process acts as a curriculum, teaching developers what is worth paying attention to and what is not.

The author warns that removing the friction for developers will have unintended consequences. Developers will become fluent in coding faster, but they will miss out on the crucial apprenticeship that comes with the hardships of development. They will learn how to make things without ever understanding when to stop or when something is truly wrong. The author argues that taste is a slow process that requires paying the full price and being forced to endure failures before arriving at the right solution.

In conclusion, the author presents a sobering reality. Developers who have mastered the art of taste, despite the challenges and the slow process, will still ship at the same speed as those who lack it. While the world may celebrate their productivity, the author highlights the quiet cruelty of the situation. Taste is a deliberate, time-consuming process that demands patience and perseverance. It is the final filter that ensures the quality of the software we rely on.