Evolving programming languages in the AI era

This article examines how programming languages may evolve in the AI era, divided into two parts: reflections and agentic tooling. The reflections section explores the impact on programming languages, their ecosystems, and communities when humans write fewer lines of code. Meanwhile, the agentic tooling section discusses how tools should adapt to coding agents as users.

In the reflections section, the author contemplates how the loss of human code-writers may affect programming language communities. Different programming languages have unique cultural aspects, such as Python's emphasis on simplicity or Ruby's focus on programmer happiness. With fewer humans writing code, these communities might struggle to maintain a sense of belonging. Should communities preserve their existing identity or adapt to a new norm?

The agentic tooling section delves into the tools that should be developed to support coding agents. The author argues that coding agents may reduce the cost of building ecosystems while weakening the factors that typically bring communities together. The author also questions whether we'll need programming languages at all if agents can write machine code directly or if programming languages will continue to evolve with new syntactic affordances to accommodate coding agents.