Agents write most of the code nowadays. So should you care whether it chooses to write Java, Kotlin, Rust or anything else? The language is now an implementation detail, natural language turned into the new abstraction. So the case is closed, right?

As we are collectively exploring how to use these powerful models best, some things seem clear: the feedback loop, guardrails and safety (aka the harness) matter. Let’s explore what our favorite languages bring to the table in that regard and whether language features are still important today.

Did the Agent indeed kill your favorite language? Or is it the very thing that helps it thrive?