This is an extremely misleading way of framing how engineers are actually using AI to generate code. It makes it sound like LLMs are autonomously writing code. But in reality, engineers are prompting LLMs to generate code. No precise prompt? No code. It’s almost as misleading as claiming that “code editors are writing 100% of our code.” It assigns agency to a computer program that is in reality a medium between the agent (the human) and the software application under development.

This kind of verbal trickery is designed by the elites to subtly manipulate people into believing certain narratives, in this case the narrative of AI autonomy in software development. As George Carlin once said: Pay very close attention to the language we’ve all agreed on.