In a recent video on his YouTube channel, antirez, the creator of Redis, introduced the concept of "Automatic Programming" to describe the evolving process of software development with the help of AI. He notes that for some time now, he has been using this term to refer to the practice of writing software with AI assistance—what he believes will soon become the standard way of developing software altogether. He emphasizes that the results of automatic programming vary significantly depending on the human guiding the process. Unlike passive AI generation, automatic programming involves active human involvement: intuition, design decisions, continuous feedback, and a clear vision for the software. This is fundamentally different from what he calls "vibe coding"—a method where users provide vague, high-level descriptions and let the LLM generate code with little to no oversight. In vibe coding, the developer often only intervenes when the output fails to meet expectations, offering minimal input into the actual design or implementation. antirez argues that when developers are deeply engaged in the process—understanding the architecture, guiding the AI’s decisions, and shaping the code step by step—it’s not accurate to say the AI “wrote” the software. Instead, the software is the developer’s own creation. He points out that even though LLMs are trained on vast amounts of human-written code, the final product is still a reflection of the programmer’s vision, choices, and intent. He also highlights that the pre-training data used to build these models was created by humans, making the AI’s knowledge a collective human achievement. In this sense, using AI to write code is not about stealing or appropriating—it’s about leveraging a shared intellectual foundation to amplify individual capabilities. He likens it to being part of a collective mind, where each person can now accomplish more than they could alone. Automatic programming, he explains, is not just about generating code—it’s about directing it. It includes decisions on how to implement specific functions, how to structure the system, and how to ensure quality and correctness. The most critical part of the process remains the human’s understanding of what needs to be built. To illustrate his point, antirez references Redis. While Redis doesn’t introduce groundbreaking technical innovations—its core is built from well-known data structures and networking code—it became widely valuable because of the clear, thoughtful vision behind it. The success wasn’t in the code itself, but in the ideas and design principles that guided its creation. In conclusion, antirez asserts that programming is now automatic, but vision remains human. The software produced through this method is not the AI’s—it’s the developer’s. And that’s something to be proud of.