Recursive self-improvement describes a loop where AI designs superior versions of itself. While I. J. Good predicted an intelligence explosion in 1966, current progress remains incremental. Researchers now debate whether current tools actually automate model architecture or simply optimize existing parameters. This distinction determines if true autonomy is near or merely a marketing trope.