An SVG is all you need

I use ChatGPT to compress SVGs, in particular QR codes. Many QR code SVG generators produce inefficient SVGs, and conventional SVG compressors often lack the understanding required for some compression techniques. ChatGPT can replace alignment indicators with <use> elements.

Is there a way to embed the data encoded in the QR code directly within the image? This would allow the data to be parsed directly by the browser, eliminating the need for computer vision to decode it again. Going further, for web images QR codes could be efficiently encoded and rendered by the browser.