I read about this yesterday on an HN discussion about a different method: https://news.ycombinator.com/item?id=36130535 https://news.ycombinator.com/item?id=36

I read about this yesterday on an HN discussion about a different method: https://news.ycombinator.com/item?id=36130535 https://news.ycombinator.com/item?id=36130535 What makes this interesting and unique: Instead of scribbling on redundant pieces and relying on error correction to preserve the meaning, we can engineer the encoded values to create the picture in a code with no inherent errors