Wasn't this just linked to here a few days ago with tests showing it has atrocious accuracy, misses a significant amount of text, and takes an order of magnitud

Wasn't this just linked to here a few days ago with tests showing it has atrocious accuracy, misses a significant amount of text, and takes an order of magnitude more time (and several orders of magnitude more energy) compared to known OCR solutions?