On Fri, 28 Nov 2025, Guy Fedorkow via cctalk wrote:
Does anyone have an approach that's more
efficient than generic OCR
and a long process of correcting typos on every line of code or comment?
My wife often gets had written notes or just as worse, photos of
screens[1] from which she has to re-type in the contents.
She uses "Claude" or "ChatGPT" with remarkable efficiency. However
note
that this is written text rather than computer code, so the results may
vary.
Gordon
[1] No. Don't ask why the senders can't send the original document. It's
just not worth it - they're as thick as mince.