README
I created this blog on the 3rd January 2026. It is 74k words, across three physical notebooks, written interchangeably by me and Heather. To transcribe it, I took photos of each page and got Claude Code to process each image using its built-in OCR and convert to text.
I would expect there to be 1-5 significant word errors on every page (multiple pages per entry). How many will depend on our handwriting on the day. Tables / images / footnotes etc will all be processed badly.
That said, for every image I checked, the start and end were right, and the overall meaning was there.
I asked Claude to differentiate our handwriting and prefix sections with "Brian: " and "Heather: ", but it seems to have done this only a tiny handful of times.
So, for the original text and/or to identify who the first person is in any given section, check the original images. At the time of writing, these are in an album in my Apple Photos collection, in 2004, side by side with the actual photo album of the trip
The Claude Code project (CLAUDE.md, python code, images and original transcripts) was added to the "Blog Diaries" folder in Google Drive, which was subsequently archived (in 03. Various)
Comments
Post a Comment