A small script that changed how I read

Last winter I wrote a small script. It watches a folder on my computer where I drop screenshots of paragraphs I want to remember, runs them through OCR, drops the text into a markdown file with the date and source, and tags it.

That is the whole script. It is about forty lines of Python. I built it in an afternoon. It has been running on every machine I own for a year now, and it has, more or less, reorganized my reading life.

What I expected

A solved problem. I screenshot a quote, it ends up in a file, I look at the file later. Standard knowledge-management territory.

What actually happened

I started screenshotting more. The friction was so low that the screenshots became a behavior, not a chore. Then I started reading differently — slower, more underlining-in-my-head, more attuned to which paragraphs were the ones I would want to come back to. The capture loop changed the read loop. I did not see that coming.

Six months in, the markdown files started becoming source material for my own writing. I noticed which authors I screenshot most. I noticed which kinds of sentences. The pattern was a mirror.

# the entire script is roughly this import pathlib, subprocess, datetime watched = pathlib.Path.home() / 'Dropbox' / 'capture' out = pathlib.Path.home() / 'notes' / 'captures.md' for img in watched.glob('*.png'): text = subprocess.check_output(['tesseract', str(img), '-']).decode() stamp = datetime.date.today().isoformat() out.open('a').write(f'\n\n## {stamp} — {img.stem}\n\n{text}\n') img.rename(watched / 'processed' / img.name)

The smallest tools are sometimes the ones that change the most. I keep relearning this.