I just got a Captcha with a phone number in it. When I googled the number, it was for Bloomfield-Montclair Ford in New Jersey. When I entered it, I got bounced for a “possible hacker attempt”
I wonder if directories are being scanned, and not just literature.
There’s lots of stuff in there. Google has recently been getting house addresses decoded this way.
Somewhere at one of the more serious online sites, perhaps nytimes or washingtonpost or newyorker there was a article within the last several days on how someone was using Captcha to correct old words that the OCR scanning got wrong. Perhaps try Googling reCaptcha for a starter should you be really keen on learning about it.
A lot of old trade directories, as well as periodicals with ads, are part of the Google Books and Archive initiatives. But all the recaptchas I see seem to amalgamate portions of words so they would not be recognizable as words. I’d be quite surprised to see seven digits in a row that were originally lined up that way. More likely to have the end of one number and the beginning of another.
Every once in a while I get one in Fraktur. I just marvel that they somehow know I can read it.