I seem to recall OpenAI saying they guarded against captcha solving...
this post was submitted on 30 May 2024
1101 points (98.8% liked)
Funny
6894 readers
377 users here now
General rules:
- Be kind.
- All posts must make an attempt to be funny.
- Obey the general sh.itjust.works instance rules.
- No politics or political figures. There are plenty of other politics communities to choose from.
- Don't post anything grotesque or potentially illegal. Examples include pornography, gore, animal cruelty, inappropriate jokes involving kids, etc.
Exceptions may be made at the discretion of the mods.
founded 2 years ago
MODERATORS
Dead internet shall become undead.
I'm just saying, but captcha had a purpose. It still kind of does. Whether solved by a person or by an AI.
I'm pretty sure that for a good while there it was using captcha to help its text recognition more accurately determine what words were from scans of books that were imported en masse to Google books as images of pages. We're talking about books published before computers were used to write them. The text recognition algorithm had an idea of what the letters should be, but didn't have a high enough confidence in the result, so it was sent through captcha to get a consensus from humans.
The humans answering the captcha would just verify whether it was one of a small list of possible matches, and in doing so, train the machine vision algorithm to better detect the letter in the future.
That's what I heard at least. IDK. I just live here (on the internet).