Hacker News new | past | comments | ask | show | jobs | submit login

We need a vintage data/handmade data service. A service that can provide text and images for training that are guaranteed to have either been produced by a human or produced before 2021.

Someone should start scanning all those microfiche archives in local libraries and sell the data.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: