Science should be machine-readable

actually the machine, and its handlers should become science literate, if there is an interest in contributing to science.

dissemination of novel idea is the purpose of publication. sharing data can be done with e-mail. if you cant work with ideas and concepts, that is your limitation, it is backwards to suggest scientific publication should just become another data scrape to facilitate AI adoption.

... so says a study funded by Anthropic, rather self-serving then ...