my hot take that the nerds don't like is that the idea of two files being "the same" is socially constructed. even on the strictest technical level, having the same bits in the same order is neither a necessary nor sufficient condition for identity. but also any digital artifact comes to us with important context that shapes our interpretation of it, of which only a small part is digitally encoded (where did i get this file, who made it, how was it made, how is it stored, etc.)
4
Replies
2
Boosts
... which is why i think it's weird when AI boosters respond to LLM bans in open source projects with the line "well how would you even know if it's generated by AI?" as though the content of the contribution was the only (or even primary) factor we use to determine the circumstances of its creation, instead of like... all the other contextual and paratextual information we have about it
