npub1yhj4ua6580h3w0kucwwrtwl4uv9pg7m2drxzplq5x3364jczp36s4h3h3t (npub1yhj…3h3t) Out of genuine curiosity, and I guess npub1slyc6w2fy6t7sk7e5caxclu2dn629l35n4tkjyxzrqndkpq86x2sfut8j7 (npub1sly…t8j7) also has something to say here: 10-20 years ago, there was the (still ongoing) debate on filesharing and media piracy. One of the arguments there was "look, you cannot meaningfully crack down on this due to technology, and if you do anyway, the societal costs would be too high".
That sounds eerily familiar to what the big AI proponents say today.
So, is their argument equally valid? And if not, why? (My gut feeling is that the societal cost for regulating the AI business models is a lot less than completely prohibiting filesharing. On top of that, filesharing is mostly only forbidden because copyright is borked - but then, isn't copyright one of the main venues of argument by those who critize OpenAI et al?