5 ms·
Chinese (and some other) AI companies buying fast access to their dataset.
by atemerev 3mo ago
Chinese (and some other) AI companies buying fast access to their dataset.
- mmooss 3mo agoSo Anna's Archive is in some ways a front for AI companies, gathering the sources they can't get themselves?
- Cider9986 3mo agoThey could get access for free through torrents.
- HDBaseT 3mo agoMeta did, numerous Meta IP's appeared in the torrent swarms. The problem with using the torrents is they are slow, most of the larger sets have <3 seeders, with many dropping in and out, on home or slow connections. I would imagine other companies have learned from Meta's mistake and don't want to appear in the swarm either, which is why direct access is preferred. 100k for unlimited books access is nothing compared to the other costs these labs incur.
- Cider9986 3mo agoI'm under the impression that hiding IP addresses is easier than financial transactions, but both should be easy for a trillion dollar company. But I suppose they could use a shady intermediary company with a don't ask don't tell policy which would make high-speed access the final result.
- atemerev 3mo agoCorrect, but torrents are slow and intermittent, and $100k is nothing. People behind Anna's Archive are openly advertising fast access for datasets. I think this is net positive.
- inigyou 3mo agoNot really, AA is the source.
- mmooss 3mo agoBy 'sources', I mean the original books, papers, etc. For example, in this context, the American Heritage Dictionary and On the Origin of Species are sources (of knowledge). I don't mean your local bookstore or library where you acquire them.